It's true that duplicate count can be useful in determining the relative visibility of bugs, but duplicate count would be just as useful with the rate of duplicates cut to a quarter of the present rate. Fifty dups means at least fifty searches for triagers (if they recognize the bug as a duplicate immediately), fifty times a QA contact has to read the dup to verify that it's a dup, and fifty bugmails to each person involved in the original bug. For some layout bugs, the QA contact might read through every dup of a layout bug after the bug is marked as fixed to make sure every dup is also fixed. Each duplicate creates work for multiple people who could be fixing bugs, testing fixes, and confirming non-duplicate bug reports.
Votes cause no spam and are easier to search for and count. Please use votes instead of reporting duplicates. On a bug with fewer than 10 comments, an "I see this too" comment might even be tolerated.
Humble request: can you make the vertical scroll bar BIGGER?
Huh? Mozilla's vertical scrollbar is infinitely large (horizontally) because it sits at the right edge of the screen. Netscape 4, Internet Explorer, and Opera (all on Windows) leave space between the scrollbar and the edge of the screen, making the scrollbar harder to click and very hard to hold onto without keeping the mouse button down or looking at it.
The value of 1 came in a patch by the PornBrowser team, did it?
Many free web hosts (not just porn hosts) require each image request to include the correct referrer. I wouldn't expect the "images only" setting to work on porn sites, because the browser doesn't know whether a link (such as a thumbnail) links to an image or an html page.
There's probably a pref in your operating system to disable smoothing or disable it for small font sizes. It wouldn't make sense for an individual app to control that.
Bookmarklets are bookmarks containing javascript code. Instead of taking you to another page, bookmarklets do things with/to the current page. Here are some bookmarklets for Mozilla that I have added since Mozilla 1.0:
Transfer cookies: creates a bookmarklet you can use to move cookies from one browser to another.
Number rows in each HTML table
Several of these bookmarklets also work in IE 5.5, to the extent that IE supports DOM Level 2 and doesn't make me go too far out of my way to accommodate its quirks.
Mozilla.org gets a lot of duplicate bug reports: 40-50% of a sample of bugs from April 2002 are dups. If you know how to search Bugzilla, you can get that down to 15-20%. (Knowing some jargon helps too, of course.) Unfortunately, the most widely advertised Bugzilla search tool, the query builder, is so complicated that many testers give up before finding their bug and report a duplicate.
There's a well-hidden search box on the Bugzilla front page that works a lot like Google. You can almost use it like Google, but there are several differences you should be aware of:
Each word is matched as a substring of the summary (and several other fields). A search for 'auto compl' will match "auto-complete", "auto complete", and "autocompletion".
Like in Google, you can use | to create disjunctions. For example, a search for 'address|location|url bar|field focus' will match "focus does not move when clicking outside of location bar". While "or" is usually unnecessary for general web searches, it is indispensible when searching for a specific bug report.
By default, Bugzilla only searches for open bugs. If you're looking for a bug that has been reported several times, it may help to include duplicates in the search. One way to do this is to prefix the search with 'ALL ' in all caps. For example, 'ALL rename exe' will lead you to an often-reported bug (120327) that I should be helping bz to fix instead of posting this comment, while 'rename exe' will not find anything.
If you know that the bug you're searching for is visible and popular, try adding 'votes:2' to your search. For example, 'ALL votes:2 context menu back' will find the newest flamewar-bug about the back command in the context menu among the 42 bugs that match 'ALL context menu back'. Searches that use votes:2 are several times faster than searches that include all bugs because bugzilla can start the search with an integer comparison.
The search includes several fields, not just the bug summary (title). For example, in a search for 'mail compos focus', the word "mail" can appear in either the product name (MailNews) or the bug summary, and "compos" can appear either in a component name (Composition) or in the summary (compose, composing, etc). To restrict a search term to the summary, use '+term'.
Other useful tools for avoiding reporting duplicates include the frequently reported bugs list and #mozillazine on irc.mozilla.org. If you find yourself working in Bugzilla a lot, you can use the collect buglinks bookmarklet to get a list of bugs mentioned in a given bug report, which is useful because many bug reports include links to related bugs.
ideally, the build will only break if two changes on the same night contradict each other
Another reason a change might break the build is if the software builds on multiple platforms. I think this is one reason Mozilla has dozens of computers continuously building it rather than one nightly build that may or may not compile the first time. This doesn't catch all checkins that make Mozilla unusable (we have a "smoketest" team that make sure each nightly works well enough to test it), but it does give everyone immediate feedback if a checkin stops the build from compiling or makes it fail some automatic tests.
Of course, it helps that compiling Mozilla takes less than an hour.
IE and Mozilla render tables incrementally -- they can show you part of a table before the entire table is loaded. That is a good thing because it effectively makes pages load more quickly. It would look strange if IE/Moz were to toss an already-displayed table because it reached the end of the document before seeing a tag closing the table. That would be kind of like displaying 90% of a porn image, one row of pixels at a time, and then suddenly replacing it with the text "This image cannot be displayed because it contains errors" rather than leaving it 90% displayed. (Mozilla currently does that, by the way.)
