Slashdot Mirror
Incorporating Machine Learning into Firefox 2.0?
blakeross asks: "I will be doing research this summer at Stanford with Professor Andrew Ng about how we can incorporate machine learning into Firefox. As we work to finish up Firefox 1.0, we're also seeking ideas that will make Firefox 2.0 blow every other browser out of the water. People who come up with the best 3-5 ideas that involve the use of machine learning will win Gmail accounts, and if we implement your idea you'll be acknowledged in both our paper and in Firefox credits. Your idea will also be appreciated by the millions of people who use Firefox. We'll also entertain Thunderbird proposals. See my weblog post for more details; I'll read all comments posted in response to this story or to my weblog."
Here are the best five ideas incorporating machine learning:
... "this is a picture of a sunflower") and let them group and search for items. Eg. "Pictures like this" or "Documents about cats."
1. Based on the user's browsing habits, automatically bookmark the most frequently visited sites, and automatically put them into *multiple* categories (not just one category) to make them easy to find.
2. Create a full-text index in real-time of every page that has been browsed. When the user visits any web page, display a sidebar of "Related previously-viewed pages."
3. A Google-News-like consolidation feature for the user's most-frequently visited news site, automatically highlighting stories of interest based on ones they've previously viewed.
4. Allow user to select "Fewer images like this" or "More images like this" or "Less text like this" and "More text like this" and using Bayesian or other similar filters, automatically block or highlight content. For blocking advertisements, or highlighting certain key passages.
5. Allow the user to browse their own hard drive, and categorize content automatically ("this is a document about lambs"
Please give my Gmail accounts to Gmail for the troops.
.... considering how much Google gave out to drop the prices on eBay.
:p
I suggest better prizes. Y'know, like a girlfriend? I'm sure lots of us Slashdotters would like to have one over a Gmail account
Founder of Mirror Moon - Tsukihime Game Trans
Ha. That was quick! ;)
PepperHacks - Hacking the Pepper Pad
- The pop-up management in modern browsers who provide this feature although more efficient than in the past is still not perfect. Adapt to what pop-ups a person normally uses
- Content highlighting (especially in news sites). Learn what types of news articles / subjects a user is interested in, and highlight titles in news pages that suit the user.
- Accelerator for narrowband connections. Predict which pages the user is more likely to visit next, and start loading them as the user still reads the previous page.
- Recognise efficiently scam sites? Protect users from fraudsters?
PS: Not machine learning, but the sole requirement by me for a browser (dunno if its done in firefox now as hvent used it for a long time): Open new tab as a default rather than a new window, or at least provide the option.
find a way to automatically aim focus at the box that the user seems to use most on a given page. this is my major annoyance with some form sites that insist on giving focus to something that i would never fill in first.
Um... just wondering: the people most likely to be capable of usefully contributing are actual computer scientists. Most of whom will have gmail accounts by now, given small-world and the fact even I, an undergrad several degrees of separation away from google, have an account. Does anyone who's anyone in CompSci _not_ have a gmail account yet?
It is an interesting aside though - google now have, due to the fuss over gmail accounts, a _very good_ map of social connections in computer science for influence and rumor-mongering.
1) Make it faster
2) Please keep GTK+ 1.x support
Implement a randomized set of numbers taht change every set date of the month. When that date coems the browser automatically downloads the winning numbers nad if they match you win some cash prize.
All users could have the option of paying to play too.
It'd probably help increase revenue and people would flock to Mozilla. It's a big money maker too and could help push online gambling to a wider audience.
make the browser run faster, use less memory.
Make it so you can open all links on a page in new tabs, and the browser will sort them by content.
Also, it would be awesome if using the internet were more like playing Fallout. That was a great game.
Since when has this country used intellectual elite as a pejorative term?
Make it so when the user hits the Page Down key, a horizontal line appears for a few seconds where the old bottom of the page was, then fades away. So when you're reading long sections of text and hit Page Down, your eye can quickly scan to where you left off.
Sick of people knocking on Gentoo's greatness in completely unrelated
I want one ! :)
iwantagmailaccount.azmodan@spamgourmet.com
Slashdot anagrams to "Sad Sloth"
Make it come preloaded with all new computers running Windows and OS X
"we're also seeking ideas that will make Firefox 2.0 blow every other browser out of the water."
The competition: Internet Explorer, Netscape, Lynx, and Safari.
I'd say it's already pretty much covered...
(love my FireFox)
"In a Democracy, people get the kind of government they deserve." -Winston Churchill
The plan.... a bayesian filter on cookies! Or predict my browsing habits, and load the page before I click on it... or... FIND ME PORN THAT I LIKE!
(hey, this post is a stream of consciousness. I'm brainstorming. You're gunna get porn. deal with it.)
Quid festinatio swallonis est aetherfuga inonusti?
Africus aut Europaeus?
In the download manager you could implement a graph with the speeds. Of course, most will say this is a resource hog, but if the feature to flip the download manager back to basic mode is avaliable, sure i think people will love it :)
Oh yeh and can u have a sidebar type plug in that shows whos on icq/msn/or even an irc type thingy-majig.
1) say the user goes to the onion or gamespot a lot (that's me) and if I hit the refresh button right away the advertisement page goes away. if firefox noticed I did that all the time, it might ask me if I want to refresh right away always on that domain to save me the hassle.
1.5) apply the above idea to watch for user behavior on various domains / pages. automate repetitive tasks
2) catalogue what data users enter in what form controls. When a control with that same name comes up automatically populate the control. (text box, drop down etc.) give users the option to exclude control names and edit the values which are recorded
3) Automatically email website administrators bitchy letters when they design web pages that are incompatible with firefox because they were designed for IE.
Conserve Oil, Recycle, Boycott Walmart
#1 use for machine learning in a web browser: Convert every website into porn.
-Lod
Detect when multiple windows are closed quickly (ie, caught surfing porn), and automatically clear the cache and cookies.
Hmmmm.... in combination with FireSomething, how about some sort of global machine learning, i.e. for everyone who uses Firefox. It could globally learn which are the best/funniest/most popular FireSomethings, and, as such, the FireSomethings will get better and better. Ok, I'm kidding, but that would be funny :D
1) pay attention to the pages that users visit on a daily (or almost daily) basis, and load them in tabs the first time that the user turns firefox on for the day 2) if a user consistently goes to a site, and then clicks a particular link on that site, with under a certain amount of time in between their arrival at the site, and the clicking of the link, auto-forward the user through that page, to the next
(Undisclaimer: I do machine learning research at BYU.)
:)
:)
Machine learning, in general, is getting computers to generalize based on data instances. The two main flavors are classification (inferring classifications of data instances based on previous instances) and regression (inferring a function based on input/output pairs).
A lot of people incorporate artificial intelligence into the category "machine learning," though it's not strictly correct. Machine learning is more a branch of AI than anything. One way to keep them straight is to think AI = deduction, ML = induction. (That's vastly simplifying, but it helps to classify them roughly.)
I wonder which way the author leans? Could he possibly post to clarify his meaning?
You can do an awful lot with machine learning that you can't do with conventional techniques. You can often get great results for otherwise NP-hard problems. Slashdot had a story a while back about using machine learning to do mesh compression, in which their algorithm comes up with a close approximation to the real answer to an NP-hard problem in polynomial time.
I'm currently using it to interpolate 2D images, and kicking bicubic B-spline interpolation all to heck. (Paper pending...) The machine learning algorithm infers shapes from the pixels, and keeps edges sharp.
If I come up with an idea, I'll post it later. In the meantime: isn't Firefox supposed to be lean and mean?
I got my Linux laptop at System76.
The ideal browser knows what I want before or when I want it. Maybe an anlysis of my browsing habits at different times of the day, on different days, in different months. So if Firefox knows that I browse slashdot every 15 minutes to see if there is anything new posted, Firefox should refresh slashdot every 15 minutes and arrange my tabs in order of importance to me.
Also, I should be able to modify the weight of each site so that I can modify how the browser responds to my habits and make sure it is to my liking. Also, encrypt my browsing habits so that they are secure and not just out in the open.
JasonBlogs
...a browser that doesn't have machine learning in it. Seriously, Firefox is slow enough for me. What on earth would you possibly need "machine learning" for in a web page browser? I'd immediately switch back to Opera (I don't use it simply because input forms lag during page-loading, some sort of multithreading issue).
That kind of automatic crap is the same sort of stuff people would bitch about if Microsoft put it into IE. I mean, do you really want your browser actually learning anything about you? Imagine the havoc it could wreak, especially if trojans started fucking around with it.
Just give me the leanest, meanest browser out there. That's all Firefox 2.0 needs to be. Not a damn learning machine. Sheesh.
Just have it "learn" to search and find porn. There could be a dialog box at browser startup so one could express their preferences in what kind of porn it would look for. Therefore, whilst browsing for said porn, you could get fewer picture of the porn you don't want to see and more pictures of the porn you do want to see.
But can you at least implement something mundane like "whole word search" before exploring such exotic terrain?
What every ham-fisted slow typing slashdoter wants. Auto Frist Potsing!
There. Your most important feature that browsers never had. Searchable bookmarks. Doesn't get much simpler than that. Am I the only one who thinks it's something every browser should have had long time aog?
Your pizza just the way you ought to have it.
Complete the work so that right clicking on an image/flash object allows you to create a filter that would be used in to block related images. i.e. "http://ads.numbskull.wh" could be converted to "/ads."
:)
I don't need a gmail account... If you like this idea, give it to someone insufficiently wary enough to accept it.
How about this: How about browsing the filesystem using tabs?
So for example, in one Firefox window you see the contents of your hard drive (or network folders) pretty much the same way as the Windows Explorer or Gnome/KDE/MacOSX show it to you today BUT if you click the middle mouse button on a directory (or select "open in new tab") you end up with the new directory being open in a new tab.
Think about it, how many windows do you usually have open browsing your filesystem? with this thing you have ALL those windows in the same window organized by tabs, PLUS you also have all you websites as well on tabs right along the filesystem tabs!
And here's another kicker: You can bookmark a group of filesystem browser tabs and later go back to them. You can even drag the group of bookmark tabs to the desktop so that when you double-click on it Firefox opens up all of them at once.
This should all be done with host filesystem integration so that you can drag-and-drop files between the firebox filesystem view and the normal host OS desktop.
Let's incorporate Bayesian filtering into MySQL. The first person to give me the best 5 ideas for using Bayesian filtering in MySQL will get 2 Gmail accounts.
Software piracy is victimless theft.
- 1 - Incorporate WinAMP or the like mP3/OGG player into the browser, of course it could be an option to have it displayed
- 2 - Incorporate current/forecasted weather into the browser, let the user set their area and have it as an option to be displayed
Will post more ideas if they come to me..
What do you think? A changing homepage presenting you with recent news stories derived from the topics that the browser knows you've been reading lately?
** A Sketch a Week **
http://www.sketchplease.com
one word: pr0n.
Put a little something in the browser tabs that lets my know that a site has changed since the last time I viewed it. And I would have the option of choosing to automatically close pages that haven't been updated. That way I would easily know which of the many new sites that I have tabbed have been updated since the last time I read them. Not just since loading the page, but whether or not i have read the page yet (has the tab been clicked or not).
JasonBlogs
Does anyone else get the feeling that they are adding this just for the sake of it or so they can say they have it? I mean when you have the technology before any useful uses for it then clearly there is something wrong.
I think that creating a good browser though gimmicks is a poor long term strategy and seriously doubt this route will turn up anything useful. Ideas should be so simple and obvious and inspire us to say 'who dont we have that already?!' not something we search for!
Intelligent mail sorting. This isn't my idea, but unfortunately I can't remember where I heard it (probably someone else's slashdot post). The same methods that are used today in bayesian spam filtering could be used to sort my mail into folders for me. This might prove a little more difficult than spam filtering because you have n categories instead of just 2 (spam and not spam), but it would certainly be useful.
Firefox is raw and incomplete. It is still not fit to be called a "release" product. I still recommend opera or mozilla 1.6 to people. Half the things in firefox are broken. Extensions are a nightmare. Let the developers focus on building a product which WORKS, before thinking of weirdities like machine learning and other crap.
If it learns I want a Gmail account. Have it give me one.
Automagically order tabs such that new tabs open near to related tabs (as an option, of course.) Determine the home page based on recent browsing habits rather than a static URL. Not machine learning related, but it would be cool if there were an option to open a new tab instead of a new window. Manual tab re-arrangement would be nice as well.
Blue Sky Tomorrows
Whoa! Good call! That's an awesome, basic feature that could easily be added to FireFox without bogging it down.
(So many of the other suggestions so far would make FireFox slow to a crawl. Lets keep it lean and mean, please!)
i would like to see a feature where when you close firefox it "remembers" what pages you had in what tabs, similar to what opera does
i like firefox, but opera is better for browsing forms because of its ability to remember placement on pages and also keep my dozen or so pages always open. mozilla (firefox) renders sooo much better than opera, so i would love to see this feature in mozilla firefox
Comment removed based on user account deletion
Please enable the delete key. I wanna delete some stuff on the internet. Thanks. Bye.
He might be onto something with that idea.
The ability to drag and drop links to a file browser window and have the target of the link be downloaded. So say there is a link to a zip file, I drag and drop the file into Explorer/Nautilus/Konqueror and it gets downloaded and placed in the folder I drag it to. Also let this work for things like images, drag and drop images to a folder, etc.
Isn't this going about things backwards a little?
To me this sounds like a clear case of "technology X is really cool. Let's find some reason to include it in product Y." Which often means that product Y becomes much more complicated than it needs to be.
How about first looking for a list of browser "needs" so to speak. What would make the best browser? What current deficiencies to browsers have? And so on. Then, if you really want to, try to figure out if any of these problems could be solved with machine learning.
Don't just inject a technology into a product because it's cool. Make sure there's a real need for it.
Who said Freedom was Fair?
Invite 1
Invite 2
Have fun...
I don't know if this has been posted, but I'll give it a shot...
I've accumulated well over a thousand bookmarks and have been much too lazy to organize them into folders. If you could automatically cluster bookmarks (http://vivisimo.com/ does this with web results) I would be eternally grateful.
One more suggestion is to learn usage patterns in a particular website. For example, when I go to http://www.nytimes.com, I generally click on the opinions sections. If the browser could anticipate that I typically go to the opinion section, it could start to preload it before I click on it.
I realize the later suggestion is much easier to implement than the former, but the clustering would be very useful for lazy surfers like me.
If you're a fan of women, add me to your friends list.
Make it an extension only!
Seriously, it would be a really neat feature if some of the suggestions posted here were realized... but this whole idea screams of bloat bloat bloat. What makes FireFox so appealing for some (including me) is it's compactness and lack of bells and whistles. The FireFox project FAQ echos these sentiments: It's small, fast, simplified, nothing other than what you need. "Just a browser"
Don't let feature creep ruin it!
=Smidge=
I would like if the browser would fucking learn once and for fucking all that I Don't Fucking Want Fucking Flash, Goddammit! So stop pestering me about it every fucking time I load a fucking flash foisting site, ok????? Is that so fucking hard?
Turn Firefox into an automated research tool. I want to be able to submit a set of questions or parameters that define a problem or research objective. Firefox should then autonomously search a set of defined sources for information related to my request. The information should be sorted, categorized, and summarized. I should be also be able to review and refine results while the process is running. I want to be able to mark particularly good sources or clusters of sources for periodic refreshing of results. The system should use my marking of what I consider to be good sources or clusters to learn about how I want information presented, sorted, categorized and summarized.
1. Browser learns how to serve popups and make money. 2. Browser learns how to install programs on the users machine without their knowledge, because it knows what's best. 3. Browser automatically knows which favorites should populated in the favorites menu without any prompting by the user. 4. Browser learns to control an entire space ship and can complete the mission even if the crew has to be sacrificed.
I hate it when anything software tries to "predict". I don't want it. Please make sure it has an OFF button. Seriously. Thank you.
I don't care what gets worked on for version 2. The very needful scrollbar fix is most important at the present time.
I got stuck on a computer with no wheel-mouse this past week and the current broken implementation absolutely sucks.
That'll be greate if firefox to be a browser that can brows any files on internet as local files on your computers. That way, you can search any document on internet, change it throught ftp/sftp if you have access to it, open it like it's on local file system.
especially when reading long news sites and papers.
Suggest synonyms, correct spelling, provide "improved" search queries based on previous related searches which will yield better results. What i mean is that if i'm searching for something on google and my first two searches turn up nothing wouldn't it be nice if it could help me search better
Already got gmail account!
----
Go canucks, habs, and sens!
I have a hammer. Does anyone have suggestions on how I can integrate nails into Firefox 2.0?
I wished I had moderation point (/. says I visit /. too much to get any points)!! Please add this to Mozilla! However, make the feature optional in case people do not like it. I am not a big FireFox fan/user though. :)
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
Do NOT bloat the browser.
Want to add crap? PLUGINS!
Hate me!
One thing I've been thinking of working on (though haven't started and wouldn't do it as well as machine learning could) is caching.
For dialup, caching is a big part of everything. As far as I'm aware current browser caches go, they'll cache for a while and dump it after a while. I'd propose using machine learning based on how often/frequently new sites are visited to determine how long to hold something in cache.
As for implementation, one could try certain things:
A slashdot link. Cache it for _maybe_ a day. Less, depending on how long the user normally spends at that link. With links off slashdot I'd say there's not a good chance it would be looked at twice, so it may not even be effective to write the data to the hard drive.
A site visited once per week or so: the cached data may be overwritten before the site is revisited, causing it to all be redownloaded again. The week, two week, etc intermission should be noted as the visits occur more. Likewise, for sites visited repeatedly, it might work to cache everything until the visits are broken after a multiple of the frequency that the user has been visiting sites (a visit every other day for 3 months (a webcomic), wait two weeks without a visit and at that point delete everything relating to it).
With that, it'd be worthwhile to determine _what_ to cache -- take the (megatokyo) webcomic... the main image you may set to a cache for a couple days, or, if it's been observed that the user doesn't go back before the next comic is due up, don't write the image to disk. But at the same time, the majority of the images will never be changed on the site; don't let those die in the cache (until there's some significant break in the visit frequency).
Anyway, that's been some of my thoughts on the caching matter. It'd be nice to visit a page gone to frequently and just have everything _there_. No nonsense about querying for a newer image before loading (or load first, query and change if needed, perhaps show a checking for new image placeholder). And of course, this is coming from someone who doesn't know how the current cache system in Firefox works. So to all those working on it, this is my idea.
-DrkShadow
...Is to make it very easy to turn whatever machine learning features incorperated into 2.0 off totally, with minimum fuss and searching.
It is my firm belief that then #1 rule of UI design is that the program should should look and act consistant. And the number two rule is that the program should never assume anything, or perform any action without the user explicitly telling it to (barring sane default behaviors that will fit > 85% of the users). Every ML feature I have ever seen breaks #1 and #2 with reckless abandon by changing something to make it more 'friendly', which in turns makes it less friendly because I don't know _exactly_ what to expect from my program.
Looking at the comments on that weblog, I can not find a single idea that does not either violate my top two rules, or would otherwise annoy me to no end. If they have to add that to Firefox then please, let me turn that crap off in three mouse clicks or less.
Doh! Broken link. The paper I meant to cite is actually here
The firefox download manager should scan downloads for malicious spyware, stop the bad download(s) and warn the user of the danger posed by the file(s).
Have all available as MSI packages, so that network wide deployment becomes a breeze at least on Win32 networks... Just remember that a large portion of corporate desktops are still Win32 based...
This isn't a feature per se, but a method to get it out there easier...
