I've found Galeon to be really fast on my computer (Pentium 200), but it still requires a lot of RAM (30 megs).
Galeon uses the Mozilla HTML engine, and it uses Gtk. And it seems really fast (faster than Netscape 4, and about as fast as IE 4 when I boot to Windows). So I have to conclude that Mozilla's HTML rendering is fast, and Gtk is fast. So why is Mozilla so slow? The main difference seems to be the XUL/Javascript based UI in Mozilla, compared to the C based UI in Galeon.
Be sure to try out Galeon and compare it to Mozilla.
Portals are really useful for stuff you don't want to have to search for every hour. Weather. Stocks. News. Comics. TV listings. The latest kernel version.
It's about convenience. Yes, I *could* go visit a news site, and then a comics site, and then a weather site, and then a finance site, and then a TV site, but it's really handy to have it all in one place.
That's the same reason I have my Slashdot settings to include links to BBC News, Linux.com Tuneup, and Technocrat. It saves time. I can scan all this information on one page without visiting each page individually.
The search engine doesn't know that YOUR browser hit a 404. It's not visiting the page -- you are. It could find out if you let them install some sort of spy software on your system that lets them track you, but normally, the search engine doesn't know what you clicked on and whether you got a 404.
Some people might want to make the user think they're still on their site, with site navigation, banner ads, consistent look & feel, etc.
If you're on RedHat.com and they just redirected to Google, are you going to still think you're "on" Redhat.com? No -- you're going to see and think you're on Google. But RedHat might want to keep their users on their own site instead of sending them to Google. So they'd want a partnership.
"I don't know of anyone doing a full text search on the whole web. "
I should clarify, I think the search engines are all looking at the full text of the page, but the search is based on words, not arbitrary series of characters (or even better, regex!!). They're indexing the web -- essentially creating a big list of URLs for each word -- and not grepping all the docs every time you do a search.
"Sadly it is still impossible to search an expression containing a minus"
Yeah, their analysis uses words, not punctuation, so you can't really use Google to search for minus signs or C#. But it seems like the other search engines do the same thing... I don't know of anyone doing a full text search on the whole web.
Can you dig up any reference to their "no ads" quote? It'd be useful to pull out an article from a few years ago and show the world what exactly the Google folk said and whether they lied.
The only thing I vaguely remember is that they didn't want to have banner ads, but I don't remember anything about no advertising whatsoever, and I don't know where I read this.
Google has had JavaScript on their home page since 1998. It moves the focus to the search box so that you can get searches done faster (i.e., without having to click on the box).
The ads only show up if the ads is judged to be somewhat relevant to your search (and therefore potentially useful to the user). They don't (currently) run random irrelevant ads. You may not see any ads if the things being advertised aren't related to what you're searching for.
Google.stanford.edu has been around for longer than the "Google" company. Back then Google was in alpha. So you could've been using it for 3, maybe 4 years..
I do notice an improvement in stop lights. Most every stop light I visit has a sensor that makes it turn green if I'm sitting there and nobody else is coming. I didn't see that ten years ago.
Refrigerators and washing machines seem more efficient too, but I can't measure them as easily.
But then, my TV software can crash, making even the power button useless. I didn't see that ten years ago either. Hmm. Maybe it balances out.:-)
Search engines aren't perfect. Some people don't *want* to see certain topics when they're searching for something totally unrelated. So NetNanny, SurfWatch, etc., are somewhat useful for that. But they aren't perfect either, and they tend to overfilter than underfilter. I think this is from market forces -- a user is going to notice if porn slips through, but isn't going to notice if the Smithsonian is missing -- so the products that succeed will tend to overfilter.:(
But what's important is that filtering has to be a choice at the individual level, not something imposed by a government. I don't think Google's saying that it's going to force all its users to use SurfWatch. (That'd be CRAZY!) It's just giving them an option.
Whether parents should be able to make choices for their children is an orthogonal issue. Should parents get to pick schools for their kids? Clothes? Religion? Family filter? Alcohol? R-rated movies? I think it's worthy of debate, but I don't think we should limit it to Internet content filter software.
Google does have advertising, and has had it for some time now. The main thing is, the way things are set up right now, Google only shows an ad if it might be relevant. Google isn't showing random ads ("run of site" ads) on every single page. So if the things you search for don't happen to be interesting to an advertiser, Google isn't showing an ad to you. Some topics just aren't commercial enough. For example, if all you search for is pages about ancient Mayan rituals, you may not have ever seen an ad on Google. On the other hand, if you're searching for airline tickets or Mother's day flowers, you might be seeing lots of ads.
The article you're thinking of is in Scientific American. It's primarily about CLEVER, a project at IBM that tries to identify "hubs" (pages with links to good information) and "authorities" (pages with good information), but it also mentions Google. It'd be interesting to see CLEVER available on the web.. the last time I looked, it was only for internal IBM use.
It's close, but not exactly doing a phrase search first. It prefers phrase matches but it also considers the ranking of the page, so a really important page with an AND match might show up before a medium importance page with a phrase match.
Slashdot Mirror
The Linux community is simply trying to make software more like the Windows equivalents, and part of that is only running on one platform ;)
And I know exactly which contest you're talking about! You know, the contest was run by a guy (Dave Taylor) went on to work on DOOM ..
.. ah, memories!
I have his business card, and got him to sign my Ultima 7 manual
(I don't remember who won the contest that year though)
I've found Galeon to be really fast on my computer (Pentium 200), but it still requires a lot of RAM (30 megs).
Galeon uses the Mozilla HTML engine, and it uses Gtk. And it seems really fast (faster than Netscape 4, and about as fast as IE 4 when I boot to Windows). So I have to conclude that Mozilla's HTML rendering is fast, and Gtk is fast. So why is Mozilla so slow? The main difference seems to be the XUL/Javascript based UI in Mozilla, compared to the C based UI in Galeon.