IE has lots of hacks that are only there to "help" webmasters who make coding mistakes and typos. Accepting tables without closing tags is not one of them.
Instead of boosting the pigeonrank of any page that doesn't validate, I think it makes more sense to list a bunch of competing sites and say how well each one validates. Here are some sites that I think do it right:
Search Engine HTML Validation Results - list of major search engines and web directories, showing how many HTML errors the W3C validator finds on both the front page and the results of a search for "mp3 rippers". The table was posted on June 29. One site (dmoz.org) made its front page validate by July 3.
Financial Institutions and Mozilla Operability - list of many banks, saying how well Mozilla works with each bank. By concentrating on the practical "Can I use this site with Mozilla" rather than the ideal "Does this site validate", this site is more useful to users trying to decide which competitor to choose. It is therefore more powerful for getting the sites to fix themselves.
and for contrast:
Free Web Hosts - list of free web hosts and whether the host makes an uploaded web page stop validating. Doesn't have enough data for a table yet, so there's not much pressure on hosts to change. Uploads XHTML test pages rather than HTML 4.01 test pages, which seems like an odd choice to me.
The list you linked to looks like a list of the most-visited sites on the web. Mozilla developers and evangelizers can use that kind of list in order to determine how important/urgent a given bug is. A list of the "most troublesome" sites would have msn.com and some banks at the top, not yahoo and x10.
The only thing you might have to redo in order to support a new browser is the web page coding. You also have to advertise, keep the prices for your products low, do graphic design for the web site, take phone calls from frustrated customers, design the backend for credit card processing, write and maintain a privacy policy, etc. I'm not a business guy, but my intuition says that web page coding does not account 30% of the money you use to lure customers.
There is an HTML tag that defines what is the "next" page in a sequence, . Its proponents like to point out that it has been part of specs since HTML 2. Unfortunately, few sites use the tag. When the tag is not present, a browser could guess that a link with the text "next" was a link to the next page in a sequence, or it could guess that incrementing the last number in the url by 1 would go to the "next" page. (The increment bookmarklet uses the latter strategy and is very useful for surfing porn.)
My opinion is that Mozilla should guess based on the number in the URL (32627) when the link tag is not present. Guessing based on link text may not internationalize well, and many sites allow users to post links with text of their choice. I have suggested the keyboard shortcut Alt++ for next (102909 #c17). Pressing space (usually pgdn) at the bottom of a page might also go to the next page (59118).
Mycroft now has a Mozilla search plugin for the version of Google with keyboard shortcuts. And to be on topic, Mycroft also has a search plugin for Teoma.
If you're using a new version of Mozilla (newer than 1.1alpha), you can press Ctrl+Shift+F to go to your chosen search engine. If you install and select the Google Keys search plugin it's easy to do an entire search using only the keyboard. You can press Ctrl+Shift+F, type your query, press enter, and then press '2' to go to the second hit or 'KLK' to go to the second adwords ad.
Teoma now has a toolbar for IE/Win, similar to the Google Toolbar. Here is a quick comparison of the toolbars:
The Teoma toolbar is missing a lot of features that the Google toolbar has.
Teoma's toolbar does not have a "Search Site" button.
Google adds "Google Search" to Internet Explorer's context menu for selected text. Teoma does not. (This feature is built into Mozilla, by the way.)
Shift+Enter doesn't open the search results in a new window like it does in the Google Toolbar.
Alt+Enter doesn't go straight to the first hit like it does in the Google Toolbar (I'm Feeling Lucky).
Alt+D (focus address bar) and Alt+G (focus Google toolbar) do not work while the Teoma toolbar has focus. Furthermore, CLICKING on the Google toolbar does not work when the Teoma toolbar has focus.
It does, however, have Alt+S to focus the search field in the toolbar (Google uses Alt+G). This is good because the first thing you're going to do after focusing the search field is type a search phrase using the keyboard.
The Teoma toolbar makes it easy to add and remove toolbar buttons. I like that.
The "Email this page to a friend" feature is useless. It creates a message that advertises the Teoma search bar and almost hides the information you were trying to send. Unlike the "Send Page" feature in IE and Mozilla, Teoma only sends a link to the page (like IE's and Mozilla's "Send Link"), but it doesn't tell you this.
Problems with BOTH toolbars:
Neither Google nor Teoma makes their toolbar Search button search for selected text. (Same with the search buttons built into IE and Mozilla.) You can drag the selection into the search textbox, but that's awkward.
Clicking search brings you to a page with a search box but does not focus the search box for you. In contrast, the front page for each search engine focuses the search box automatically, saving you a mouse click.
Pasted Japanese text comes out as a bunch of question marks, and clicking search actually searches for question marks (%3f).
The only version available is for Internet Explorer for Windows.
Now...I have to say that Google doesn't appear to specifically target me with 'ads for Canadians', but there is nothing to stop them from doing so, and no way to tell if that is indeed what they are doing at some point in the future.