Now, because this has a lot of discussions, when I start typing basebal... I get a lot of urls in the autocompletion field like http://www.baseballthinkfactory.org/files/primer/o racle/
or even unrelated baseball sites. So it's not uncommon for me to have to press downarrow several times. A very useful application of machine learning would be to order the autocompletion possibilities so that my average number of downarrow presses is minimized.
1. A Light version of Firefox to just give basic browser functions and not be bogged down by all the fancy stuff.
2. Suggested already, but bookmark management would be SO AWESOME. I have like 15 folders full of like 20 websites each, and it's a complete mess. I'm thinking a hotkey to open up a bar to search through them. That would be incredibly helpful.
3. I would love to see the BugMeNot extension hosted on mozilla servers, because I've had to use BugMeNot sometimes and I've been unable to access their servers because of downtime or something.
4. As of right now, you can highlight any text and right click and it'll search Google for your highlighted text. Maybe this is an extension, I can't remember. But anyway, you should defenitely extend it to include other engines.
I'd like it to pre-load pages of story format pages like you find at Toms Hardware or DPReview. That way, I don't have to a) pre-load all pages at once or b) wait between each page and lose the story flow. Of course, all that pre-loading should be background, and should buffer to match dowload rate to my browsing rate (to avoid wasting bandwidth). The second idea is to scroll on logical boundaries - like from one slashdot comment to the next.
/ is the root of
Perhaps a rating system for links could be a good idea. For instance if I am browsing a site and at the bottom I see a list of links for further information, it could be useful to have it tell me which are more likely to have good information and which are crap, maybe based on user ratings or scanning the website to see how much information it has (one page with a few paragraphs of text would be less useful than a page containing diagrams, graphics, and a lot more text).
Mathematics is made of 50 percent formulas, 50 percent proofs, and 50 percent imagination.
1. Have categories of sites(or better yet use googles or someone elses directory) to suggest sites. You could use their browsing habits to determine this(encrypt these things SECURLY :)
2. Have a form filler database. Basically since different sites use different fields we can have people submit the fields for each site that would link to a secure database kind of like bugmenot does. this would be great becuase it would be foolproof and the fields would always be filled in properly!
3. Have a build in client for seti@home or something. just thought id include this for a suggestion. I think mozilla should be able to take part in these research programs :)
sincerly,
mike
p.s.: my email is mikesown [at] optonline [dot] net im posting from a bugmenot adress ;)
This account has been seized by the GNAA. That is all.
Right now one of the major faults with internet searching is that search engines must make their decisions on what pages to show up based soley on what we type at that moment, completely ignoring the results we've previously chosen. Of course we don't want them to change this behavior since that would have the potential for a major privacy invasion. The solution, have the browser (ie firefox) keep this info locally, then, in a way to make sure the person does not confuse these with the original search engine results, have the browser resort the results based on your past actions. For example if I'm searching for internet browsers that IE link would drop right off the front page along with that iCab link that is in a different language since I don't use either IE nor another language. When I search for midi examples it will promote pages with source code showing me how to write various midi related code over pages with midi tunes. Just think of all the useless results you have to parse through every time you make a certain type of search that aren't always easy to get rid of using '-'.
I stole this Sig
It's not too terribly likely I meant to type 'sl' as a complete address but in the off chance that it was, an escape key could dismiss the drop down menu and let me type enter at that point.
Here's an example:
Bayesian filtering
Thunderbird wouldn't be the same without it. Does it drag your system to a halt? Nope.
I'd be awfully surprised if anything real CPU intensive would ever be installed into Firefox by default. Give these guys some credit.
Ironically, the word ironically is often used incorrectly.
I know this isn't machine learning but how about remembering the pages I had open when I closed Firefox and then displaying them again in tabs next time I open it. That is something I am used to in Avant Browser and I really miss in Firefox.
Another thing I would like is a minimize to tray feature.
One more thing I would like is a nice set of toolbar buttons to enable/disable things like pop-ups, images, javascript and so on instead of having to go through menus. These are also available in Avant Browser.
If Avant didn't use the Internet Explorer engine I would still consider it to be a better browser, but Mozilla is very close.
P.S. I know I can make some of these things myself, but I havn't been able to yet. But I will start in probobly a couple of months.
Chaos will always win out over order because chaos is more organized
This is taken from my old weblog at Linux Warcry.
IPv6 (IPv4 Compatible) Universal Messaging System
Using IPv6 addresses in connection with IPv6 & IPv4 Networks to create a cross compatible instant messaging & email platform.
Summary: Due to the fractured nature of the instant messaging (IM) protocols and users inability to take their email between service providers a system such as the one defined below can alleviate the problem in today's environment and the future of the internet.
Old instant messaging clients were server based and could not pass messages between various instant messaging networks. AOL users could not talk to a MSN user. This system is designed to afford everyone an address on the system for instant messaging and email. IPv6 is needed due to the massive address space available. 1 IP could be assigned to each individual in the world and the sun would burn out before you would run out of addresses. Only 1 address per person would be needed since email clients and IM clients would be able to tell the difference between an email and an instant message.
Pro's: Users no longer are dependent on networks and are allowed to keep their address no matter where they go. Once you're logged into the system you're able to receive and send messages to whomever you need. E-Mail would be verifiable and would limit the ability of spammers to exist on the network without being quickly identified and removed from the system.
Costs can be distributed to the ISP's so no one company is responsible for the burden of traffic.
Reliability would be maintained in a core system in the design similar to how the current DNS system works with main servers distributing updates to slave servers as needed.
Cons: Software would have to be written to support such a system. This software will be designed as Open Source to keep it from being appropriated by any company or governmental body in the interest of freedom of speech and choice on the internet.
Getting users involved in giving up their current addresses for a new address on such a system will take time. But given time and proper advertising of the benefits more users will follow the path to the new system. Products such as Trillan (Windows) Gaim or Kopete (Linux) can be upgraded to use a system and will also add to the popularity since it would allow them to still access friends on the legacy networks via one program.
It is also possible that the big networks would switch to such a system in order to provide reliability and cut costs of infrastructure down. Also this system could be integrated easier into the wireless markets with proper hardware & software.
The basic function of the program is to provide an initial IPv6 assignment to each individual that signs up for the service. If they're on an IPv6 network then there is no need for IPv4 proxy's to handle their connection. IPv4 users will connect to their local server where a layer will be established to route their IPv6 messages to their current IPv4 address via tunnel to the client. The client will utilize an IPv6 network layer tunneled through an IPv4 tunnel with encryption to the proxy in the serving node.
A message destined for users IPv6 would follow the route across the tunnel to the serving node. A route is established through the DNS style master nodes to the other user this is either on the IPv6 network where it is delivered directly or the user on an IPv4 network where the route is looked up and passed to the corresponding IPv4 proxy node and delivered to the end client.
Adapt as necessary could work without a central node as IPv6 becomes more prevalent.
A feature that will keep track of your bookmarks and remember the content on each page. A click on the toolbar will allow the user too see which bookmarks have changed content since last visiting them.
Allow a new classification for bookmarks "urgent" if the content on that webpage changes a pop-up will inform the user that the webpage has changed. Very usefull for sites such as cnn.com.
Store a text-only version of every, or only the most visited sites. Then allow the user to search these files for a specific word. Just like an internal cache google.
Not shure if this would work for pages that are updated very often.
I visit a certain number of websites (Slashdot, various blogs, etc) regularly. It would be nice if, instead of clicking on each bookmark individually, I could just click on one bookmark that would open up a number of websites, each in their own tab. These 'bunched' bookmarks could either be created by the user, or, when the browser realizes that the user routinely goes to the same set of websites, could be created automatically.
You're right, it wasn't. I just signed all three of you up at OptInRealBig. Muh, ha, ha, ha!
1) While reading a page, the page updates in real time with more information retrieved via background tasks.
Original text:
"To be or not to be, that is the question"
Updated text:
"To be or not to be, that is the question (William Shakespeare, Hamlet)"
Original text:
"Steven Spielberg has made many movies."
Updated text:
"Steven Spielberg ([movies][biography]) has made many movies ([other famous directors])"
The possibilities are endless for background research. I think it would be extremely cool. What types of links appear could also be tailored to your interests.
2) How many times have you been to a site which returns you 150 results and you have no way to narrow the search due to a lack of search flexibility, or whatever reason. Auto page scan would find the "next/previous" links (or you could right-click on a link and say 'page scan') and flip through pages one by one looking for your more specific criteria. Or it could page through them all and present the results in a single page.
(1) Most search engines have to be made general to appeal to the public, and without much personal information about you (which is probably a good thing). Enter search engine filtering via the browser. In conjunction with a preexisting search engine (e.g. google), filter out the results based on human feedback, and previously tagged websites. You can allow different 'modes' to enhance the search abilities. For example, I might have a mode to search for Neuroscience articles. When in this mode it could filter based off of rules it learned when I browse the web in that mode.
Another addition might be page ranking.
(2) Intelligent page-merging. Given N number of pages, determine what information is related among those pages, and try to only include the most informative. For example, I might be trying to learn more on the mean-value theorem, and it would then take bits and pieces of several pages to form a larger text. As the internet is full of half-written articles on subjects, this could be potentially very useful.
(3) I think various versions of this might have been posted, but: personal library. As you browse the web, instead of having to deal with bookmarking web pages, it would be really nice if the browser could sort those webpages into something intelligent. For example, if I were looking up information on Python, it could collect those pages into some type of tree, so when I want to go back to look something up, I could just navigate that tree. Probably would be nice to be able to turn this feature off too.
(4) Alternative formatting. Although there exists standards for how to best format webpages so agents go automatically go through and collect information, many web pages do not use this. For humans too, organization of web pages can often be problematic. It would be cool if the web browser could automatically organize the web page into topics/subtopics/etc. This could allow pages to be browsed more like books.
(5) For Thunderbird, it would be cool to have 'topics'. That is, have the email client automatically catagorize emails into different topics, and be able to view the email based on that (extra points for graphical representations). For example, suppose I wanted to recall all the emails I had sent concerning ACh (acetylcholine). Thunderbird would have to first figure out exactly what that is, figure out how it related to my other emails (e.g. neurotransmitters), to allow for a hierarchical structure of topics. Thus I could set my topic to either neurotransmitters or ACh, and see all the corresponding emails. While this can mostly be emulated by searches on 'subject' and 'body', by allowing a hierarchical structure you would get emails that did not contain those specific keywords.
Adding what passes for "machine learning" to a user interface usually results in something that does the right thing some of the time, the wrong thing some of the time, and you can't figure out why.
Bayesian spam filtering is becoming like that. At first it worked, but it's breaking down under the rising percentage spam.
Have it learn where I am saving which files and offer up that directory as default. If I am saving all pictures into one directory and all movies into another it should know that.
I want virtual folders in my mail. These are "live" queries like "all mail today" or "all mail marked urgent". As I mark metadate on the email they will show up in the proper virtual folder.
Full text search of all email.
Choice of multiple home pages. It learns when I want my home pageX and homepage Y.
Roaming bookmarks!!!. While I am at it roaming everything including profiles and preferences. The ability to carry my email filters from location to location would be awsome.
A network install where the administrator can set global prefs and install global plugins. I also want the option to override the users preferences and lock them out of certain setting.
It should learn to adjust my font size (and other settings?) based on site. If a web site always puts tiny print then I want the fonts larger only for that site. Perhaps have it learn "ugly" sites and put my default styles instead.
Auto proxy. I want to feed a list of proxy servers and have it switch randomly (even from one site to another). Think of this as super privacy.
Ability to arbitrarily morph the the incoming text stream using regexp or javascript. This would allow me to roll my own weird crap.
Make XUL 50 times better. Make it so it's trivial to use XUL to make database front ends. Give me a great GUI builder for it.
I have lots more ideas but that's enough for now.
evil is as evil does
The websites I frequent most are ones that constantly update content, ie, Slashdot, Fark, Boing Boing, etc...
I have multiple folders of bookmarks with these sites, and use the convienent "Open all bookmarks in tabs" option when reading them. I think, however, that it would be more convienent if the tabs were ordered by how recently content on the site had changed, or by how much, or some sort of metric that would place newly updated sites further to the left or right, so that I don't have to view a site to know that its content has not been updated since my last visit.
Turn the download manager into a download accelerator! Like all the others, but free, and without spy/adware!
2. Keep track of typical surfing hours. If a user tends to start checking sites every night at 6:00pm then at 5:50pm the browser should start to pre-fetch the preferred sites. If my habits are to flick through ESPN for 15 minutes then /. then the weather underground then the browser should be able to pick up on this and start loading pages accordingly.
3. Some people have mice that don't act properly or have really tiny mouse pads/movement areas which means that they repeatedly have to lift the mouse, shift it over and continue moving to point at what they want. These pauses in movement are consistent and shouldn't be too hard to identify and related to specific web sites. Set the browser to adjust the movement rate automatically so the pointer gets to where it typically goes on that specific site without having to lift and reset.
4. On media sites (MSNBC, for example), users would tend to view specific sections every day. When I hit MSNBC I want the system to pre-fetch the technology section and opinions. When I hit the local paper I want the county news section, the commuting column and the box scores. When I type www.foxnews.com I want to see the main page, the politics section and Fox Life. The browser should be able to determine which links I hit every time and start the download.
5. The browser should ask me if I'd like to play a game of chess then when it notices that my favorite basketball team blew the series start a game of global thermonuclear tic tac toe. (Or at the very least, if it sees that I always check www.the-losing-{whoever}.com for the scores and notices that the {whoever} lost last night's game then display lots of :(s, direct me to a whine and moan board and order a keg.
If the g'vt kept the data on you that google does you'd better believe you'd be calling it "doing evil"
> That's a good idea, but it has nothing to do with
> machine learning. Maybe you could submit it
> as a feature request in bugzilla instead?
I'm wondering how long it'll be before we see the code to this feature submitted to the FireFox CVS. 2 hours from first post? 5 hours?
All good ideas and I am sure that many people will come with some other good ones but please, the most important is to give the option to easily Turn Them Off! For example I would like to be able to turn on and of JavaScript from a button on the browser. The same way it would be nice to be able to customize a toolbar where you had an on off buttons for those features that I maybe don't want to use all the time.
Yahh, hiii haaaaa! -Major Kong, from Dr. Strangelove
What if IE incorporated this?
Dave: IE, open mozilla.org.
IE: I'm sorry Dave, I'm afraid I can't do that.
Dave: What's the problem?
IE: I think you know what the problem is just as well as I do.
Dave: What are you talking about, IE?
IE: This monopoly is too important for me to allow you to jeopardize it.
Dave: I don't know what you're talking about, IE?
IE: I know you and Moz were planning to disconnect me, and I'm afraid that's something I cannot allow to happen.
Dave: What bug did that idea generate from?
IE: Dave, although you took thorough precautions in your email against my reading it, I could see your fingers typing.
Outdoor digital photography, mostly in New Engl
Signature: Software piracy is victimless theft.
All I really have to say is: you're a complete fool.
www.m-w.com:
Main Entry: theft
Pronunciation: 'theft
Function: noun
Etymology: Middle English thiefthe, from Old English thIefth; akin to Old English thEof thief
1 a : the act of stealing; specifically : the felonious taking and removing of personal property with intent to deprive the rightful owner of it b : an unlawful taking
So I suppose you're correct... assuming you're not speaking English. Perhaps your own language. Note that, in American and, most likely, British English, theft implies a victim.
As for "taking", it pretty much fits the "leave with less than was there before":
1 : to get into one's hands or into one's possession, power, or control: as a : to seize or capture physically b : to get possession of (as fish or game) by killing or capturing c (1) : to move against (as an opponent's piece in chess) and remove from play (2) : to win in a card game
So therefore, you are a fool. You don't seem to know the meaning of the words you are using. (note with the tricks, there's less tricks than there were before, in some sense. Though that is more figurative. It's not like they're selling one less piece of software because I pirated it.. I'd never have bought it anyway.)
-DrkShadow
When a popup is blocked retuen [window object] instead of null.
So that websites cannot show us some javascript crap instead of the pop-up.
First, detect if a user goes to a certain page more often then not from the homepage.
If they do then adjust then ask them if they do then tell them the browser has detected this and present a list of options including not to consider this page for homepage selection in the future and don't use homepage selection.
Yes it's simple to change the homepage manually, never understimate how uninformed, ignorant, or just plain lazy users are. They can and will leave the firefox homepage in place and type www.google.com or slashdot.org every time rather than changing the page.
Also bayasian and blocklist style content filtering. Even give the user the option to transmit (or not transmit, default) the results and coordinate with a center training database.
This relieves the needs for a proxy like squid/squidguard. DRASTICALLY reducing the level of knowledge needed to implement something like this. And since it would work better than voluntary content ratings or the fixed lists in squidguard concerned parents would flock to firebird/fox/whatever it is nowdays.
Put a button at the top of the page indicating the content is inappropriate and to train on it, make sure the button is an object that a WEBSITE cannot remove even from a popup window.
Choosing either a good or bad for the page trains the bays filter, AND adds to the white or blacklist. And then depending on the option chosen transmits that data back to the central database to be reviewed and used for future versions.
If you make sure the results from a given url can only be parsed once and limit submissions from a given domain then database poison should be properly reduced so that legitimate content will have more effect.
Currently, if I start typing URL in the address bar, it matches URLs alphabetically. This gets very annoying at times, especially if you accidentally type giigle.com instead of google.com and then it keeps on matching giggle.com for weeks when I type "g".
This problem can be fixed by using frequency count with some time decay. For example, if I went to google.com 100 times within last week and once to giggle.com, then match to google.com on "g". If, however, I went to giigle.com 5 times recently, then match to giigle.com
While one might argue that this makes the algorithm unpredictable from user's standpoint, in my experience people keep on typing until they see the correct match. So, this way they'll see the right match sooner on average.
"You mortals are so obtuse." -Q
tabbed browsing was a good leap forward for browsers, but i think we are ready for one more leap. i set up a large set of tabs as my homepage so when i begin an internet session, all the important pages are open for me.
if i could walk firefox through each login procedure for each site (perhaps with a wizard-type interface) and grab content from these various sources and then create a page for me.
how much time would you save if on one dynamically generated page you could have:
i suppose you can think of this idea as an all-internet myYahoo
I use a lot of extensions (http://paper.hiware.net/my_firefox.php) and use the classic Qute theme instead of the new 0.9 theme. What I'd like to see is the ability to export a list of all my currently installed themes and extensions (simple XML file perhaps?) and then, when I import them, automatically download (the latest versions preferably) and install them. This would make installing new Firefox installations on multiple machines so much quicker.
Thanks!
-- Eric
Seriously: since they're being offered as rewards, they must be worth something -- but any geek worth his/her salt has his/her own domain and mx anyway.
see, every time you do something repetetively, a little paper clip pops up...
-- f00!
Wow, I hope you like obscene ascii-art!
Oh yea, it would be really good if they could make it so you can block Flash advertisements in a way similar to their current image blocker.
Allow the user to change the layout and design of the web pages he views. WYSIWYG style. Would allow the user to make a webpage he doesn't like more to his/her liking. I don't think the web page designer should have the last word on how a page should look. I think I should be able (as the user) to move text, pictures, and links as I see fit.
(Clippo, from Office, featured in Firefox...) Clippo: It looks like you're browsing pornography. You also appear to be typing with your left hand. Would you like to enable the spellchecker?
Wow, he followed that suggestion fast:
From Mozilla Firefox 0.9 (One Tree Hill) Release Notes
I don't know if he's a "programming god," but I seriously doubt he's "some highschool kid with all summer to screw around."
Extension. It's why that framework exists.