Be sure to try out Galeon and compare it to Mozilla.
.. in a game I wrote:
s .h tml
http://www-cs-students.stanford.edu/~amitp/game
It really makes life nicer.
Portals are really useful for stuff you don't want to have to search for every hour. Weather. Stocks. News. Comics. TV listings. The latest kernel version.
It's about convenience. Yes, I *could* go visit a news site, and then a comics site, and then a weather site, and then a finance site, and then a TV site, but it's really handy to have it all in one place.
That's the same reason I have my Slashdot settings to include links to BBC News, Linux.com Tuneup, and Technocrat. It saves time. I can scan all this information on one page without visiting each page individually.
Check out:
l
http://www.google.com/mozilla/google-search.htm
I'm using it on Linux with Mozilla and Galeon. (Galeon supports the %s thingie described at the end.) It's not as full featured as the IE toolbar.
Let the "Netscape 6" flame wars begin!
I've put up some design notes about SRE here:
c le s/SRE-Design.html
:-)
http://www-cs-students.stanford.edu/~amitp/Arti
I hope someone finds it useful when cloning the game.
The search engine doesn't know that YOUR browser hit a 404. It's not visiting the page -- you are. It could find out if you let them install some sort of spy software on your system that lets them track you, but normally, the search engine doesn't know what you clicked on and whether you got a 404.
I only use it when I'm pretty sure I'll get what I want -- and that's usually for company names.
Google's + needs to go inside the "" (weird, but they want to know exactly which words you want + on, not just all of them), so do "spike +and lee".
Some people might want to make the user think they're still on their site, with site navigation, banner ads, consistent look & feel, etc.
If you're on RedHat.com and they just redirected to Google, are you going to still think you're "on" Redhat.com? No -- you're going to see and think you're on Google. But RedHat might want to keep their users on their own site instead of sending them to Google. So they'd want a partnership.
"I don't know of anyone doing a full text search on the whole web. "
I should clarify, I think the search engines are all looking at the full text of the page, but the search is based on words, not arbitrary series of characters (or even better, regex!!). They're indexing the web -- essentially creating a big list of URLs for each word -- and not grepping all the docs every time you do a search.
"Sadly it is still impossible to search an expression containing a minus"
Yeah, their analysis uses words, not punctuation, so you can't really use Google to search for minus signs or C#. But it seems like the other search engines do the same thing... I don't know of anyone doing a full text search on the whole web.
Can you dig up any reference to their "no ads" quote? It'd be useful to pull out an article from a few years ago and show the world what exactly the Google folk said and whether they lied.
The only thing I vaguely remember is that they didn't want to have banner ads, but I don't remember anything about no advertising whatsoever, and I don't know where I read this.
Google has had JavaScript on their home page since 1998. It moves the focus to the search box so that you can get searches done faster (i.e., without having to click on the box).
"dressed in European grey"
:)
brings up some Ultravox pages on Google, too.
AND AN AD! (for a music site)
"but where are the ads on google? "
The ads only show up if the ads is judged to be somewhat relevant to your search (and therefore potentially useful to the user). They don't (currently) run random irrelevant ads. You may not see any ads if the things being advertised aren't related to what you're searching for.
Google.stanford.edu has been around for longer than the "Google" company. Back then Google was in alpha. So you could've been using it for 3, maybe 4 years ..
Google *does* have a search within search! It's at the bottom of the page, to the right of the serach box. Just click on "Search within results".
I do notice an improvement in stop lights. Most every stop light I visit has a sensor that makes it turn green if I'm sitting there and nobody else is coming. I didn't see that ten years ago.
:-)
Refrigerators and washing machines seem more efficient too, but I can't measure them as easily.
But then, my TV software can crash, making even the power button useless. I didn't see that ten years ago either. Hmm. Maybe it balances out.
I think OS/2 sponsored the Fiesta Bowl, not the superbowl.
Search engines aren't perfect. Some people don't *want* to see certain topics when they're searching for something totally unrelated. So NetNanny, SurfWatch, etc., are somewhat useful for that. But they aren't perfect either, and they tend to overfilter than underfilter. I think this is from market forces -- a user is going to notice if porn slips through, but isn't going to notice if the Smithsonian is missing -- so the products that succeed will tend to overfilter. :(
But what's important is that filtering has to be a choice at the individual level, not something imposed by a government. I don't think Google's saying that it's going to force all its users to use SurfWatch. (That'd be CRAZY!) It's just giving them an option.
Whether parents should be able to make choices for their children is an orthogonal issue. Should parents get to pick schools for their kids? Clothes? Religion? Family filter? Alcohol? R-rated movies? I think it's worthy of debate, but I don't think we should limit it to Internet content filter software.
Google does have advertising, and has had it for some time now. The main thing is, the way things are set up right now, Google only shows an ad if it might be relevant. Google isn't showing random ads ("run of site" ads) on every single page. So if the things you search for don't happen to be interesting to an advertiser, Google isn't showing an ad to you. Some topics just aren't commercial enough. For example, if all you search for is pages about ancient Mayan rituals, you may not have ever seen an ad on Google. On the other hand, if you're searching for airline tickets or Mother's day flowers, you might be seeing lots of ads.
The article you're thinking of is in Scientific American. It's primarily about CLEVER, a project at IBM that tries to identify "hubs" (pages with links to good information) and "authorities" (pages with good information), but it also mentions Google. It'd be interesting to see CLEVER available on the web.. the last time I looked, it was only for internal IBM use.
- Amit
It's close, but not exactly doing a phrase search first. It prefers phrase matches but it also considers the ranking of the page, so a really important page with an AND match might show up before a medium importance page with a phrase match.