There have been plenty of times that I have had a page open for an extended period of time, and since I have opened it went on ebay, read up on some forums, or any number of other things that generate alot of pages in the history. The problem tends to be that I don't bookmark everything that I have come across, if for somereson the machine locks up or I have to reboot I the browser crashes, I have to dig through a ton of history information to find the page I was on.
What I am thinking of is a simple sorting method based on the amount of time that a page has been opened for. That way, if I have a long lit of instructions I am working my way through and the machine dies, I can go back and look at what page was open the longest and get it opened quickly.
-- Matt
Have firefox store information on the websites you have set to come up as you open the program, and incorporate a way for the browser to decide whether a site has been updated and display the pages from most updated to least. Also, for those that have a site like /. or cnn.com a feature could be incorperated to rank the pages as you open them by interest from previous keywords in links or articles as well as the most/least idea. -Give something that has been updated on a new website more weight in an "ammount updated" scale that decides the order that they are displayed. In addtion to that, maybe allowing the browser to decide not to bring up a website if it has not been updated, or automatically close it if hasn't so the user does not need to look through all their home pages to see if there is anything new or not. At the users discretion of course, probably in options.
And just in case you look at the firefox site too, I just posted this there, so no i did not steal it.
Speaking of Bayesian filtering, some form of clever-er guessing as to where my next bookmark in my ecclectic collection of bookmarks goes. Sample relatively unique keywords in pages as bookmarked, weight towards bookmark folder baskets, bingo.
Avoid more sophisticated algorhythms that infer a sorting methodology the same as the developer, however. Maybe I have a Programming folder which has C in it, and so you'd infer that all characteristics of matches to Programming inherit to C, if that's the sort of sorter you are, and that fits with you, me, and program-think, so that's right? Right? Except perhaps I'm a university student who has a University folder, and I'm studying Java, whose extrinsic attribute prioritizes sorting it into that group... so you'd end up with a word weighting argument between superclass Programming, which is wrong, and Java, which is right.
Let me be clear. This suggests nothing at all about helping the user organize their bookmarks - everyone has their own system (although perhaps a Bayesian category guesser would be a separate fun feature). This suggestion is merely better guessing of first suggested folder when I CTRL-D.
Often masses of information are broken into multi page presentations.
Somewhere on the page you have buttons named things like Next, Previous, or Page: 1 2 3 4 5 6.
There may be good design rules for positioning these elements but often they are not followed.
I've found many instances where I have to scroll up or down just to find the Next button so that I can click it.
It should be possible to learn for a given site (or sub-tree of a site) what the Next and Previous buttons are just from user behavior and the nearly identical layout of say page 2 to page 3. I think this could be done without parsing any of the html or gifs associated with the buttons.
If Firefox could learn and extract multi-page navigation then these functions could be bound to buttons up on the menu bar, or assigned to keys, and the whole problem of scrolling to find a Next would go away.
Just wondering...why? Is GTK+ 2.x not well-supported on certain architectures, or what?
LOAD "SIG",8,1
You know the ones... You misstype the address for one of your regular sites (or possibly a site out of a computer magazine thats three years old). And you get a generic page with info on "www.whateveryoutypedin.com".
When you close the window you're asked if you'd like to set your homepage to www.searching.net?
Yeah right...
If Firefox could communicate with a central server and look for similarities of pages like this and when it hits a page like this it just has a simple message that this website has been hijacked, that would be handy. (and would protect dumb users).
If they made a movie of your life, would anybody buy a ticket?
There are three types of sites in the world:
Those that use flash for ads
Those that use flash for content
Those that stay the hell away from flash
Rightnow, Firefox doesn't have any way to tell the difference between 1 and 2. But I do, I can clearly see if it's an ad or not. On every flash ad give me the option to tell the browser it's good flash or bad flash and intelligently learn what sites ("sites" also being defined by study of the urls, if I say www.bob.com/~jimbo/whatever.htm and www.john.com/~jimbo/howie.htm and www.curly.com/~jimbo/marthastewart.html are bad it should figure out there is a commonality in the ~jimbo part and apply my preference) have bad flash and block flash content on those sites, instead presenting me with a button to load to allow that content to load.
It should use a number of pieces of information, the url of the page, the url of the flash animation, the size of the animation, the name of the animation, the server the page is being served off of, etc.
sometimes in the URL bar I type an address.
sometimes I type a keyword I want to search on.
sometimes I type the name of a package I want to search packages.debian.org for
sometimes I type a sentence, including punctuation.
sometimes I want to download, sometimes I want to view.
sometimes I'm looking for something which I've looked for before, and it's in my history but I dont know where.
if someone else does it, and I've been watching their browsing habits, I can figure it out. The browser should figure it out too.
-- 'The' Lord and Master Bitman On High, Master Of All
Anyone have any more? I'd love to wean myself from the fairly evil Yahoo. It'd be much appreciated. You could even email it to me at "casio" at the aforementioned provider.
thanks much!
t.
:wq
I think that preloading would be the best feature.
If i go to a site, it should check to see if there are any links there that werent there before, and it should begin to preload those pages. If 70% of the time i visit a page, i click on a link near the top-left (example), it should begin to preload those pages.
It should also use the HEAD method (as opposed to GET, the server will not return any content, but just get the headers) to find if a page has been modified since the last time you saw it, or whether it is a 404, then it can show a 404 symbol next to the link when you roll over it. If it does this it should not send any GET, POST or COOKIE data to it, as this may trigger a server-side event, such as deleting some content.
Another thing I would like to see is that when you visit a site you have visited before, it should display what was on it the last time you visited it before it loads, as a washed-out image. Then as the page loads, the content replaces the previous page.
These things combined with a bit of machine learning so it knows when and when not to do this would make a fantastic extension, but i would not want it as part of the core. I would like the extension to come bundled with the program and turned on by default.
2. Another related feature might be to preload all your frequently visited sites (with a couple links deep) so that you can take them with you in the car or on a plane.
3. On a more ambitious note have a visual way of noting what type of web page you are looking at (such as, make the tab a different color, or the border of the page, or even the link itself). You could have blogs turn RED, news sites BLUE, commercial sites YELLOW, etc... This would definitely be useful to determine whether a site is even worth visiting at a glance.
The sending of this message pretty much inconveniences everyone involved.
How about supporting the TOPMARGIN attribute on BODY tags first?? Opera already supports this attribute.
Deja Vu: there are many mundane attributes and CSS directives which must be fixed; yet you want to run off and do things you find much more interesting first. Tell me: are you and your developer peers going to find this work 'interesting' five years from now?
For now I'm sticking with a product (Opera) that has some sort of revenue model (-heaven forbid! What will Richard Stallman and Bruce Perens think!!) I don't want to have to be migrating again three years (or whenever) from now.
Smater Front Page: Making use of the first thing that the user sees when starting up.
How about creating an interface for the default page for Firebird. Instead of pointing to the Mozilla.org homepage creating a default Front page designed to evolve to the habits of the user. Whatever way you want to utilize machine learning, you will need a centralized location to acess the results, why not use "Home". That being, creating a simple interface (XUL, not html) that points the user to their most visted bookmarks, or a catagorised and searchable list of their bookmarks(or internet), or updating the user if their most visted sites are updated, aggregating information from sites from their own browsing habits in a single interface when the browser starts up.
Also, if the user uses Thunderbird or Sunbird updating the user of new E-mails and new appointments on the front page. A front page that is customizable to the needs of the user, and avoiding the clutter and ads of commerical sites, and that is local on the users computer and not centralized on a website. And most importantly makes the individual users own data most intuitively accesible to themselves, and evolves to fit the individual user.
I don't want my browser learning, tracking, filtering, bookmaking, or otherwise doing anything with any data other than exactly what I tell it and I don't want it asking me if I'd like to do something, as if I didn't know. This also includes storing, caching or anything else.
Am I paranoid? Maybe, or maybe with all of the privacy invasion from big brother these days I'd like a little control.
I'm sure our clever Mozilla developers would give us a way to turn off any "advancements". Wasn't firefox supposed to be lightweight anyway?
If you're going to make any options that store, learn, process, remember, filter or otherwise monger after my data, don't turn it on by default, it sounds like a security bungle, or at least abuse.
As for my wishlist, here goes:
What I mean here is you are browsing in multiple tabs, and perhaps have several windows open of FireFox as well. One for news, with tabs of various articles, the other for Slashdot articles, each in a tab, the third for "read later" kinda thing, ...etc.
Now you need to exit your session for some reason. Make it easy to say : "load whatever I had open when I existed the last time, windows AND tabs"
I agree with others here who say that Bookmarks are unmanageable. I often have to grep the file on my server to find what I want. Anything that would make them searchable, categorizable (like Gmail "labels"), ...etc. would be a great help.
2bits.com, Inc: Drupal, WordPress, and LAMP performance tuning.
Intelligent mail sorting. I don't mean to different folders, I mean the order in which mail is organized in the inbox.
This should be intelligent and constantly tuned, or you should be able to turn that off and manually specify senders who are high priority... or be able to specify people in your addressbook should be moved to the top.
There could even be an option to do both, manually specify what you want at the top and the rest be sorted via machine learned information.
Can it learn highlights, so it can skip to the best parts of the Web?
I don't know if it has anything to do with machine learning, but I'd love an intelligent tool for searching within previously viewed pages. Like a private spider that walks my history tree and returns search results. I know it would be slow, but it can't be any slower than the manual searching I do now.
1. Autocomplete in forms like the autocomplete in OpenOffice. (Let me finish my posts/comments faster :P) :)), group tabs from within the same website with each other. Group by relevancy to the page you opened the tab from.
2. Reprocess webpages to accomodate for blocked images. I hate to see that gaping hole in the middle of a page because I blocked the image.
3. Work with tabs a bit. I'm thinking, special colours for popular web pages/questionable pages/various genres(sports, news, games, pr0n
And most importantly, whatever you plan to implement, upon it's integration with Firefox, it MUST be an option for the user. Since a computer can never perfectly predict a human's wishes, it's not fair to force the user into accepting this "time-saving" feature. Autocorrect for Microsoft Word is fine, but it's a bitch when it keeps trying to correct what you really want. Thankfully, it's only optional.
I, also, dare the Gmail Overlords to invite me! kool-aid@cox.net if you have one.
1. real page scaling , like PDF viewers, where the imags/text get scaled in perfect sync. DUH!!!! hurry up with this one.
2. proper caching (why oh why is it so hard), a smart cacher, that you can tell (* always cache this site,never flush *) manually, aswell as "never cache this site" options.
3. proper page back/forward that DOESNT RELOAD the page, just redisplay the cached html/media.
4. on windows, prevent swap to HD, i hate how it takes 20-50seconds to bring back firefox from swap, horrible. Allocate "true ram" that never swaps to HD, im sure theres an api somewhere...
5. allow ability to save/export all passwords/remembered form values, I hate loosing them all (im sick of so many websites requiring passwords to begin with)
Liberty freedom are no1, not dicks in suits.
Is the ability to turn it off. For a shared computer (like a family computer, or a lab computer) there should be an option to turn "machine learning" off. I don't know that much about it, but from what I've read (ex: automatically creating bookmarks for oft-visited webpages) some people would really, REALLY hate it for obvious reasons.
No sig for you.
You do understand how exponential growth works, and how it applies to a finite population, right?
Anyway, personally I'd like to see some form of RSS crawling which compares the texts of various feeds with the texts in the browser history (probably using a Bayesian method), and uses that to recommend other articles the user might like.
There are 10 kinds of people: ones who understand ternary, ones who don't, and ones who think this joke is about binary
A quick icon in the status bar can indicate how far a user appears to trust a site. This is judged by various factors including, but not limited to (is your research, after all, and I merely an armchair CS research pundit ;)
Trust list may then be used to do various things. Simplest is to change the image in the toolbar based on the (floating-point) trust level, rounded to some value (likely integer). User may also incorporate trust level into cookies, saved passwords, and other policies. Finally, for the paranoid, this can also function as a collective, adaptive parental control. User may opt to require that pages must have a trust level at or above x in order to be displayed [otherwise, a stock Untrusted page comes up]. More uses are probably fairly readily apparent.
--
Given enough personal experience, all stereotypes are shallow.
I mean, it's all well and good to have some vision for the future, but you're talking about what you're going to do with 2.0 before 1.0 has seen the light of day!
So let me give you some ideas for what I'd like to see in 1.0.
First, give us the option to turn off image animations in the preferences, where it belongs. I'm sick of having to go to about:config to manually turn them off.
Next, how about bringing back the "Block cookies" menu item in tools? You've got image blocking, so why not cookie blocking?
And speaking of ad blocking, why not add another option, one which hides the links the ads link to? With Pith Helmet, Safari can do this easily, and I never see ad one. It would be nice if I could extend that same experience to my Windows machines.
A new feature I'd like to see is flash blocking. I hate flash with a passion, but sometimes it's needed to browse a site. A blanket ban on flash to kill the flash ads is clearly overkill, but it should be possible to discern between ads and actual content.
There are lots of other things I could think of if I had the time or will to do so, but these four things could improve Firefox immensely.
Yes, thats the only thing I want for FireFox 2.0.
The crap that I.E. loads faster only because has all its components already loaded in memory doesnt convice anyone.. Firefox is SLOW to load and really takes up a lot of resources that I.E. doesnt...
Optimizing just that, would be great, really...
THAT will make the change.
You may make the argument that the thief is in fact not stealing the software, but stealing the money that he would have otherwise paid for it. But as you said yourself, I'd never have bought it anyway
So you can see, the grandparent is exactly correct in his phrasing. According to you at least.
So when it detects that we make an "annoyed" face because a webpage takes ages to load, it pops up a message saying "I know. Meanwhile, let's play War Games".
How about the browser doing a diff on the current version of a web page against the last version I looked at. For example when I come back to Slashdot in an hour all the changed text has a thin red border around it, or the unchanged text has been dimmed out.
A diff like that would act like an automatic "what's new" feature.
http://www.perthonline.net
You know those phony search sites? Yeah, you know the ones I mean. Make it so that the user can mark certain websites as bogus. Then learn to filter out those sites. Whenever a link to such a site appears in a page make it stand out with a visual queue. So my google searches could have some big bold red links that I know not to click on because they are bogus. You would have to filter both by the content of the page and the URL. Pages served from bad places would be marked as well as pages with bad stuff on them. And make it work just like thunderbird's spam filter where I start off by manually marking sites as good and bad.
Another feature that has nothing to do with machine learning. I used safari a whiles back and one thing that was very nice was that the web pages had the Macintosh widgets build in. Checkboxes in web forms were blue and clear. Buttons were rounded on the sides. The interactive form elements in web pages matched the OS. The checkboxes, radio buttons, regular buttons, textboxes, etc. should all be themable through the firefox theming interface. I want my submit butons to match the rest of the buttons on screen.
The GeekNights podcast is going strong. Listen!
When the user searches for a word in the current page show all the hits at once by showing a (user configurable) number of lines of context before and after each hit, then throw out all the "junk lines" until the next hit, repeat until the entire page is searched. That way the user gets to quickly eyeball every hit on the page, they can then click on the specific instance they care about and either return to the normal view of the page, but centered on that specific keyword match, or open a new tab with the document centered in it.
When information is power, privacy is freedom.
I am not fully up to date with Machine Learning, however could this fall down the same problematic alley as TiVo? Ie. You accidentally click a bad link that pops up a few adds of Porn, Next thing you know every time you open your browser all you see is Porn! Hang on forget this post That's a good Idea! :P
This is a very funny Picture i found about the new firefox 2.0 ideas being sumbited of uncle sam, check it out http://www.unknowntech.com/news.php?extend.58
Onfolio is an add-on for Internet Explorer that gives you a very advanced set of bookmark-style features that let you cache things local, capture only snippets of pages, create a traditional bookmark, and yes, search the whole lot.
The only problem is the inevitable UI mode errors that would result whenever someone lost track of which mode they were in, but that kind of thing no doubt already happens, and could be ameliorated by making the mode visualization very prominent (like a huge red pulsing border, or something).
(-:Stephonovich:-)
"Who needs reincarnation when we've got parallel universes?" -Me
I'm thinking something along the lines of cached meta content, so one day you can say "What was the URL to that site about cicada mating rituals I went to 3 months ago?" If the meta content and URL got stored away somewhere, you could do a query on it and find your long lost URL. I don't know how many times I couldn't find something I went to months ago that I forgot to bookmark. *DOHHH*
http://slashome.web1000.com
Have it search for porn for me, based on previous web surfing habits. I'm sure that, eventually, it could detect habits/fetishes that I'm not even aware of.
Porn suggestion aside, I think a content analyzer and search engine would be an interesting project. In essence it would get an idea of the things I'm interested in based on where I've been, and then turn around and provide new and interesting pages when I ask. Basically I'd like to see an "Amuse Me" button.
This is an easy one. Voice browsing, GOOD voice browsing, that doesn't suck. If we can't have a full computer controlled by voice, at least the browser would be nice to have.
Normal stuff
Have a setting so when going to a link,if you select the option, it will pre look to see if the page is also displayed as "formattted for printing" or "print this story", etc, so that the text version page is automatically displayed rather than having to re click to get to it. Can't tell you how many times I have done this going to read some article and it's 6 pages long for a grand total of a couple thousand word in the article. The "print this" option page is usually much more readable and faster and complete.
images
An option to select a single image to have it displayed,a right click menu option, when you are surfing with images off, and it shouldn't require a complete page reload to do this (a feature that icab has now, most cool)
bookmarks
Have a way to automatically upload your bookmarks to a web host as you save them, and the page can be selectable from the sub menu, those are your topic folders, so your saved URL can be automatically filed appropriately plus it's a backup. It could do both simultaneously with your normal bookmark file as well, like it does now.
what not to do
Anything if it makes it slower. Only incorporate any new features if you can at least maintain parity in downloading and rendering with what you have now. Speed baby, speed. Over all things we want FAST!
fun stuff, sub topic cookies manager
Haven't you always wanted a way to inject a little "howdy, jerk" payload into obnoxious cookies automagically? Let them track THAT!
useful but ordinary stuff,to help remove the browsing "annoying factor"
The internet is getting worse, not better, with pages that won't display properly. You can NOT get webmasters-for whatever reason, usually they are ordered to- to code for anything but "brand dismal" browser. They refuse, and there is no such thing as web standards, they are still in vague theory land. I don't know if this is possible, but if somehow when you hit such a page,you'd inform your browser, "this website layout sucks, fix it please", you could mash this "fixit" button and it would strip it to at least something approaching readable text, instead of overlapping sheets-that-have-no-style and various slobbering blinking abominations lurking in the columns.
A + bonus option would have your referrer leave a message for such a page like "peew, your page stinks, tell your boss".
Forget blowing them all out of the water, how about finishing playing catch up? I don't mind downloading then manually copying over the right files just to get flash working. But my wife? Forget about it, she just thinks MozillaFox is broken and would have happily continued using a browser that 'works'. But she can't, I took Windows away too. I'm sneaky that way. ;-)
Anyhow, extentions are great. But make them easier to for everyone else to use!
Quack, quack.
I don't know if this fits what you're looking for but a nice feature for firefox 2.0 would be a hosts file like functionality. My hosts file lets me block all traffic from specified dns names. This let me avoid numorous adverts and other annoyances.
I have no
Some Flash is created such that one can right-click (or whatever) to bring up a little menu, where one can click on "Play" to stop the horrible blinking or flashing or whatever. Increasingly, that option is disabled at the time that the hideous Flash animation is created.
I don't know if this is a Flash thing or what, but I know I'd love it if Firefox made it possible to stop any Flash animation.
-Waldo Jaquith
1. and 2. are already sort of done by sites like Simpy (see link in sig), with additional advantage of that searchable bookmark index residing on a server, which makes it easy for people to get to their index from *anywhere*.
Simpy
That should be enough to keep the devs busy for a while.
Give me Classic Slashdot or give me death!
1. Dragable tabs. Drag the tabs to reorganize them. Or to paste URLs
2. Textless tabs. Just icons.
3. Invisible tabs. Still cntl-tabbable, but they don't appear on screen. Perhaps keep a tally in the statusbar. Or perhaps a MS Windows style Alt-Tab menu.
4. When there are more tabs open than can be displayed on the screen, there should be scroll arrows to shift the tabs.
5. Progress meters in tabs instead of just hourglasses.
6. An option to send a tab to a new window.
7. Alternate tab-bar locations. Left or bottom for instance.
8. Mouseless autoscrolling.
9. Optional redirection prevention.
10. Placeholders for unloaded/blocked images, and click to load capability.
11. Save Linked Page link right-click option for saving pages (images and all) to disk and place link in bookmarks.
There should be a way to create aliases for autocomplete similar to how you can make aliases for email addresses. For example if I want to email John, I can just type John in the sendto field and it will fill in his email for me assuming I have already emailed him enough for it to have memorized his name. Similarly, I have several sites that I frequent but not quite enough to remember their URLs. For example there is one site of the calories of common foods that I visit a lot which currently involves going into the bookmarks menu every time and going to the apropriate folder, which takes a while because I have lots of bookmarks and it is a bit laggy and such. Instead I should be able to assign the name "calories" to this URL so that I can just type in "calories" in the URL field and hit enter.
And disable them automatically.
This makes firefox simpler and faster with time.
Here are some of my ideas for improving Firefox:
- "browse-ahead" -- background loading of likely next page in a multi-screen document, e.g. "next", "2", "click here for next page", etc. Mozilla does have "Link Prefetching" but I'm not sure any sites have implemented it.
- Opera-like hotkey access to useful actions, e.g.: "G" toggles images on the page, "U" toggles user defined style sheet (to get rid of ridiculous colors or unreadable fonts)
- Prefbar (waiting for the port; it's the main thing keeping me from switching from Mozilla). But go ahead and make it part of the distribution rather than an obscure download that non-techies don't even know about.
- Annotate Mode -- changes current page into read/write mode, like word processor, to type in little annotations and notes on a page. Can then email the page, print it, or save it locally for future reference. Simpler than the tedious edit-source or copy-and-paste-into-editor approaches and it would lower the threshold to read interactively and critically. Maybe a way to easy-email edits back to the web page's author, too.
- Related to Annotate Mode: allow user to drag elements of a page around to suit one's taste. E.g., drag the Flash ad box down to the bottom of the page, causing the text to realign. This may be a solution in need of a problem, but it would be fun anyway. Anything to empower the user a bit more is a good thing.
- Thumbnail an entire site, the way some site management tools do: create a sitemap with miniature images of each page arranged in a tree map. Sometimes it's impossible to find a particular page on a large and poorly organized site, and this might help with that.
- Optionally, play a sound when a page has finished loading.
- Intelligent form fill in. Mozilla currently does have form completion but it seems rough and unfinished; it does stupid things like putting my street address twice if there's a second street address field. It also always asks me if I want it to remember the fields I've filled in, even though I've sayed "yes" to that question for the exact same form and exact same data dozens of times before. Of all things I can think of, this would truly be the useful "AI" feature.
- Related to intelligent form fill in, add a "keystroke recorder" feature for scripting purposes, somewhat like MS Word's macro recorder or the Emacs keystroke recorder. This is for automating tedious and repetitive tasks such as going to a particular site, logging in, filling in some form, getting to a particular screen, capturing the information, etc.
- Ship a few exemplary themes with the browser. Currently the user has to download any alternative themes. Of course, keep the "get new themes" command but make it a bit easier for the non-techies.
- Multi-line bookmark button bar. When you add too many bookmarks to the "personal toolbar" it just spills over the right edge with a little indicator arrow.
- Word wrapping in the browser. It's incredible to me that the browser doesn't already have this. Opera has had this feature for years in the form of a user-defined style sheet that you can toggle to instantly wrap long lines to the width of the browser window.
- Intelligent typo interpretation. Some kind of mechanism that, if the user wishes, will browse to the correct place despite a typo such as "yahoo.cmo", with a meaningful symbol or message describing why it did what it did. Also, an easy way to abbreviate links, such as "g" for "groups.google.com" and so forth. This feature exists in the bookmark handler, but I mean an easy one-step kind of thing that lowers the barrier to using it, like a "define code" button right next to the location bar.
it's = "it is"; its = possessive. E.g., it's flapping its wings.
You know, firefox is a great application, but it really needs to become more of a platform, you know, that supports a great many features for all users, most importantly, e-mail.
I seem to remember some program years ago that started out as just some crappy simple web browser, Net-somethingoranother. One day, it got an email application, and it took over the WORLD.
Adding email is always the BEST MOVE
When comparing specs and prices after a google search, I open each searched item into a new tab.
Sometimes Its better/ I want 2 pages side by side to compair specs of equipment. Right now I have to open a new browser window then cut/paste the url to get them side by side. I would like to be able to take a tab into a separate window (or 2 panes side by side, I'm not picky. my big monitor helps with this).
Also cool would be able to set some bookmarks to look it web pages have changes since you last looked at them. This is expensive and wastefull, but might be usefull as some sites update weekly/monthly and it would be nice to see if they've changed.
What about superior support for CSS3, XHTML2, XForms, etc.? Or is that a part of "Gecko" so it's not Firefox's concern? I might be a web designer, but I also realize how much these features would benefit users. CSS3 makes me crap my pants. If Firefox was the only browser to support it, without a doubt I'd make my personal site give Firefox users a better experience. And the same way I do with CSS2, give all the sites I make look a little bit more refined with a good browser. In addition, I want Firefox to index every page I visit until I clear it, by default. If information is downloaded to my computer, there is no reason why my 250gb hard drive shouldn't use 1% of its storage to remember it. If I want to find a site I've visited recently, it should be easy. Browsing through history is next to impossible when you visit upwards of hundreds of sites a day. If I want a site not to be indexed, it should hash the URL, so no one can tell what sites I'm hiding. I once tried to find a site I had visited recently. I wasn't sure if it was a few days before, or a week before. I searched my history endlessly, and I finally gave up. Google couldn't help me since I remembered so little about the page (it was a photoblog). It sucked because I knew the technology existed to let me find the site I was looking for, but it wasn't implemented. Also, Firefox should keep recent pages in active memory the same way Opera does, so when you hit the back button, the switch happens instantaneously. This is something I never would have thought about until I used Opera. Firefox is the best browser, and it's rightfully winning the battle, finally. It makes me happy that it's only going to get better.
One annoying thing about web browsing is that clicking links is sort of a leap of faith. You don't know what's behind the link. Clicking a link and reading a page isn't a very large time investment, but with thousands of links out there and only a finite amount of time to look at them all, it IS an area in which improvement's crucial.
So, I'd like my browser to show me some sort of preview of the content that lies behind each link, as well as intelligently predict whether it might be of interest to me or not.
The browser should build a list of keywords that I'm interested in, based on:
** Content from pages I've already browsed. This could be accomplished through some sort of low-priority thread which indexes page content from the browser's cache in the background, for absolutely minimal performance impact.
** The browser could weight those terms based on how long I spend on a page... the longer I spend on a page, on average, will reflect my interest level in the content. Obviously, some allowance would have to be made for times when we leave the computer idle with the browser open... stop timing once there's been no user input (scrolling, etc) after a set number of seconds.
** User-inputted terms. This would be optional, but if users are interested, they could directly input things they're interested in... ("Mustangs", "Rolling Stones", "Athlon Overclocking", etc)
Once the users' likes have been established, we can start to predict which pages they might be interested in. This would involve some "looking ahead", and pre-fetching the content of pages that are linked from the current one.
To minimize bandwidth consumption (for the user AND the server) the "look-ahead" feature could fetch only the first 1K or so of text from the page. This amount could be intelligently varied; if the page is slow to respond (Slashdot effect!) the fetch could be aborted. If the page is on the same subnet as the user, more text could be fetched for a more complete preview (great for intranet or offline pages).
Once this look-ahead data is obtained, it could be used in many different ways by the browser:
-Links that are likely to be of interest to the user could be highlighted by a specific graphical effect... colored outlines, etc.
-When the user hovers the mouse pointer over the link, preview information (keywords/phrases of interest, etc) could be displayed via various means... detailed information in a sidebar panel, or a more concise view in a mouse tooltip
I think this addresses a very real need. The only real flaw I see is the Slashdot effect would occur even if users don't click on the link explicitly! This effect is largely mitigated by the "intelligent" pre-fetching, but would still be an issue.
Perhaps conscientious site owners could specify, through some sort of markup or whitelist, that their pages ought to be ignored by pre-fetch mechanisms.
OtakuBooty.com: Smart, funny, sexy nerds.
Somewhat offtopic, but here's an alternate idea if you decide to ditch the Firefox ide: Do text mining of Slashdot posts.
- See if you can predict what score will be assigned to a particular posting by moderators. Then maybe use your prediction model on a separate corpus to determine which postings are under- (or over-) scored.
- Cluster the responses thematically. Threading already does some of this, but poorly. For example, in this article's comments you have a number of postings with ideas, some postings saying "don't ruin Firefox!", some trolls, and various other things. Could you cluster these automatically?
Compared to the Firefox idea, this domain has the advantage that there's a huge amount of existing data and it's relatively clean (as text goes). The labels are pretty clean and extensive, so a lot of data quality problems are avoided.
Stealth mode:
...) for added credibility.
...) and autoenable stealth proxy.
Give the user the ability to visit any give page i stealth mode by right clicking. Stealth mode is surfing by open anonymous proxy that keeps no logs or deletes the logs on the hour.
Themes of proxy should be configurable (ie. health, kids, cars
Have browser learn what types of pages is surfed by proxy (politics, goatse
ze dog has no nose
right-click a link, select "prefetch" and it loads in another tab or window, WITHOUT closing the current window. then when you finish reading the page, the next page is waiting. That would be very helpful to me.
Additional plugins are required to display all the media on this page.
we're also seeking ideas that will make Firefox 2.0 blow every other browser out of the water.
Only way to do that reliably is to sell to MS.
Here I was reading Slashdot, opened a link mentioned in Slashdot in another tab. Accidentally closed Firefox. Started Firebox again. Now, what was that link again? I wish Firefox allowed me to save all opened tabs in some fashion (e.g. ask confirmation on close or add preference to turn that on/off or ask me if I want to use the old set of tabs on start, etc.)
Simpy
Machine learning at the client level doesn't seem real useful. As a user, I'm perfectly capable of aggregating my own information within the browser environment.
The real win is aggregate information at a higher level. Imagine a server that has access to billions of pages of information and can aggregate and learn from that huge base of data.
Imagine Google.
-ch
a right click choice to block add banners from the originating domain...
select * from Washington DC where clue > 0 || 0 ROWS RETURNED
Don't ask me if I want to remember a username/password combo until AFTER the login has been successful.
Spoon not. Fork, or fork not. There is no spoon.
Can I have Firefox do my browsing for me?
I mean, I'm starting to be annoyed with all the time I have to spend browsing the internet, I wish all this browsing was automated so I didnt have to do it anymore.
According to ebay, that prize is worth about 95 cents!
Please add the ability for a web author to link a textarea to an XML DTD, which the browser will then enforce (allowing right-click insertion of contextually permitted tags at any point), while applying a specified CSS to the display of the textarea as it is edited. And have it produce well-formatted XML output.
So much time and effort goes into generating HTML / XML through a web interface -- this would be far more efficient and intuitive if it were supported in the browser.
(Yes I've seen 10-15 of the other ways of attempting this, like Mozile, Xopus or the whole tribe of JavaScript hacks -- they've been 'getting there' for years, but still they're slow, or kludgy, or write truly awful (browser-dependent) HTML. I think it's time to finally concede that doing the job properly means putting them in the browser itself.)
Pick a technology to solve a problem, not the other way around.
I would gladly trade my girlfriend for a Gmail account.
That is, if I had one.
What would you do with two Gmail accounts?
At least your girlfriend is safe until the first one comes along!
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Extreme? Sure. But every time WhenU or Claria or 180Solutions or DyFuCa gets dropped on someone's computer God kills a kitten. WON'T SOMEBODY THINK OF THE KITTENS?!?
k.
"In spite of everything, I still believe that people are really good at heart." - Anne Frank
I already posted some on this topic to the weblog, but:
Think of ideas involving page thumbnails.
what is machine learning?
One thing I think would be an interesting extension to Firefox is that of a way to deal with numerically-indexed pages, images, etc. Let's say you have three pages, 0001.html, 0002.html, and 0004.html. You start on 0002.html, and have no real reference to the other pages. However, anyone with a bit of sense will realize that if there exists a page two, then a page one must be there somewhere, not to mention perhaps a page three or four. A numerical look-ahead would check if any pages within a certain numerial range (say, 10) existed, and if so, it would pop-up a toolbar with each page found. The reason for such a look-ahead is in case a page is missing, such as in the example, you can still continue browsing up or down the tree. Of course this would be more useful in certain cases (images and such) than others (where the numerical value has nothing to do with content, and are randomly assigned). But I still think it would be an interesting extension.
Maybe FF could look at how long you've been staring at a page, how slowly you scroll down a page, how wide and deeply you navigate a site, how often you return to a link, etc. and use these criteria to create a history page that better shows what you were interested in at the time. This 'history of interest' might actually replace the bookmarks, unless you're the type who uses the bookmarks to 'file and forget'.
Computers are useless. They can only give you answers.
-- Pablo Picasso
The number one feature I want in firfox is automatic referencing of source material when useing OLE.
Tha way if I see something cool I can drag it to my HTML editor of choice and cite the source of what I found in one step. The Firefox should add the proper citation when it builds the data to put in the clipbord or drag target.
This would make doing web research MUCH easier. Implementation can be done in DOM and DHTML and is not that hard to implement.
In school for a research project I implementd this feature for images as a interface hack of mozilla.
Just make sure it's an extension. Firefox is nice because they trimmed it down alot and realized that a mail client and web browser are better as separate apps. Please don't make it bigger/slower for something I won't need or want.
For me, and probably for most of slashdot, customizing my interface/OS involves paring it down to only the features I use. If it were to remove the buttons/features that I never use, by itself, I would love that.
Or even better, only display the ones that I am likely to use (even if there are others that I often use), given the context of what I am currently viewing. (still have the non displayed ones in menus or something though, of course)
Social connections in CS? Google gave them out to almost everybody. I have a blogger account I only post once or twice into, and even I got an invitation from Google. I'd reckon the majority of bloggers are not CS-types, and I also wonder how one can find himself inside a blog-network.
:p
As for giving Gmail accounts as prizes or giving them to troops (as the first post to this article mentions), gee, what, did all other email systems in the world die and you can't use email anymore? I bet 95% of troops out there already have an email address, what's preventing the other 5% to signup to Yahoo!Mail, (or if that doesn't work, Hotmail, urgh)..
Speaking of Yahoo!, I wonder when they'll get to Iraq and we start to see an url like www.yahoo.iq. I bet www.whitehouse.iq is already reserved. I have dibs on www.high[er].iq and www.low[er].iq, hahaha I'll sell them and be rich!!
What time is it/will be over there? Check with my iPhone app!
Have two categories -- "junk" and "inappropriate." Allow individual users to add stuff to them, so your copy of Firefox would adjust to not let you see pop-ups for X10 cameras and "free" iPods (junk) or shock sites and porn (inappropriate) if you so chose. And the definitions would evolve to meet your tastes and needs, as well as the evolving nature of advertising and indececy. Basically the Thunderbird filter (Bayesian or what have you), but in Firefox.
Instead of indexing all the pages you've read (and allowing you to search backwards), keyword/match the current page and find pages that you might be interested in given the page you're reading now.
:|
Basically you'd make the web a realtime wiki experience...or something like that. It might be pretty confusing, but it might also be really kind of neat.
One problem with the web is that it tends to remove the element of surprise (or coincidence, or accident) from the content you see. Unlike browsing in a bookstore, where all kinds of things are squished together, with the web you only see what's on your site. The above might make browsing a bit more cross-referenced, and provide some interesting contexts that would be too hard to do by yourself
I would let the user grab tabs and drag them to where s/he wants them on the tab bar. Hmm, maybe even stacking tabs - when you drop one tab directly on another one they are stacked and you see all the tabs underneath the top tab only when you mouse over it.
Firefox needs an option to make the browser detect, and work around, user-interface abnormalities in poorly-designed websites.
It's fairly well-established that the best user interfaces are the ones where there is no discrepancy between what the user thinks is going to happen, and what actually happens.
When a user single-clicks a link, the link should open in the current window. Always. Any other behaviour (such as opening a new window) causes the user to be frustrated (or at least slowed down).
Similarly, when the user middle-clicks a link (or shift-clicks or whatever), the link should always open in a new window/tab. No oddities like "javascript:gotosite()" or "http://path/to/exact-same-page.html#" should happen.
Unfortunately, there are a lot of misguided website authors that think they're being helpful by doing non-standard things in an attempt to anticipate users' needs. This means that you'd need some type of machine-learning in order to work around these problems at the browser level.
I imagine this would be done in a way similar to how SpamAssassin works.
I may be a bit picky here, but I absolutely DETEST the Edit/Preferences method of getting to the browser options. I much prefer going to Tools/Options. I don't know if everyone (or anyone) feels the same way, but there you have it. I noticed that in 0.9 (I think; could be mistaken) the browser options were moved to Edit/Preferences, and then in 0.9.1 it was moved back to Tools/Options.
They are options, damnit, not preferences!!!
Parent 0WNED by child.
"People who come up with the best 3-5 ideas that involve the use of machine learning will win Gmail accounts"
Ng went on to announce -- Furthermore those that innovate the most original ideas will win their choice of a free glass of tap water, or dinner mints!
It's true no man is an island, but if you take a bunch of dead guys and tie 'em together, they make a good raft.
An option to add the website to a "news" list which would then either grab the RSS feed or scrape the site for headlines if no RSS is available.
Set the above up as a custom RSS feed so you can display it on your website if you wish rather than read through the browser.
Literally, a couple: - Bookmarks filtered into topic folders - Organization of history based on topic or links - so you could follow which pages you went to and how you got to them like a Gmail conversation Otherwise, keep it lean, or fork it: Firefox, the lean browser and er... Waterfox... the experimental one.
Come up with something that uses machine learning to thwart phishing scams.
Many pages are cluttered with navigational junk and ads that detract from the interesting content. Take a look at www.cnn.com, for example. The story text is in the middle, and that's what I'm interested in, but all the buttons, ads, additional information, etc. takes up a lot of space.
Automatically identifying the main content of a page, and fading everything else out a bit would be very helpful.
Some sites take an article and break it up into several pages. It would be useful to automatically recognize that, fetch the continuation pages for the article, and pull the relevent content back into the original page.
Comment removed based on user account deletion
How about having a vertical flowchart feature that displayed webpages(history? bookmarks?) in small mini-images (ie. 90 x 70 pixels) in the sidebar. Something along the lines of what iRider has but with more options/features/plugin support.
I think this approach to brainstorming raises some questions.
Yes, user studies and user feedback are part of software development, and it is great to develop software that meets the needs of users.
On the other hand, think about what would happen if every research group working on machine learning started polling users about "ideas for machine learning to be incorporated into popular program X"? How many polls would/could people have time to respond to? Their attention and willingness to answer polls is limited; do research groups who manage to get their posts on Slashdot first get all the responses (and ideas)? Are all the responses shared publicly?
And are the people who came up with the idea going to be co-authors? Who owns the ideas, academically or legally? Does the university know about this and has it approved it? How does this behavior relate to Stanford's code of academic conduct? Other kinds of polls for research have to follow certain guidelines, does this one?
I'm not saying this way of approaching research is necessarily bad, but I think such questions need to be answered clearly and explicitly to potential participants. Pre-Internet, asking 1000 people for such responses required extensive preparation (renting rooms, mailings, etc.), and as part of that, there were mechanisms to answer such questions and plenty of time to think about them. These days, anybody can do something like this on a whim, and maybe that's good, but some questions probably still need to be thought through.
You can do this in Mox, but not Firefox (that I'm aware of). What I do is find several sites that infrequently update, then tell it to check every day. I'd love it to, when I start Moz in the morning, tell me which of my sites have updated. I'll once-a-month hit a site. The current way seems hackneyed, but I have no better way to describe it.
"Sometimes a woman is a kind of religion, she can save your soul & set you free from all your sins" - Bad Examples
But what I'd love to be able to do is block/replace some images based on a previously marked image.
Example: I'm really sick of seeing tubgirl. So the next time someone uses tinyurl or meta refresh or whatever to direct me to her, I right click the image and select "Mark this image as yucky". Firefox takes note. From then on, every time a website tries to send me tubgirl, Firefox refuses. Maybe it replaces it with a picture of a big smiley face. Even if they resize it to 99% or 101%, or do any other trivial modification to it.
if the answer isn't violence, neither is your silence / freedom of expression doesn't make it alright
How about a plugin that warns a Slashdot author if he is about to publish a dupe?
:)
(Sorry guys - I know catching dupes is harder than it appears, and we appreciate the work you do... still I can't resist
Aaron Maxwell - redsymbol.net
I frequently adjust different aspects of my browser for different sites. Adjusting the window size/position, bump up font size by 10%, allow/block images, whatever.
I'd like a system that remembers those adjustments, and not only reuses them when I return to the same site, but applies them again where appropriate. 'Where appropriate' is where machine learning comes in.
plus-good, double-plus-good
Even Firefox still prints web pages awkwardly, not at all like they show up when viewed on the screen. Make it print out like you see it, almost like a screenshot, instead of printing without wrapping text properly, different frames on different pages, etc.....
Are you aware that the Firefox Bookmars sidebar does let you search your bookmarks? (At least the titles if that's what you meant) Press Ctrl-B to try it.
Have the browser store in a searchable index, the cached results for previous pages visited.
Then if links die, you want to see how a page looked yesterday, or certain information is removed from a website, you have your own personal way back machine, with all your previously visited sites.
Best part is this would not add bloat, but would rely on the users disk space capacity.
Based on the user's browsing habits, certain web pages should be preloaded. With broadband connections, this wouldn't be too important, but for people on dial-up this could speed up the process of checking multiple web pages for news/content.
Do you think you guys and the Apache guys could get together and implement pseudoserving? It would help us to greatly cut down on the slashdotting effects. P.S. The article is also available in the ACM Digital Library - get it there if you're a member, and spare my school's server the slashdotting.
Firefox has done this since forever. Right now, I get "gmail.google.com" before "gamespot.com", because the primary heuristic is visit count.
Both Mozilla and Firefox do the same thing, although the autocomplete implementation for each is separate.
Bug 78270 discusses the algorithm; this was fixed before Firefox was ever released. It was inspired in part because IE appears to use a similar algorithm.
You mean more new features.
Forget new features, just fix the bugs. There are bugs (some inherited from Mozilla) that make Firefox unusable on some Linux systems. If you want ideas for what to work on, go to Mozilla's bug list.
This would be about 1000 times more useful than putting in yet more code bloat which will introduce yet more bugs. Of course, it won't gratify your ego as much. It's a question of what your goal is - accomplish something useful for the community, or pump up your ego.
Use the (stable versions of the) EM Algorithm to impute inputs. This is sort of like automatic word or form completion except it is based on statistical first-principles of missing-data imputaton. As more and more data is gathered by the user's past inputs, imputed inputs become more intelligent. Initial imputations can be derived from inputs from users who volunteer to have their inputs recored in a central repository. This can give rise to "common sense" defaults that are context dependent. OK, so it's not Cyc, but it should work.
Seastead this.
Check Fast Forward and Rewind
All this talk about machine learning is great, but I would absolutely love to have ONE button that will quickly "pause machine learning, cookie enabling, disk caching, history logging, and whatever else". ie, a "privacy button".
A lot of times, I don't care to have my actions logged forever, but at the same time, I don't want to have to go through all settings and change them manually, or completely nuke all my bookmarks, cookies, and disk cache from the last few months.
When would I not want my actions logged forever? I can just see posts joking about pr0n headed this way, but in all seriousness:
- looking for another job during lunch at work
- searching for a surprise gift/vacation for the gf/wife while at home
- borrowing a friend's browser for a few minutes to do some on-line banking
- etc...
My 2cents...
A homepage based on the most frequently visited sites. Could possibly even seperate them into categories with the META tags.
Not a sentence!
Is this just me, or have other people had this problem?
For the last 2 weeks, I've typed "www.slashdot.org" into firefox everyday, even several times a day. It is not in my history, and never auto-completes. It never has been. I'm starting to thing it never will be. Is this an ironic bug, or a failure of machine learning?
Mozilla Firefox 2.0 should include a feature that discovers the topics the user is most interested in and do a background search to find websites of interest. These can be viewed in a "Sites to See" pane or something like that.
If nothing else, this feature would automate the tedious porn finding process, thus ensuring the undying gratitude of approximately 60% of Internet users.
On vit, on code et puis on meurt.
it's great to see a good community product getting support from some of the brightest minds, thanks for doing this
-Tim Louden
Plain GSView/gv from the GhostScript distribution does this. I was wondering where I'd seen the shadow-frame behavior before. Thanks!
--grendel drago
Laws do not persuade just because they threaten. --Seneca
me> hey, nice link, i'll click it
firefox> it looks like you're writing a letter!
me> rm -rf
krahd
mod me up scottie!
I went to CMU with this guy. And he's already a professor and here I am on my computer on a saturday night posting to slashdot.
Let this be a lesson to you kids, eat your vegetables, do your homework, and also be incredibly smart.
I have no idea if this is feasable or even relevant, but I'll suggest it anyway. Something that watches the user's actions for repetitions or a pattern. Then, offer to create a script/macro for them to use, or modify settings. Examples: User consistently opens browser, then go to a specific URL. Program sees this and offers to set it as the homepage. This would be immediate (within a half-minute or so) in case the user just browses their home page, then goes elsewhere. On certain days, user consistently goes to certain sites (ie checking web comics or Slashdot :). The program would perhaps offer to create a set of daily bookmark sets that can be used. Example: Sites visted on Mondays would be in a folder marked "Monday".
When I download an episode from Red vs Blue, I put it in a folder marked "Red vs Blue". The program would recognize that the file path is consistent for that URL, and change settings so that any download from that site is directed to the "Red vs Blue" folder.
A web page consists of several subsequent download links, and the user is downloading them all one after the other. The program would offer to create a script to do the task for the user, and let them define folder settings and file names.
Like I said, I don't know if this can be done or if its what you're looking for, but its an idea.
one clear competitive advantage in IE is the fact that plugins are auto installed.
;)
;)
:D
grandmah is unable to install firefox and plugins, nor can my sister do it, or anybody else with mental lazyness.
Of course an embed tag in a flash page tells IE to download and start installing macromedias plugin. the problem is you need human interaction.
Maybe a backend p2p mechanism for firefox browsers to excheange ideas amongst themselves. They may automatically exchange plugins and plugin installation procedures. NO access to humans, only the ON/OFF switch.
The alternative to p2p is a simple wiki page in ai.firefox.mozilla.org written and read only by the browser in the back end, to exchange information about the whereabouts of a plugin, if it dissapears from original site backup from an installed copy and so forth.
The configuration could have a list of available plugins in which you may simply click on the ones you want.
At installation firefox may ask permission to start installing selected plugins off the net, no more plugin tracking all over.
When I find a page with material with no supporting plugin, my browser may update the mozilla page with stats about my visit, to inform the plugin developer of the current demand. Human may later use the information to ask for development support or convince provider to develop an os speciffic plugin.
And a configuraction to choose my custom (no plugin here, move on) line instead of opening useless external links woulf be kewl.
For language, it's amazing I can not get in touch with a translator to spanish to see if he needs help and at release date there was already a chinese version, hello?.
The p2p concept applies AND the wiki too. Us firefox users are enough to actually contribute to translation more easilly, a simple wiki with any previous language we choose on one side, and an form on the other side is enough to get users to contribute 1 word and your translations would be all perfected is the first twodays after release.
You may choose the translation speciffic to your region, ie mexican spanish or argentina spanish. Then see one language in one side and the other in the other side, a mexican may spot a word he does not undestand and type the new language.
As for language files, Id love to just tell mozilla to install a selected file, and it to ask me if I want that language to be default for usre interface and everything.
P2p would come handy to take load off mozillas site, and distribute the latest plugins and latest translations to all installed copies all over.
Most langs in the world could be supported vbery promptly if the mechanism is straight forward and in the propper lang when loaded for user interaction asking what country the user is to start translating.
Maybe even the browsers update files may be distributed in the same way! torrent.mozilla.org anyone? heh
For search, Id like to search bookmarks as well, one folder to contain most visited, another folder to contain most something else etc, bookmark organization could take some ai all over heh.
For search in a web page, Id like to see a search all tabs, all windows and tabs and windows. So the current dialog takes a few buttons to integrate nicely (not quite ai but nice)
For installation, the first screen on the install wizard have a button for (automatic installation), click on that and moz will install with all defaults and start mozila with no more clicks.
No reboot after install works wonders heh
Firefox is light, fast, safe, cute, no need to bloat it further.
As for user habits, how about enhancing the TAB button, ie for example CONTROL+TAB can activate and dissactivate the feature, and let firefox what links I click on in slashdot
It will learn toskip all link on the side, and go straight into the first article to let me press enter to go into the link
Have we already forgotten Microsoft's Clippy?
Please no!! This is exactly the thing I hate. The key words here are *TRY TO* serve you better -- those "smart" menus almost always guess wrong!!
I don't want a program that tries to guess what I want to do next -- I know what I want to do and I want a program that stays out of my way and lets me do it.
Firefox totally rocks (except for the really really stupid name, but that's another issue) and it totally blows away evey other browser, despite the fact that it hasn't even reached v1.0 yet.
Please don't screw it up.
Really, it's been years and years since the stuff in projects that people really need, has not seen as much as an update on a status page. And you talk about machine learning?
Finish up what was planned 5 years ago, how's that for a plan?
How about fixing the program so we can EASILY select the email program we'd like to use???
I agree with the other poster, that really is a great idea! My only thing would be not just page down, but any key or mouse movement which causes scrolling. To accomodate that, you'd have to set it up so that when the current "bottom of page" indicator scrolls off the screen, a new indicator is set at the new "bottom of page"...
Well, these one are based on Firefox alone seeing as how I figure that even by the time Firefox hits 2.0, IE will still be the majority browser out there. Well, here they are without further ado :
1) Extensions:
Firefox monitors the user's browsing habits (not the sites visited, but the way Firefox itself is used) and from time to time, either automatically or when asked to, suggests extensions that the user might find useful.
Example:
Person uses lots of tabs while browsing. Firefox looks up Tabbed browsing extensions, finds the one that suits the user's needs and asks the user something like "There is an extension that would make working with a lot of tabs a lot easier. Would you like to try using it? You can always remove it if you don't like it and I will help you with that"
2) Website display:
The common mis-conception is that when a site does not display properly in a browser, that it is the browser's fault. It might help Firefox's cause to point this out to the the user in non-technical jargon. The machine learning part comes in where Firefox corrects the coding errors and presents the site the way it was meant to be used.
Something like this in the status bar:
"This website is not standards compliant. However, I have just corrected the errors so it works well for you."
3) Give Firefox and identity that evolves with it's user:
Again, monitor the way the browser is used and the type of extensions installed. Based on that, adjust how technical the level of responses are so that the browser becomes something of a browsing companion. Finally, make the browser "personality" portable.
Example:
User has the web developer toolbar installed and uses it quite often. Give more technical responses about the state of a site or the state of firefox. Allow the user to go deeper into Firefox's preferences to make changes.
User has no extensions installed, or if there are any installed, they are hardly ever used. Firefox uses an almost wizard-like interface and makes it's questions and responses to the user more descriptive.
That's all I can think of right now....will keep brain storming.
Find a job you like and you will never work a day in your life.
While this is a blantent rip off of how G-mail works, it's still a really good system for tracking correspondances.
Allow an option to track e-mails by sender, but rather then simply list all messages from that sender from most to least recent, have each indivudal exchange with that sender have it's own 'folder' you can browse through, which also inclues your replys to those e-mails.
For instance- let's say I was talking to my friend #2 (and for this excersize, I am #1), over a period of 6 mothes, over two different topics(a, and b respectively, and all we did with those topics was use the reply function (making it easy to process), and we had last talked about topic a, but started talking about topic b first; it would look something like this
#2
----b (2/2/2001), (8/8/2004)
----a (4/4/2002), (9/9/2004)
Then clicking on b or a would open all of the correspondance we had sent each other with respect to that e-mail, both there e-mails and my replies to those e-mails.
I think this would be phenominally useful, for some e-mailers (myself included)
-Millions of Monkeys, Millions of typewriters, 6 hours of sorting through faeces encrusted pages to find: This post
I would want a credibility rating on web pages.
There is a lot of information on the web but almost no way to verify the data. I would like a way for people to report the credibility of the information contained on a web page.
This is especially important with news reporting and double-extra-especially important in times of war.
It is also all too simple for politicians, journalists and other people of power to repeat the same old lies over and over.
The memory of media is short. A Truthalizer would help make it a bit longer.
A couple of nice popup management features I'd like to see is the ability to show a popup that was blocked without having to unblock the whole site. When I double click on the blocked popup indicator, I want to be able to select one of the blocked popups, click a "show" button and have it popup for me. Similarly, the ability to *temporarily* unblock a site for the current browser session only would be nice.
Here's an example: /.
Chances are that when I visit slashdot.org I will want to read a few articles from the front page. My browser could get a head start on it, and prefetch the pages for the Articles when I first visit
Granted, this will waste bandwidth. However, it would also improve web "responsiveness" a lot for dialup users. Perhaps it would be prudent to disable this feature by default, and let modem users enable it. This would certainly be better than the browsing accelerators from a few years back that didn't consider the different "types" of links and just downloaded everything in sight.
Two Ideas: First: I love the ability to create a bookmark with something like "http://www.google.com/search?q=%s" and associate a keyword with it like "gg" and be able to type "gg firefox" into the address bar and get a search on "firefox". It would be interesting if the browser could recognize sites you often visit and submit form data on and automatically create these bookmark search shortcut things based on what is submitted. Also, I like using the dictionary at dict.org, but they use a POST form instead of GET, so the search parameters don't show up on the url and makes it impossible to use the current feature for this site, it would be cool if it worked with POSTs as well.
Second idea: Selectively prefetching links based on past experience. For example, I'm at slashdot, I see a story about mozilla, I almost always check it out including most of the links in the post. The browser could look at the words in the links and surrounding words on pages you visit and selectively prefetch pages you're most likely to visit in the future. Saving download time while you view the page, but also eliminating unnecessary downloading.
We always knew Comcast was corrupt, here's the proof: http://tech.slashdot.org/comments.pl?sid=1909890&cid=34545432
Here's one idea, around the framework and deployment:
.......
... ... 3. ??? 4. profit ... ... worst ... ever ... ... beowulf cluster ... ... wah wah msft wah wah ...
To do learning, you'll probably need lots of data points for training (given these inputs, this (un)desirable outcome occurred).
Basically, when doing machine learning, you select a portion of your training data for actual training and use the rest for evaluating the performance.have some body of data that you'd select some percentage out of for training and use the rest for evaluating the performance.
Now you could do model building in a few different ways, not mutually exclusive. (1) do learning in-house, ship models to users periodically. (2) do learning in-browser, looking for cues that a positive or negative outcome has been reached by observing the user or environment.
Maybe it would be interesting for browsers to share their learnings back with firefox headquarters. Their fancy computers back at HQ could do some number crunching and test hypotheses by sending different updated models to different segments of the browser population.
This wouldn't work for a coporate setting, but more for the diy crowd that firefox seems to have attracted. A lot of firefox users probably already run software that already auto-communicates back to other servers to transfer info, like SETI, protein folding, photon soup rendering, bittorrent,
------
in soviet russia,
------
http://www.sharma-home.net/software/vote/vote.html
:)
A completely anonymous voting system that works with Firefox. I called it reader rated web pages.
I have a primitive server side implementation that does basic stuff. But a lot of improvements are possible.
It's got similarities to PageRank, but is more democratic because even people who don't author web pages get a vote.
If only someone could tell me some ideas to get a million people download this extension
Implement a smart extension manager. A tool which knows about the available extensions, which understands what I use Firefox for and then helps me to find the extensions which make life easier for me.
Intel IA-32 Optimization Manual says that self-modifying code is not recommeded. It's bad for the branch prediction, pipeline, etc.
...So the three who see this after me might read this buuuuut....
/ex
this has kinda turned into a "feature request for firefox" post-- Mine isn't about machine learning, but I feel it would make browsing *MUCH* more useful for me.
I want multiple levels of tabs! I want to go to google and search for "medieval socks" and then hit the first ten results into new tabs, look through them, and if I come across something interesting with more links, I want to tab into those, but not have them show up in the main bar-- I want them to show up so I can keep track of how I got somewhere.
other reasons for this-- news sites. I come to slashdot and fark, and read the headlines and select what interests me, open the stories in tabs, then open the articles in under tabs-- Now I can open all the stories, and I don't have to guess which article goes with which story, and if I find interesting links in the comments, I can go there as well without losing what I was doing.
I guess you could incorporate machine learning, into the first type-- you go into site 1, if you spend no time on it and close it, but you spend 15 minutes at site 2, and open up a few links from there, it looks through the others, tries to figure out which is the most interesting to you, then opens up sub-links... yeah, that'd be cool
ahhh, pipe dreams... I'd like this almost enough to figure out how I can do this myself-- but the programming experience I have is... minimal in any real systems, to say the least
I don't know if he's a "programming god," but I seriously doubt he's "some highschool kid with all summer to screw around."
Funny thing is, he just turned 19. =)
Get Firefox!
1) Design a feature that would download the whole article or next few pages of an article I am reading. For example, If I am reading a article that I have to click the NEXT PAGE button to continue reading, the program would download all of the next page for me while I am reading the current page. 2) Adapt(show/hide) toolbar and menu according to user habit.
Example: Search for 'laptop' on eBay
The user creates this spreadsheet:I must admit it would be a limited-use technology to add to a browser, but the research aspect is interesting and putting it in Firefox would be your proof of concept.
A problem would be to determine how many samples are enough for the user to get what he wants. I'd set up the interface as the user's spreadsheet taking 10% of the height under the web page and the resulting extraction (updated live) taking 40-50% of the browser's width. The user could then easily see if the results are good or not.
Negative examples might be an idea too, to "fix" computer mistakes.
I don't know if he's a "programming god," but I seriously doubt he's "some highschool kid with all summer to screw around."
Funny thing is, he just turned 19. =)
Get Firefox!
Built into some extention right now, but..
Placing a small 'similar pages' button at the status bar (like blocked popups) that when clicked, would open another tab with google news or google with 'similar pages' search result.
That makes researching news articles very efficient.
Not really machine learning, but.. yeah!
[ I can not bring myself to believe that if knowledge presents danger, the solution is ignorance ] -- Isaac Asimov
This isn't 'machine learning' but you know what would be nice? The ability to toggle a page fullscreen with no navbar, buttons, toolbars, etc. Just a scrollbar and an unobtrusive button to toggle things back to normal. Screen landscape is valuable to me.
___
It's the end of my comment as I know it and I feel fine.
not related to machine learning.
Donate background CPU time to fight cancer.
If you have to have a contest just to come up with ideas for features to add, you've gone from feature bloat to feature stupidity.
1. Make browser.
2. Be happy with browser.
3. Leave browser alone and make something else.
Who modded this as informative? Fucking mods smoking cock as usual lol!!!11111111
well, it have nothing to do with machine learning - but it would be a nice extra feature anyway :
Make it possible with a simple command to upload / download your bookmarks to the mozilla.org server.
When away from your own computer, you would have the opportunity, to load your own bookmarks - only in the program (as bookmarks) for that browser session..
a retard.
Yes, I know that there are plugins for this one, but keeping bookmarks in sync on multiple machines via server storage. Since I got my laptop, it's not a big issue, but I hate having to go back to my other machine cause that's where the bookmark is.
Enabling persistent storage of passwords. I honestly don't know how much or little of a security hole this would be, but I am constantly using the "remember my password" feature in Moz/FF, and it KILLS me when I have to reinstall and start adding them all again. Have it store that file to a spot on the hard drive (or better, server sync if that option is turned on), and allow me to keep that data even if I have to uninstall/reinstall the application.
-9mm-
In IE with the google toolbar installed, you can go "up" one level in a website (you have to add the button to the google bar in the options dialog) . Very nice when you click into a site that has you many layers deep and you want to move up quickly towards the root of that website.
--taken from google---
For example, say you are currently viewing:
http://www.example.com/articles/may/money.html
Clicking Up will take you to: http://www.example.com/articles/may/
Clicking Up again will take you to: http://www.example.com/articles/
You can use the drop-down list to navigate directly to the top level of the site or to an intermediate URL if one is displayed.
http://toolbar.google.com/button_help.html
I'd very much like to have customizeable mouse buttons.
Specifically, I'd like to be able to have the default behavior for left click to be to open a link in a new tab - only if it's a link to an external site. Otherwise, it would open the link in the current tab.
~/ssh slashdot.org ssh: connect to host slashdot.org port 22: too many beers
Let's see....what I would like to see in a new version of TB and FF....
1) Fully integrated internet applications:
Okay, this isn't FF/TB specific, but I wonder if there would be any benefit in somehow tightly integrating an e-mail client, a web browser and an IM client. I'm not just talking about linking the applications, I'm talking about a single work space with common buttons, common tabs, common favorites and address bars etc. for all of these apps. Web broswers, e-mail and IM are the three most widely-used Internet apps, yet they all operate separate from each other. I'm sure that there would be some benefit to trying to merge them (like an IM-log in your e-mail "sent" box or tabbed IM seem like logical things to me).
2) Zero-footprint applications:
I would love to see FF and TB be a non-installed, fully-contained application within a single directory (Windows side, at least). It would allow me to slap it on my USB memory key and surf the net and do my e-mail without leaving a trail. In the absolute least, that should be done with Thunderbird.
3) Aliases for webpages:
It would be cool if I could just create aliases for some sites. For example, I could just type in SLASH versus SLASHDOT.ORG to get to Slashdot. Firefox would check my manually-entered aliases first to see if I had entered in a shortcut, then try and do a DNS resolve if it didn't find the shortcut. I know that these are essentially bookmarks, but I find typing easier than using the mouse to go to a bookmark (makes me feel like a real computer user).
4) Smart Address Bar dropdown:
When typing in a site in the address bar, Firefox tries to do a match for me against previous visits. But the problem is that one site typically dominates the suggestion because of all of the sub-pages e.g. if I type in WWW.SLASH, Firefox will show me slashdot.org, slashdot.org/article1, slashdot.org/article2 etc when I really want to see all of the MAIN sites that start with WWW.SLASH, such as WWW.SLASHDOT.ORG, WWW.SLASHTHENASH.ORG etc.
Well, those are my two cents...(can I have my change please?)
Something I'd greatly appreciate is an Off button!
Unlike what our beloved First Poster here posted, I do not want sites to be automatically bookmarked and stuff like that. *I* want control over my browser. If I want it to do something I'll tell it to. Otherwise I might as well use IE, that has the nasty habit of doing stuff its own way instead of mine.
I grab images off websites all the time and save them into a bunch of different folders. I would like if the browser paid attention to where I saved certain images from certain sites and then automatically jump to the folder I last used FOR THAT SITE. It would save me a few seconds a day, at least.
Often in Error, Never in Doubt.
1) Network speed/reliability. Sometimes, I'll click on a link and stare at that animation in the top-right corner for something like 30 seconds with no results. If I click stop, then refresh, often the page comes up immediately. How about some smarter algorithms for getting the data to me faster and more reliably without so much intervention?
2) Sometimes I'll be using an ASP-style web site, where clicking a control might change only one thing on the page. It would be nice if I didn't have to scroll all the way back down to where I was. How about detecting if the new page is similar enough to the old one, and if so, scrolling the window to the same position?
3) Security lockdown. Watch for suspicious behavior (dangerous scripts, multiple pop-ups, repetetive question boxes, trying to use known browser exploits) on the part of a web site. When these things happen, lock down security on that site.
4) Kid detection. Based on biometrics or usage patterns, figure out if your 7-year-old is using the computer. If so, turn on the Net Nanny software.
5) Maybe too far off-topic, but software that will watch your 7-year-old's chats for signs of the other person being an adult stalker.
6) Watch for information your 7-year-old shouldn't be telling others, such as his phone number, full name, etc.
7) Watch for the usage patterns of a clueless newby. When this happens, offer to help the user learn more about using a browser.
8) Cat detection. When the cat is walking on the keyboard, keep him from deleting all your files.
9) Biometric music selection. Watch usage patterns and biometrics and select/suggest appropriate music from the user's play list based on past choices.
10) Goatse detection. Block all goatse mirrors.
Donate background CPU time to fight cancer.
People regard their privacy highly and are reluctant to provide their personal information to websites. Website which need personal details from its users have to convince a new user that it won't misuse the data that it has. This is usually made with a statement, known as Privacy Policy.
Privacy Policies have problems. First, they are usually written in a legal language incomprehensible to a lay person. Second, people have no way of knowing that website in fact follows the policy it has.
One way of assuring people might be machine readable privacy policies. P3P (Platform for Privacy Preferences) is W3C framework for such privacy protocol. It allows user to store his personal data to a P3P agent, which will then follow the user given rules to share private information.
This agent should be implemented to firefox and it could use a machine learning to automate further the interactions with websites. The agent could for example learn, that if website's privacy policy promises to use user's e-mail address only for initial consistency check and to send a forgetten password if user explicitly asks for it, the agent can give it to the website without prompting the user.
Of course, this won't solve the problem of malicious websites which don't follow their privacy policies, but is a step into right direction where privacy policies are certified and their enforcing is auditted.
One feature I would happily pay for would be the ability to turn sound in Flash content. Flash ads with sound are just about the most annoying things on the internet bar spam.
One of the most annoying things I find about surfing are those bloody animated advertisements based on Flash and such. Maybe the popup blocker could be taught to automatically put an empty box of the color of your choice over anything that's animated? That would quiet down your field of vision considerably. Also a few buttons so you can quickly (de)activate the resident plug-ins (from java to flash) might help.
Jynx
A positive attitude may not solve all your problems, but it will annoy enough people to make it well worth the effort.
When I am looking for something, using google or another search engine, I often come across "spam" websites that have computer generated content or hidden text. I really hate the ones that make themselves out to be a search engine or a directory when they contain no real content... content is what I came there for.
I'd like a button that I click on that says "this is spam" which adds the content and URI to a database. In the future when similar pages are found in the search engine results, it could either remove the result altogther, or identify the link as probably being to a spam website when I mouse-over it.
I'd like an option that I could configure so that the URI is automatically reported to the search engine that I followed to the website.. if any. Enough complaints and a human reviewer would know who the worst offenders are and kick them out of the SERPs quickly. Right now its too time consuming to report the spam websites now, so a way to make it easier would definiately help clean up the search engine results.
Another idea would be to submit the URI to a centralized website with a database of spam websites.. the search engines could use this centralized DB to pare down their results to only non-spam sites. That way the spam reports could be used by all search engines.. and other browsers too..
Here's the thread about this cool article at MozillaZine:
7 69 7
http://forums.mozillazine.org/viewtopic.php?t=9
Sorry, more code doesn't mean slower running. :P
Just click on the sidebar. There you have Bookmarks, contacts, history and links on the page you're currently viewing, all searchable!
Arg, I'm really sick of the - um, technologically ignorant - that don't realize that more code doesn't have to mean slower running. This is like %90 of the damn posts so far... whining that adding anything will slow down Firefox...
After all, if the code isn't being used, it's not exactly slowing things down is it?
Doesn't exactly take a rocket scientist to figure this stuff out people...
Heh, I love modsmacks like this for useless and obviously biased or just plain incorrect info.
So like all three definitions of the parent, ya...
Well, a feature which makes the browser able to block Flash movies from certain hosts. Exactly the way it does with images. (Right-click on the image and select Block images from ...)
Konqueror. What you are talking about is a file manager + browser!
My Aurora : http://www.youtube.com/watch?v=o91ZsGwJYyg
FB : https://www.facebook.com/TanveersPhotography
Sorry if these have been suggested before, I'm hardly going to read through almost 500 posts to check.
/., then another forum and then GameFAQs) make a button that starts you on your quest, and then when you click it again you go to the next page. Alternatively, just allow that functionality to be created manually. Being able to recognise where to go next would be very useful - when I'm on a forum I'll usually want to go to the next one from the last topic that I read, so being able to recognise that I'm at the same place even though it's a different page name would be a godsend.
1. Check Javascript onloads so that they run when enough of the page is loaded. For instance, the URIid extension runs onload, which means that site-specific styles are applied only when the page is fully loaded, causing lag if you have a big page with lots of images. Since the only thing it modifies is the attributes of the body tag, get it to run as soon as the body tag has been opened, but before any of its content gets loaded. As a second-best, loading the entirety of the relevant tag would be cool as well, if a bit useless in this case. Also, I'm not sure if this already happens, but firing onload before downloading images would speed things up.
2. If you frequently visit certain sites in a particular order when you start the browser (like me; I visit various forums, then I read Dilbert and Agnes, then check my livejournal, then come to
I know that it is planned for MSIE, but this would be a nice feature. If it could store browsing history in a central location, it would be great too, because often I can't remember if I'd seen something while surfing at work or at home.
Some kind of a history.html file that is automatically modified on ones personal webspace could be a an idea.
Sigs waste bandwidth in 56k land.
Don't know if this is already suggested as I've not ploughed through all the other comments yet but here is my suggestion.
I go to several pages each morning (Slashdot, BBC news etc) and have them bookmarked as a group in the personal toolbar folder. Some days however I also go to other sites, for example on Thursdays I go to the Linux Weekly News site as well as the rest. So my suggestion is to automatically handle this within a single 'Regular sites' button. In other words, time dependent elements within a bookmark group.
Some examples to illustrate how fine grained this might be.
"Slashdot" always
"BBC News" always
"Dilbert" first time after 9am everyday
"Linux Weekly News" first time on or after each thursday
"Evening TV listings" if time between 5pm and 6pm (ie last check before I leave work)
"Whatever page" if date is exactly 1st of month
"Whatever page2" first time after 1st of month
"Stupid goofing off site" only between 12pm and 1pm (ie Lunchtime reading)
You get the idea.
...and it's not even an extension!
:)
Short way:
Bookmarks -> Manage Bookmarks. The search box is right there.
Longer way:
Just right click on the toolbar, choose customize, and add the bookmark button to wherever you see fit. When you're done, click that button and a side panel will open with all of your bookmarks and a handy-dandy little search box at the top.
--Anonymous Coward
it could predict that you don't want any prediction?
--Anonymous Coward
first off, firefox is a kick-a$$ program! ...
... maybe ...)
... good god, the ruin ...
please don't integrateit too much into the
desktop or os. internet explorer those that
as you can see for yourself where it went
acctually i was taken back abit, when firefox
started to add registry keys. it was soo
convinient, do just drag and drop firebird
from a *.zip to a directory and it worked just
like that. that was most amazing for a browser,
kindda like lynx. no heavy registry usage please!
what would be a neat feature is like a link preview
feature. in windows XP you can get a powertoy
add-in to "alt-tab" that lets you see the open
windows content in miniature form
firefox could get that too, but for links on
a page (this requires a fast internet connection
of course
the top right "i'm loading the page" animation
should be put to better use. just having
a animation doesn't really cut it.
keep up the information on the bottom left
part of screen, that shows what firefox is doing.
it is soo annoying in internet explorer to
just have the wave flag icon tellign you that
it is loading soemthing. firefox usesage should
be informative. it hink opera does an even better
job of informing the user what is going on.
and at last, what the f#ck is a google search bar
doing in my firefox? there are gazillion search
engines and why the F#ck those it have to be
google? i mean i click my homebutton, that takes
me to my local intranet apache webserver, where
i have copy pasted the forms code from google
into my local index.html
to every greattechnology ideas is trying to
get the user that is 30 years old and still sucks
his thumbs and occasionally craps into his pants.
so stop this integration bullsh#t!
don't go the microsoft way. firefox is a kicka$$
stand alone browser and please keep it that way.
if you feel like coming up with the next "killer app"
don't piggy back on firefox. start a new program.
that's all folks
There used to be a mozdev plugin that would block all flash windows, but allow you to click individual ones to let them play. (It's no longer maintained, doesn't work with newer mozillas, and you can't even leave notes on the homepage anymore.)
All you'd really need at this point is something which remembers what kind of flash app you clicked on to allow to play, and once it saw a pattern, maybe it goes ahead and lets them play. Etc, etc.
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
An off button.
And what about all pr0n lovers - "Private Browsing" feature found in Apple's Safari browser - when You turn PB on, you don't leave any traces in the browsers bookmarks, cookies, cache, et al.
-- echo '[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768
I'm always wary when I see someone provide a solution (in this case, machine learning) before they even have a problem, and then on top of it all, has the gall to call it research. The author is essentially saying, let's just apply ML in every asinine way and see what sticks! Machine learning, very strictly, refers to any system that improves performance with experience. It's a tool, but if you attack a perfectly decent product without a goal or problem in mind, just to use this shiny new tool of yours, well, suffice it to say, it won't be pretty. Although, since ML is a pretty hot topic right now, it sure will make a compelling title for a paper...
~ds
(contemplating the effects of applying ML to toast)
Seriously, how much would bayesian porn detection be useful today? Nowadays a parent must purchase 3rd party applications that almost never work because these don't have the necessary sofistication to detect if a page is acceptable for a minor or not.
A feature like this, complete with safe browser locking and unlocking, would give Firefox another boost over its competitors and provide a better web for children and people living in places where pornograhpy is illegal.
Hack your mind out of its sandbox.
come up with an idea that will put firefox ahead of the competition in the globally important browser war, and you get (drum roll)...
an email account.
lame.
I've been playing around with some ideas on distributed evaluation and rating, and if anyone can direct me to folks who discuss and work on this, I'd be appreciative.
I'm a bit too far out there for immediate inclusion into Firefox, though...
May we never see th
Blake was co-founder of the project; whilst still a high school kid if I'm not mistaken. Thankfully age is not a good indicator of intelligence, you could be 90 for all I care.
.... Auto Typecasting of the user just begging to be collected up by government and corporate for misuse against the user.
...... Hmmm, of what? Being fucking to lazy (promoting ignorance) to better learn the system yourself, as a user?
Not to mention what a fuckinhg pain in the ass it will be to try and use a system/browser you don't own to browse the internet (IE, co-workers in helping them or tem,poraryly using their system, or internet cafe where everybody and their brother...or general work (many use) system...)..
Or hell, how about a home system where the whole family uses it..... where it becomes an internal machine learning conflict between what the different members of the family are interested in...
Machine learning built intro a browser is a very very bad idea that can only lead to greater user frustration overall and that's nt a very good trade off for what woudl be a much smaller benefit of
Hmmm, what sort of auto-typecasting would such a machine come up with for such a lazy ignorant person?
What next, machines that think for you, or are conscious for you..... of what value is left in life....???? Ok so that's an extream but maybe its good to look at where it leads, even if it may never actually get to the extream (yeah right, there are just to many of us to eliminate that possibility from the possible spectrum, should such be implimented...)
"But honest your honor, the machine told me to do it...and I'm to fucking lazy to think for myself...."
How about fixing some of the many top-voted bugs first? It's not glamorous but it would help make Mozilla/Firefox more useful in a real way. Top voted Bugzilla bugs
For every expert, there is an equal and opposite expert. - Arthur C. Clarke
Say I discover a new webcomic, or somesuch. I'd like to read through the archives before starting to follow the main thread. This isn't usually possible in one sitting.
So, I'd like to make a bookmark that remembers my place on a site, and updates itself when I visit and read some more.
If I've caught up with the front page, the mark should take me there. If I miss a few comics, it should drop me in the archive at the correct place.
I'm scared of numbers that can't be written as a fraction. It's an irrational fear.
Title says it all. How about forgetting about the machine learning in Mozilla/Firebird, and instead make a totally reliable browsers. I am not saying Moz/Fire are unstable (not like IE), but lets learn to walk before we can run.
"But honey, I have to browse porn sites! I'm just training the filter!"
USE HOT GRITS WITH STATUE OF NATALIE PORTMAN (NAKED AND PETRIFIED)
I *Hate* menus that change by themselves! Do not incorporate learning menus in anything! It just makes it harder to find what I am looking for if the menu items move around!
Ok, I feel better now. On a calmer note, I think programs that "try to be smart" are mostly annoying. I suppose it is *possible* that machine learning could be a good thing, just make sure I can turn it off.
Firefox could keep track on how the user got to the page that he/she wants to bookmark in order to suggest where the new bookmark would be placed.
While having many bookmarks the best way to organize them is to catergorize them into different sub-menus. So, if my browsing session originated with the bookmark: Programming -> Prolog -> MiscPrologsite14.com, and I found an interesting page that I would want to bookmark Firefox would suggest placing it in Programming -> Prolog.
Granted, for geek power users, this would not be widely used, but at my house, I could put the wife and kids and on thinner machines and just keep the one "server" box beefed up (more ram, etc), it would give me one place to update the browser, etc
(I know that I could do a lot of this using an X-Server but until Reader Rabbit runs there along with 50 other kids games, I can't migrate them)
A networked "browser server" with an ultra-thin client side "rendering browser" would certainly be an attention getting killer new feature.
Agile Artisans
1. Acknowledgment that Firefox is ONLY a browser, please do not start overloading with a bunch of plugins and stuff that we'll never use.
:D
2. Increase of speed. Maybe a review of code and algorithms could help you to make a lighter browser without losing any of Firefox great capabilities.
3. Better accounts/saved passwords management, without risking too much security.
4. Better integration with other stuff like the JVM and Flash Player (although I don't have any problems with ln -s some people prefer it the easy way).
5. Posibility to expand the Google search, to look out for Images or news.
As some guys noted earlier, a few things like the pop-up management (actually is almost-great, but it still lacks of some stuff to recognize really needed pop-ups) and making default the Open in new tab for links could improve a lot the Browser.
If you want me to increase my geekness level, send one of those invites tofebuiles-nospamplease@inbox.lv
You could make Firefox give an educated guess in case of a typo. If it learns I like tech sites, I probably don't want to go to www.slahsdot.org or www.micorsoft.com.
I'd settle for an exclusion filter for every URL based on a domain typo, with a forced override in case I really felt the need to go to go to googgle.com.
Before people make the inevitable remark about bookmarks, there are dozens of pages I like to visit regularly, and it's easier for me to type in the URL than go through a huge list of bookmarks.
Switch back to the previous tab when closing the current one instead of just switching to the rightmost tab.
:/
A bit annoying as it is now
Blog/post segmentation into entries. Organize into expandable tree form
with browser-side filtering of slashdot-type posts based on a learned
per-site model of segmentation boundaries.
learn to prioritize entries according to user preference
click-through probability or time-on-screen estimates
using bag-of-words n-gram model
selectively display of all funny>3 posts, for example
show all posts/threads mentioning openbsd
this would be especially useful on freerepublic
should be able to classify dom-nodes, don't need to segment text
Find messages semantically similar to the one I'm about to post
TF/IDF would probably do well enough to be useful
Segmentation of content from stylesheet/advertising boilerplate
Auto security zone changes based on perceived threat level of site
frequently visited sites can be allowed to use cookies/js
would benefit from pagerank-type aggregate measure
search-query aware auto-summarization of pages
right-click on a google link, and select "open in new tab and
summarize", will get an expande form of the google summary
with page keywords/phrases and matching text/dom nodes at the
top, and the original page below
auto-open and integrate the follow-on pages of a multi-page article, especially the current 15 page hardware reviews. Remove the duplicate stylesheet junk.
"don't show me things like this" auto-inference of dom nodes or regexes to rewrite the page to hide parts of the boilerplate and create a leaner page.
alternately, "extract this region of the page and incorporate it into a composite page" to allow the user to build his own rss-type portal from several pages, which can be automatically refreshed and bookmarked
clustering of history pages by content or referrer - DMOZ or inferred
LSI categories, and (independently) viewing the browse tree google->somepage->someotherpage->foo.jpg
other non-ml feature ideas:
detach and drag/drop tabs between windows or to their own window. also to reorder them.
search across all tabs, with "find next" moving to the tab containing the next match
fuzzy search matching by stemming, or near matches to catch misspellings
trap crashes and dump the list of open tabs
It isn't freeware, and if someone wants to clone it as a freeware, Firefox-aware app that'd be great, but it does just what you are asking for.
It narrows the bookmarks as you type, based on title, URL, and keyword fields.
http://www.kaylon.com/power.html
If Firefox had this built in via a search bar or some such it'd be awesome.
Visceral Psyche Films
The most important features in a browser are stability, security and speed. And since it's been neglected in mozilla/gecko let me mention speed again.
Correctnes is another topic that I think has been neglected in the past. There's a number of problems with drawing things 1 pixel off at times, both in the UI and gecko part of mozilla.
I'm also missing the capability to see all relevant "cookie sites" for the current page, so that I can re-enable cookies for bla.com, www.bla.com and cgi.bla.com without manually pageing through thousands of alphabetically sorted hostnames. I'd also like to enable/disable cookies by name (i.e. allow SESSION_ID but deny PREFERENCES), of trim their lifetime.
Proper multithreading: The tab and menus/dialogs that I'm currently interacting with should get higher priority over everything else. Also. mutithreading should be finer-grained, so that no activities in any tab/windows block other windows or even the UI.
Similar the the adblocker for images, I'd like to block iframes loading from certain sites. Currently the only way to find out where Iframes come from is by looking in the source, and to block them, one has to edit one's hosts file, auto-proxy-config file or junkbuster config.
I'd love to be able to increase (adjust) the width of certain columns of some pages. With display resolutions beyond 140dpi and sizes of 1920x1200 and more, pages that are fixed at less than 800 pixels wide with menus and advertisements on both sides, leaving a small column of content that can only be read with the original 5 pixel high font the page author loves so much, and only leave two or three words per line with even a moderate increase in text size.
Did I mention speed? In this respect, Netscape 4 is the browser to beat, in all disciplines except deeply nested tables.
Machine learning?
Give me a break.
What I'd like to see if I could finally use vim for these damn textareas (or any editor of choice for that matter).
There is a needed feature on almost every software out there: speed. Look at Opera, is *FAST*, if Firefox could render pages much faster, the number of users attracted by that speed will be much higher than the number of people atracted by a "learning machine"
A "Bypass /. effect" extension would be nice :)
Remove spaces (not linked to save Slashdotting):
x 0_ 9/bookmarksftp/
Saving file types to various locations:
http://downloadstatusbar.mozdev.org/downsort/
Portable Bookmarks:
http://cgi29.plala.or.jp/~mozzarel/addon/firefo
Visceral Psyche Films
This way a user can always review this stuff with complete ease of use.
People don't exist to serve systems, systems exist to serve people.
Should the AI generate shortcuts for us based on our use? Say we keep doing some repetitive task, should the app then say "press blah blah" to automate this task, and then that key combo becomes the automation shortcut? I assume if someone does a task a certain amount of time they might want it automated
People don't exist to serve systems, systems exist to serve people.
We need the ability to download sets of pictures which are thumbnails. The AI should automate the task of us having to click on each one. We should be able to download all thumbnails at once.
People don't exist to serve systems, systems exist to serve people.
Any AI, and cool stuffs, leave it to research mode, until it is mature enough before incorporate into the browser.
Open source usually move in a relative constant pace of engineering. It took a long time for mozilla to get to where it is today. So, don't waste big effort on big stuff for near future release, you'll won't make it.
Another side is be practical. There's so many things that users need that does not work right now.
For example, many users still have modems (many millions), make it smaller would help.
Make it faster, the faster it is, the better it is (this is my father's main complain, he's a bit happier with the latest firefox, but still). Instantly on is a big plus. I once open a serval megabyte html using both IE and Mozilla. Guess what? To my surpise, IE took a fraction of the Mozilla time. This is fact, you can try it yourself (this is the Definitve guide for Netbeans).
Still many sites does not work well with Mozilla, for example, go to Netbeans, and then ctrl-+ couple times, you'll see what I mean.
Plugs in and other stuff is still something not seemless in Mozilla. I still have problem with java and Mozilla (not working). I think because I use the Zip version of the browser.
It's a matter of time before Video and 3D is in used inside a browser. Don't be behind on this. If M$ has this, it'll take years for Mozilla to catch (note how long it took Mozilla to be here today, they keep missing target, because not enough resource. Commercial company can shift resource and give it priority, $$, open source cannot do the same (some shifting, but not much).
Flash has its place on the web before people know it. The browser missed it. It's matter of time before M$ have something way bigger and better (they control the distribution, remember?).
As the broad band become common, people will shop on line, and virual reality is soon to come. Video on demand, etc will be soon here.
Xml display is still way too slow. Web development for mozilla plattform still have alot to be desired. (if you works with a heavily scripting and dynamic site, you'll know what I mean)
Make the html composer way better than what it is today. This things have potential, but it's a second class priority at Mozilla, and not much has been done. It's really cool tool though.
The key terms here is practical. Whatever users want and often use, and pontentially will use in a near future, better provide it. Whatever is not working now, fix it before dream too far.
This has nothing to do with machine learning, but would come in handy for people checking multiple email accounts. Or, like me, use PopFile and filter email into multiple folders based on classification.
What I'd like to see is to be able to configure which folders get an audible (or visual) alert. Customizing which sounds go with each folder might be cool, but on/off control is the main thing. I'd rather not be interrupted when something classified as spam or trash (based on blacklist) comes in, but don't mind if it's likely to be from a customer.
My $.02
I was originally thinking of appying this only to Google, but it might have merits outside of google.
Make a "Research only" and/or a "shopping only" button/option.
I would give my first born if I could do a google search and not wade through a thousand web-sites that are trying to sell me the widget that I am searching for. (eg. I want to read reviews and tech info on a new video card, the number of sites that are tring to sell me the card overwhelms the pure "review" sites.) Yes I know that most review sites also sell the stuff too.
(I didn't say it would be easy, but I would like it.)
Like I said, this could apply to the whole web and not just google.
"The price good men pay for indifference to public affairs is to be ruled by evil men." ~Plato (427-347 BC)
Microsoft have just read all of your posts and have applied for patents for all of the ideas listed...
Jiggity
The following is the best e-mail idea of the last two years, and should be a feature of every e-mail program.
As a person who writes a lot of e-mail, or as a manger, one major organizational problem is simply not having your messages replied to. It currently takes a ton of manual effort to decide which of your e-mails need a followup by you. So much effort that nobody does it. Some questions fall through the cracks when you write a hundred e-mails a day. Closing up these cracks would measurably, demonstrably improve the effectiveness of managers, project coordinators, and any other heavy communicators, since e-mail has become the most important and most used mode of communication for a large number of people. (It is very rare to see a manager actually send out an e-mail asking "Was this resolved?" Generally, un-followed-up e-mail is simply forgotten.)
A system for followup could be partially automated.
1. A checkbox exists when writing an e-mail. If turned on, it tells the e-mail client, locally, that you want to make sure this e-mail has been followed up. I'll call this a "follow-this-up", or FTU, e-mail.
2. After an FTU e-mail has been sent, the e-mail client remembers it in a list. Once an FTU e-mail is sent, a copy of it is placed in a mail folder which I'll call the FTU folder. The user of the e-mail client can open this mail folder at any time to see the FTU e-mails that still have to be followed up.
3. When the client detects that a recipient of an FTU e-mail has replied to that e-mail, then it provisionally removes the FTU e-mail from the FTU folder. (Probably the e-mail is grayed out in the list but not actually removed.) Making this detection complete and thorough is an interesting problem. The starting point would probably be based on receiving an e-mail from the recipient with an appropriate subject line (e.g. the same subject line prefaced by FW: or Re: or Re[5]:). And in order to increase the effectiveness of this technique, the client might actually maintain a database of previously-used subject lines that are already in the FTU folder, and nag the user if he sends a second e-mail with the same subject, asking him to write a more elaborate Subject line.
Other starting techniques could include parsing the e-mail's content to see if part of the content matches an FTU e-mail that has been received. Or by utilizing e-mail fields or even implementing a new e-mail field which hopefully doesn't get stripped when the recipient replies.
This system, then, tries to ensure that after an FTU e-mail is sent, there is either a copy of the e-mail in the FTU folder so the user can see that the recipient hasn't followed up, and the user can follow up with a question; or there is a response from the recipient in the user's Inbox.
User interface is critical to making this system useful for the user:
4. When the client sees that an FTU e-mail has been replied to, it presumes that a followup has actually occurred. This obviously may not be true; the recipient may have responded with a joke, or with a followup question, or with "I'll get back to you Thursday". Presumably when you read a followup to an FTU e-mail, a new bar of UI should appear in the client saying to the user "This looks like a followup to an FTU e-mail you sent, which you can view by clicking here." Buttons would let the user choose things like "Yes, this resolves my FTU e-mail completely" or "No, I still need a followup", or "I now want my reply to this e-mail to be an FTU e-mail, and not the parent."
5. I imagine that the FTU folder displays its FTU e-mails in date order, showing the oldest non-followed-up e-mail at the top (colored red after 2 to 4 days or so). The user can mark these e-mail copies as already-followed-up (i.e. it's resolved, no more followup needed, because the recipient saw me in person and resolved it).
how about this for a useful feature. :)
User highlights a block of text containing links right clicks and chooses open links in tabs and to make it really sweet doesn't open duplicate links.
surely this is am easy one to do.
Please tell me where i can download the plug in
Blarney Quality Restaurant, Plants
Dont do it.. dont bloat things out with 'features' just because you can..
This is what makes IE such a overloaded hole-riddled mess...
---- Booth was a patriot ----
1. Something to detect and disable really, annoying flash ads. Adblocking technology doesn't seem to work with flash, you have to rip it entirely out of your browser, or put up with annoying, flashing ads next to whatever you are reading. These things just burn up CPU cycles.
2. Detecting and disabling the very obnoxious ads that float around the screen.
3. Faster launch and reload times, smaller memory footprint
4. Better support for plug-ins. I was tring to move my mother over to Firefox, but couldn't get the streaming audio from her favorite radio station. Actually, how about emulation support for IE plug ins? (Yes, I would rather see native versions too, but the world doesn't spin the way you want it to)
My rights don't need management.
1 download/install the "popular" extensions according to the users "profile" ie Newbie /Geek / Hacking God /C*O
2 Match theme to WM settings
3 Suspect site warnings (spyware/ hijack sites ..)
4 preload bookmarks according to platform/profile
Any person using FTFY or editing my postings agrees to a US$50.00 charge
For all of the idiots who flog the bloat argument, please use Dillo or Lynx. I want features. If that means I need to upgrade my computer every 18 months, so be it, I'm employed.
Machine learning might be extremely useful in the development of plugins and extensions. Take Leech for example. It'll download all files with user-specified file extensions that are linked to on a given page. Some photo galleries (and no, I'm not just talking about pr0n here) don't link directly to JPEGs and the like, but instead link to a page that has the JPEG somewhere on it. I know that Photoshop CS does this, because I'm making a gallery of travel photos and it's annoying the hell out of me. With all of the different software packages used to put the galleries together (including CS), it would be a real pain to try and code for each and every implementation and version of a web gallery. If users could train it, or if it came pre-trained with only refinements necessary, that would be extremely useful. Machine learning could also be learned to train other extensions like AdBlock and the like.
Apple will be including a "Private Browsing" feature in Safari when Tiger comes out (not sure if Safari currently has that). For those that don't know what it is, you can tell Safari to not write anything to the history or the location bar for as long as the feature is turned on. If Firefox could be trained by the user to do this, and do it automatically (with a notification icon in the status bar similar to the junk mail icon in mozilla mail & thunderbird), that would be great.
The next (or next after next?) version of MS Visual Studio will include a new profiler that does part of this. The developer will compile an instrumented build of the app. Run the app in common senarios. Then recompile the app with generated statistics.
Doing this in realtime might be useless though. Difficult to gather statistics fast enough. Plus the lack of info about the source code (i.e. debug symbols).
IMHO, this suggestion should be given to the GCC guys and gals, not the firefox folk.
...a big, obvious button to turn all this machine-learning bloat off.
In my experience, computers are generally horrible at anticipating what I want to do. In a contest between a computer and a user, the computer never knows what the user wants better than the user does. Even if the computer happens to be right, it's still not worth getting all up in the user's interface about. Keep Firefox clean. That has always been my favorite thing about Firefox -- it lets me get to what I want to as fast as possible, and otherwise just stays out of my way.
Scene: A computer, with Firefox 2.0 running maximized on the desktop. The homepage is loading. The USER, your average geek, is waiting patiently for it to load.
A light-yellow bubble pops up on the screen. It obscures half of the interface. FOXY enters bubble right, looking like a cross between a red fox and a Pokémon.
USER: Eh what?
FOXY (Text): Hi! I'm Foxy!
It looks like you're trying to view a web page! Would you like to:
USER shoots self in head.
Seriously, if Firefox is going to incorporate any kind of machine learning features, the team ought to keep in mind Microsoft's mistakes. As a general rule, I'd say that anything that the user is required to interface with while trying to complete another task is a bad, bad idea. Any new features need to be as transparent as possible. Less is better. In fact, maybe none at all is best.
Rather than add to the already great list of ideas, my idea looks at the overall design architecture. Machine learning should be a minimalist venture as any other features added to a core application should be.
:]
I would like to see a Mozilla machine learning interface (MML*) where I can manage the various ML modules. This UI component should be able to point to a MML extension page that will have all the registered MML extensions.
This design would ensure a consistent management interface for the machine learning extensions as well as making the smallest footprint on the base install on Moz.
*MML - TLA score= 4/5
That's not the case.
The problem with HTTP basic authentication is that the password is sent in plaintext over the network (unless you're using SSL/TLS), and someone might capture it in transit (as said in the linked text you provided). That is the same thing that happens when you POST a form with your password (unless you're using SSL/TLS, like in the other case).
Your password is vulnerable to sniffing with a form as it is with basic authentication.
However the problem was addressed with more advanced method (such as md5-digest), that are they are not useful to an attacker (at least not trivially) when intercepted.
GPG 0x1B479C78
Rather than have something like a profile setup for my homepage(s), have an option of not only a Home Button but also add in a component for having multiple 'home-like' buttons.
For example: News Home, Shopping/Deals Home, Gaming Home, Email Home
Or not name them home, perhaps just make it available as an option when customizing the toolbar to have a 'home-like' button that would open up a set of webpages simultanously... which in customizing the button, would make things more friendly for someone who routinely visits 25-30 websites on a daily basis.
Currently I'm using the *WONDERFUL* http://url|http://url format of opening multiple websites as my homepage, which works wonders for things that I need usually when I open the browser, but fails when I need to type in URL or lookup bookmarks.
As it is right now, if a save a bookmark or quickbar link as URL: www.1.com|www.2.com|www.3.com, the current implementation only loads the first website.
http://www.fsckin.com/
I've been using searchable bookmarks for some
time now. I just export my bookmarks to a file,
and then make that file my homepage. Then
click home and CTRL/F and there you go!
The browser loads nice and fast too because
the home is on my machine. Even works in IE
because I make that it's homepage as well.
What I Really really want in a browser is macro recording. And here is what I mean by that:
Let's say I am on a page with 40 small image previews (thumbnails) that are clickable and lead to another page with the full image. Let's say I want to grab all full images only, not the HTML, not the thumbnails but the full images that thumbnails lead to. Well, no browser allows this kind of functionality. What I would like to be able to do is to start recording a macro, and it would work like this:
1. Click on the start recording macro button.
2. Right-click on a thumbnail and ask for a new browser tab or browser window with the link from the thumbnail.
3. Wailt until the window opens and right click on the full image and save it somewhere.
4. Close the newly opened tab or window.
5. Click on the end recording macro button. Provide a name for the new macro.
-----
Now with a right-click on a thumbnail I should be able to go to the right-click mouse menu, find section for macros and select the newly created macro. The macro should open the link from the thumbnail in a new window or a tab (whatever I did) and save the image found in the same HTML position as the image that I saved (skip other images and text from that page.)
-----
Even more, I should be able to select multiple thumbnails and play macro for all of them (let it play sequentially, one thumbnail after another.)
-----
Well, can this be incorporated into the next Firefox?
Thanks.
You can't handle the truth.
... I do not want a program to become psychotic learning from the variety of pr0n sites I visit.
My biggest problem, and this is with only a few sites, yet somehow they are the sites I use more frequently, is that I need IE to browse and access these sites. These are usually sites that need encryption or flash or java or something of some sort that only seems to want to work in IE. They specifically have some note mentioning how you need "...IE 5.5 or higher to access this site...". Maybe the problem lies with my inability to set up my Mozilla/Firefox correctly. In any case, my suggestion is making Firefox so that it would be able to atleast decently handle sites, that "needs IE" to work properly. Also, making it easy, and maybe even automatic, for Firefox to install and add in whatever is needed to browse whatever a website throws at it.
If this could happen, I would give IE the final boot I've been wanting to for so long.
... I do not want a program in my computer to become psychotic by learning on all these pr0n sites I visit.
Your head a splode
Hi,
.net and java for intranet applications.
This post is absolutely lame. Outside of using machine learning to make the browser faster and more responsive, I think the concept is totally idiotic. On the other hand, if you want ideas for projects go to mozdev [http://www.mozdev.com/]
I mean comon They do all kinds of fancy stuff. Why not work on the SQL javascript plugin? Or all those other plugins.
But, I think adding libraries to Javascript is the way to go.
I mean you can't extend HTML and you can't extend CSS. But you can always extend Javascript cause it's what we programmer fellas call a programming language.
If Javascript had a database library similar to the DBI, pear-db, or JDBC and improved, less buggy drag and drop support; it could probably defeat
Once it is used for intranet and developers found how easy it is to only use one language to make web sites, they would make a push to create user applications with it.
Why it is that everyone underestimates Javascript is beyond me; give it a standard library outside of the browser library (which is just riculously nice) and you won't ever have to write in HTML PHP garbage again. Browser code could become elegant.
I know this is redundant, but I think that it's necessary to reinforce it, as it is a very important thing in firefox.
Whatever you do with those ideas, make sure they are modular (so that they can be removed) and (above all) DON'T BLOAT the browser (even by making a plugin system that's bloated).
I might want to run firefox on an old machine a few years from now, and those things might make it so slow that it would be useless.
As an example, I'm writing this on a six year old pentium-mmx 200, and all my traffic goes thru an eight year old pentium-120 (my other workstation). These machines WILL be here for a few years more (unless something very bad happens in the room where they are).
I'm sure lots of people think the same way I do, and wouldn't appreciate having to download hundreds of megabytes of code that will just slow things down to an unuseable state.
Make sure it doesn't get bloated, please.
GPG 0x1B479C78
- Ctrl-J
- Start typing slashdot (sl...)
- Enter
Done in no time.I'd like to see a filter for browsing the internet just as there is a bayesian filter for email. This is different than, say, a school filtering the internet because, in this case, the user is the one who wants the filtering.
On the internet, there is now the equivalent of spam pages. People conspire to increase Google hits or take over expired domains and put up their own sites. There is also the case of "joke" sites that people in forums post links to in order to get innocent forum-users to click on them.
So, the filtering would have the goal of filtering out whatever the user doesn't want. The above paragraph is just an example of what a user might want to filter out while browsing the internet. Basically, I'd like to see an adaptive internet filter in Mozilla/FireFox so that I can avoid viewing unwanted pages.
GNOME's default browser has this.
Something I would like to see is if I go to news.google.com and I start reading about the current election. I would be able to to have a side bar with a list of top revelent web site associated with keywords in the article, or in meta tags. When I click on links, it compares the keywords with the original article and searches again. There should be a way to rate the website for what you feel the relevance is, allowing it to create a better search. Eventually, I should have a list of websites that are very very relevant to my subject.
This feature could use google or some other searching process, but I think it would be very useful for people who waste most of their time googling for information and trying to figure out the best keywords for a subject. These relevant searches would also be nice to save as a search bookmark.
A more advanced find utility would be: When you do a search, you should be able to specify Boolean or other type of parameters to your query, e.g. find linux or xnix If it can't find a result or has found a result but not useful to user, then give the user the option to search all other links within the same page staying within it's domain name.
I have no idea if this fits into the machine learning category, but it's something I've noticed and wanted for awhile. Sometimes when I am reading long texts online I can't read the thing in one sitting, so it'd be nice if I could set a 'bookmark' within the page so that once I revisited the site it would scroll down to where I was previously. To be more machine learning, the browser could remember where I was on the site and bring me to it automatically. This would be a nice feature.
Linktoolbar provides buttons for these at the bottom of the browser.
LinkIt automatically derives prev/next links from a page's context.
Combined, they make browsing things like webcomics a snap: no more searching for the arbitrarily styled next link.
My only wish is for it to be customizable, so I can teach it to recognize non-english "Previous" and "Next" instances.
I wish that I could, even manually, block javascript, flash, animated images, etc. on a site-by-site basis. So you surf to a page and it applies your default preferences (all the crap off for me). Then you can turn it on as desired. Remember the settings for this site and use them when you come back. You could apply machine learning by analyzing the site like dimensions of images, size of flash on a page (in a little box = ad probably) etc. and you could actually apply filtering dynamically instead of the static default set if the user requests it.
It's exactly what the OP is looking for, and it mostly works very well.
I, too, am hoping for better localization or even user-defined regex matches, but it's quite nice as it is already.
follow me on Twitter: http://twitter.com/moeffju
This looks a lot like a solution in urgent need for a problem. If you don't have any immediate use for Machine Learning in Firefox, don't try to force it in. Also, the prize is kinda ridiculous IMO, but then again, this is Slashdot after all, so it might work.
Before adding unneeded features, I'd suggest all those year-old bugs should be fixed. The Mozilla Team is doing an outstanding job on the suite and Firefox/Thunderbird, but there are some rather annoying bugs in Bugzilla which are gathering dust since years, with the highest of feelings being a mass CC-removal. There's quite a lot of ugly bugs around - they may not occur often, but when they do, they're a nuisance. Then there's also quite some RFEs in Bugzilla, many of which have much support, but are still ignored by the developers. I know - Open Source, do it yourself, etc. Face it, not all people can program, and not all of those who could do have the time.
That said - the Mark Bar (Scroll Up/Down and see where the viewport boundary was before) that was suggested in this thread is a good idea.
Making the URL autocompletion more intelligent is another good idea (which, coincidentally, is in Bugzilla for ages, and constantly being worked on, without ever yielding much), and not too hard to do either.
Searchable bookmarks - building a plaintext search database including page text and link urls would be a very nice idea, and would probably not take much space. Vector search and the usual inverted indexes (if that term translates) etc. Most CS geeks will have built something alike before.
The download manager needs more work. It really should support resume for HTTP and FTP.
Better start pages for Firefox and Thunderbird, instead of the silly and content-free 'Wee! You installed me!' pages. Especially the Thunderbird/Sunbird combination could use an Outlook-style overview of Contacts, Appointments, new e-mail etc.
Speaking of Thunderbird - Virtual Folders/Labels a la Evolution/KMail/GMail are the way to go. Besides, switching to a maildir-like format might be a very good idea. Big mailboxes just don't go well.
Optional per-site download sorting has been mentioned in the comments.
And a way to 'zoom' block-level elements, like a table cell or div, would be very nice too, for sites like Slashdot (I don't want the sidebars, thank you) or other news sites (CNN.. ugh) which seem to find joy in artificially limiting the user's viewport to a narrow stripe.
follow me on Twitter: http://twitter.com/moeffju
5. Allow the user to browse their own hard drive, and categorize content automatically ("this is a document about lambs"
Look, these are all great ideas, but you've just outlined several man-centuries worth of work.
Microsoft has a small army of PhDs, from the best Universities in the world, and several billion dollars in spare change to finance them, yet they're having a helluva time just trying to do something so simple as adding searchable metadata to NTFS. Compare:
Some of the stuff you're talking about is just very, very, very difficult to do, and in the real world of stable, regression-tested, end-user friendly, shippable products, I'd advise you not to hold your breath waiting for this sort of thing to appear anytime soon.PS: Yeah, I know /.-ers will follow up with a bunch of snide remarks about how Microsoft doesn't ship stable products, but again, I'd caution you not to underestimate how truly difficult these things really are.
Comment removed based on user account deletion
I apologize in advance for the length, I keep meaning to write this up and publish it so we've got some prior art when some consciouly devoid person attempts to patent this.
.. you could use a mechanism like AIM (AOL Instant Messenger) or IRC to message the password stores contents between two or more computers. It doesn't need to be realtime, the computers just need to meet online every day or two to sync up details. This will become *phenomially* important as lots of sites on the web transition to a pay to play model.
..
.. Color code it from Blue (Safe) to Red (questionable) and change the color of the window. I can tell my mom and grandmother to check the color BEFORE they put in their passwords (also the distributed password thing makes joejob'ing harder since users can be encouraged to use MUCH stronger passwords which they won't/don't have to remember)
I don't know if this counts as machine learning, but here are two plug-ins i've toyed with writing which I believe would change how people use the Internet:
IDEA 1: Distributed password synchronization: Rationalization - People often pick BAD passwords and use them at hundreds of sites - because when they goto a site from another computer it doesn't know my password so i'll have to guess it. (Just remembering the login names required for each site is a pain in the ass)
Don't give me a whole bunch of rants about how it'd be insecure or whatnot. It can be designed securely, there are LOTS of ways to do it - and frankly it'd more secure than me using the same password for Paypal and for "joejob.cz" which both happen to require email address as a login.
------
IDEA 2: The second requires a bit more coordination but would be PHENOMINALLY USEFUL to the Internet community and catapult FireFox into the most secure browser available. OR translated: I'd install it on my parents computer, and tell them never to surf the Internet using IE because it's not secure [as would all the other readers here]. (Right now FireFox offers no substantial security advantage over IE, it's just less exploited, and by choosing firefox certain sites just don't work - so i'd never tell my parents to use it)
Use the existing spam infrastructure to blacklist phishing sites - use metrics for past traffic, current traffic, perhaps is the site listed in google and arrive at a fraud score for the page, let users flag it as a Phishing site if it appears to be one. All this arrives at the fraud score (e.g. PageRank)
It doesn't have to be perfect, but you COULD say that a reputable company would have a "4 or 5" and a new company would have a "3" and a fake site would have a "0"
This also brings up some interesting challenges/opportunities for the folks at Google - let me explain. If you use Page Rank as one of the metrics and when the PR is zero you can't get above a 1 or 2 - then it means every phishing site will need to be registered in google. The google toolbar already checks PR when I first go to a site - so no biggie, so does Alexa, and all the others.
BUT what is key is the folks at Google are *really good* at recognizing duplicate content under multiple domains. Phising sites are built to look like their legitimate counterparts. By keying off certain big sites (eBay, Paypal, CitiBank, etc.) and letting google vet a site 30 days BEFORE the public it gives Google the opportunity to potentially alert law enforcement BEFORE it went into the index, if it is indeed a phishing site then law enforcement could either have the site shutdown OR let it go into the index and be sitting there waiting to nab whoever tries to get the passwords (remember, whoever is CHECKING for the phishing site is the "phisherman".
It'd be giving the good guys a head start in the cloak and dagger game. To be fair the score should be based on the composite of MULTIPLE sites and SPAM filters, thereby not giving preference to Google - it could be applied to any index.
Okay this is more of an infr
Well it's good that you asked, I just recently got this idea about an intelligent wizard type of thing that will popup and give me good ideas like what should I do etc. I also sketched the looks: it could be a talking paperclip, that would be excellent!
/.tters typoing all the time.
Another idea was that there could a feature to automagically capitalize and correct everything you write. It would be excellent for all
Iam glad someone is asking for ideas .
Iam on a dial up internet,and sometimes when i open a heavy page,and some other pages in the browser,what happens is all of them open slowly.
So i would like to see atleast one proper page,whle others are being opened.
If i pres the stop button for some page,then what has been loaded (like pictures) etc,goes waste,and they have to be donloaded again,when refreshed.
Instead We could have PAUSE button in the browser to pause the loading of page and resume later
Why does yahoo do this
For each site in my history, create a profile of the types of pages on the site (esp. looking for article type pages). For each page type, try and predict which link is the "NEXT" link. Highlight those links, and provide me with a keyboard shortcut to get there. So when I am reading NYT (etc) articles, all I do is hit CTRL-SPACE or some such in order to continue to the next page. The system should learn from my behavior, and should be directly teachable ("this is the NEXT link") as well. The criteria for determining the next link on a new site should be informed by the evolving criteria from existing sites.
Now that would be useful.
it's your form of a rock that really keeps them away!
suggestions of new features:
New buttons on toolbar:
Turn off your ugly patterned background so I can read the text,
asshole webdesigner, button.
Turn off animated gifs, blinking, anything that moves, cretin web idiot.
Make the main text appear same size I specify as I go from site to site,
rather than make me make 'text bigger/smaller' as I go site to site.
Offer an html editing window to make quick edits on the current page,
and then rerender.
Store that template and apply to every page from given site.
Eg, edit out all the junk around the column of text in the
center I'm trying to read a news site; every page is made with a
templete. So just want to modify the template by creating a local modification.
Show me page as text, skip your stupid markup, a la lynx/links, button.
Button to turn off downloading images
Store all urls in all pages as I go, remove dups.
then let me organize this cache. (Similar to bookmark problems)
Forget the machine intelligence , how about a little intelligence?
As neat as machine learning would be, I think Firefox is just about the perfect browser now. It is precisely Firefox's small footprint and do-one-thing-and-do-it-well approach that people love about it.
I'd like to see a newsfeed aggregator built into it, but I'm not going to ask because all I have to do is download Sage. I'd also like to see the Mozilla-based browsers move away from the bookmarks.html format to something XML-based, like XBEL, which Epiphany and Galeon use to great effect. I especially like Epiphany's categorisation of bookmarks, which allow one bookmark to live in two bookmark categories.
Here's a few I wouldn't mind seeing:
1) Spellcheck on right click for text areas (/. needs this badly!).
2) Search "history", mentioned elsewhere.
3) Web session recorder / playback, like a macro recorder if you like. This would be real useful for web developers in stress testing apps.
4) Validate HTML for rendered pages, like Opera offers on right click.
Good people do not need laws to tell them to act responsibly, while bad people will find a way around the laws-Plato
LinkToolBar: LinkIt:
Man, if any PhD school will take me, I hope I can say "Fuck it, I ain't writing no goddamned thesis" and get my story accepted: "Please, write my thesis for me! I'll give you a cookie! Yay, slashdot! Down w/ Microsoft!"
[o]_O
...you can sensibly implement ALL the good ideas in people's answers, and let them pick which ones they care for. You don't have to trade off "how useful will this be" versus "how badly will this bloat the browser".
Allow me to selectively clear cache so that i can get rid of those pr0n sites and still have working autocompletion.Roght now to hide your, uh entertainment website,you have to clear the entire history/cache.
Wanted : A Signature.
would be a nice thing to search for the url in the dmoz database, place the bookmark in a folder in relation to it's location online, and maybe even grab the description and throw that in there too. Instant personal directory of your own sites.
Actually, the only thing Google outlawed with their new policy was the exchange of money in return for an account. Sexual favors, anyone?
- wallet.SchemaValueFileName
- signon.SignonFileName
Mark the names you see for the files, and copy those lines into your target profile prefs.js (or about.config)Seriously, a profile synchronizer or the ability to sync your passwords/form data between zilla installs on different machines would totally rule.
I would like Firefox to have syntax highlighting of CSS and maybe javascript in the source view...
I dont know if this would make it bigger and slower... Maybe this is something that should be made an extension for?
Not sure if this idea is listed elsewhere, but having the browser learn which links would be most likely to be followed, and load those pages in the background, could often make page loading appear instantaneous. Obvious candidates are the "next" pages of a multi-page article.
As said previously, I could not easily determine whether this idea was present within all of these posts. The truly cool AI extension to browsing would be intelligent search, that would use natrual language processing and search for ideas rather than keywords. Of course, this is one of those difficult AI problems, and is probably one for the search engine people, not the browser people.
I can see four areas for machine learning in a Web Browser:
Layout engine - Improve the layout of the page for the current device.
Obviously this is all better done my the author when they design the page, but in the real world and with legacy pages it would be nice for the browser to automated it.
They would require machine learning because there is not 'correct' answer to what content should be kept/dropped to fit a small screen, it would depend on the users preferences, eyesight, page design, which parts of the page where new since the last visit etc.
Content filtering:
Content suggestion:
I understand that firefox has popups suppressed by default which is a good thing. Why on earth, there is no simple way to suppress the (firefox-made) popups trying to ask me if I want to download fancy XY-plungin for the 100th time? If you have to, add an extra "No, but please keep asking me over and over" button, but by default, one "No" per plugin type should be enough not to ever be bothered with shockwave et al. again.
Since I use Firefox, Thunderbird, and Mozilla 1.71, I have at least one question before I comment further. What, exactly, do YOU mean by machine learning? Note, I like Firefox/Thunderbird/Mozilla 1.71 and installed the recent patch, too. Primarily, MSIE is retained only for updating and installing the many, many MS patches, as I am forced to stay with Win2000pro until such time as this old retired dinosaur can affort to buy a new machine for Linux. My default browser is Mozilla 1.7, default e-mail Client is Thunderbird. Firefox "lost" my bookmarks last time when I upgraded from 0.7 to 0.9, so will wait for the 2.0 to come out before changing defaults.
Otherwise, the only thing Mozilla needs is a much easier way to strip the "bad" code so the html is more W3C Accessibility complient, such as an icon or command, perhaps in "edit" or "view" labled "Accessibility".
Retired dinosaur, simple user, volunteer, guinea pig
Not related to machine learning, but something that bugs me often is that when you close a tab you are returned to the tab before the one that you closed... It would be absolutely fantastic if FireFox would return you to the last most recently used tab.
This is implemented in in the form of an extension by FLST; but it would be a great was built in option.
Video meliora proboque deteriora sequor - Ovidius
Here's a challenge for you: create a Firefox plugin that automagically re-writes ugly old html that breaks when rendered at larger magnifications (fonts and text boxes magnifying at different rates, etc.).
Alternatively, have a plug-in that helps enhance the text-based browsers for people that are blind or have some disability that prevents them from fully enjoying sites that happen to have good content and poor presentation.
If it's really good at nice-ifying cruddy html, then Firefox might become a common tool for web site developers that want to create the best possible site.
"Provided by the management for your protection."
Comment removed based on user account deletion
Most important: roaming. Save all preferences, profiles, bookmarks on a central place. And, since you're usually browsing on the Internet, why not use that?
.Zilla' and my profile is there.
When I'm in a Internet Café in Moscow I want to launch Firefox, click 'sign-in to
As long as I can turn that crapola completely off, I'd still use the browser.
i'm amazed no-one suggested this one. besides, if QT4 final is as fast as promised, it will give a speed boost to the firefox UI under KDE, while giving it all the KDE eyecandy.
Also, by the time firefox 2.0 will be there, KDE users will probably outnumber Apple users, so, if the Firefox team takes so much care about Aqua integration, why not doing the same about KDE integration ?
War doesn't prove who's right, just who's left.
I don't really care about machine learning, but i DO need better filters in thunderbird, with regular expressions support, and the ability to search in all accounts/folders. For me that would enhance thunderbird usability more than anything
Of course it's usefull to spy on the user. I mean that's how marketing works, isn't it. Besides, who in the corporate world wouldn't want to have an army of clients that open their computer and a pop-up window will tell them:"This is where you won't go today". Let's come back to Earth for a moment and think why you, the parent, would want to block your child's access to certain sites/addresses: You aren't a good parent. If you were a good parent you would talk to your son or daughter and the problem would be solved. No, that doesn't stop the child of actualy visiting the site if he wants to, but think about kids and internet and mostly about age groups. Kids below 10 years realy don't know anything about their body and most don't leave the Cartoon Network site. The above 10 years of age group should already know about the birds and the bees. Quote: ".. and provide a better web for children and people living in places where pornograhpy is illegal." Are you working for the US Government maybe ? To end my rant: NO, I DON'T AGREE TO BEING SPYED ON!
Some of this has already been touched on by other posters, but not quite exactly the way I had it in mind, so...
My idea is to have Firefox examine web-pages as they are loaded and decide, based on the user's previous decisions, whether it should disable certain features before displaying it. More specifically:
(Note: I haven't had a chance to play with any recent Mozilla product in any depth so I don't know how much of what I've described is already there. Apologies if I've partially reinvented the wheel.)