Is that a bad thing? Google prides itself on serving relevant ads that users actually want to look at. For some ads, location matters when deciding whether an ad is relevant. When I put up my ads, I had a choice of languages and countries. I chose to advertise to everyone.
Btw, www.google.com probably gives you ads targeted to Canada if it can tell you're from Canada. And it can tell you're from Canada:)
They could have a penguin among the animals waving the American flag around the Google logo. I don't expect Google to do anything, though. It's too bad the Windependence Day people chose July 4, since they get mentally grouped with all the corporations abusing the American flag in ads and product labels since September.
You have it backwards. Google slashdots other sites. Have you ever seen a Geocities hit at the top of a search for song lyrics? Those links never work, because Geocities blocks links to sites that get too many hits.
You might be able to slashdot Google with a real-life event that everyone learns about at the same time, such as an earthquake or a big breaking-news story like 9/11. But even then, I'd expect Google to be able to handle it.
1. Record URLs visited by Google Toolbar users. Google's Privacy Policy for their toolbar seems to allow this.
2. Get Earthlink to record all URLs visited by Earthlink users and put that database in a format that Googlebot understands.
3. Get Dreamhost to let Google index directories even where index.html exists. Let Dreamhost call the feature "[x] Submit all pages on www.squarefree.com to Google once a month" or "[x] Allow Google to find unlinked pages". Make it optional so that rare pages protected by "javascript passwords" can remain unindexed, but enable it by defalt.
4. Give Google users a place to upload their browser history files and say the information will only be used to find unindexed pages.
5. Don't worry about it, since unlinked pages will rank poorly anyway, and are likely to be old versions of pages that were renamed but not deleted.
Slashdot Mirror
It's true that duplicate count can be useful in determining the relative visibility of bugs, but duplicate count would be just as useful with the rate of duplicates cut to a quarter of the present rate. Fifty dups means at least fifty searches for triagers (if they recognize the bug as a duplicate immediately), fifty times a QA contact has to read the dup to verify that it's a dup, and fifty bugmails to each person involved in the original bug. For some layout bugs, the QA contact might read through every dup of a layout bug after the bug is marked as fixed to make sure every dup is also fixed. Each duplicate creates work for multiple people who could be fixing bugs, testing fixes, and confirming non-duplicate bug reports.
Votes cause no spam and are easier to search for and count. Please use votes instead of reporting duplicates. On a bug with fewer than 10 comments, an "I see this too" comment might even be tolerated.
The scrollbar is infinitely wide if you maximize the window or align it with the right edge of the screen.
Humble request: can you make the vertical scroll bar BIGGER?
Huh? Mozilla's vertical scrollbar is infinitely large (horizontally) because it sits at the right edge of the screen. Netscape 4, Internet Explorer, and Opera (all on Windows) leave space between the scrollbar and the edge of the screen, making the scrollbar harder to click and very hard to hold onto without keeping the mouse button down or looking at it.
The value of 1 came in a patch by the PornBrowser team, did it?
Many free web hosts (not just porn hosts) require each image request to include the correct referrer. I wouldn't expect the "images only" setting to work on porn sites, because the browser doesn't know whether a link (such as a thumbnail) links to an image or an html page.
- Jesse, member of the unofficial Mozilla porn-browser team.
There's probably a pref in your operating system to disable smoothing or disable it for small font sizes. It wouldn't make sense for an individual app to control that.
Bookmarklets are bookmarks containing javascript code. Instead of taking you to another page, bookmarklets do things with/to the current page. Here are some bookmarklets for Mozilla that I have added since Mozilla 1.0:
"Fixing" annoying web sites:- Zap event handlers: removes event handlers, including those responsible for blind links and exit pop-up ads.
- Zap embeds: removes java, flash, background music, and iframes from a page.
- Zap colors: makes text black on a white background, and makes links blue and purple.
- Zap: combines "zap embeds", "zap colors", and "zap event handlers".
- Test styles: type in CSS rules to experiment or to create a temporary user style sheet.
Web development:- View Style Sheets
- View Scripts
- View Script Variables
Other:Several of these bookmarklets also work in IE 5.5, to the extent that IE supports DOM Level 2 and doesn't make me go too far out of my way to accommodate its quirks.
Mozilla.org gets a lot of duplicate bug reports: 40-50% of a sample of bugs from April 2002 are dups. If you know how to search Bugzilla, you can get that down to 15-20%. (Knowing some jargon helps too, of course.) Unfortunately, the most widely advertised Bugzilla search tool, the query builder, is so complicated that many testers give up before finding their bug and report a duplicate.
There's a well-hidden search box on the Bugzilla front page that works a lot like Google. You can almost use it like Google, but there are several differences you should be aware of:
Other useful tools for avoiding reporting duplicates include the frequently reported bugs list and #mozillazine on irc.mozilla.org. If you find yourself working in Bugzilla a lot, you can use the collect buglinks bookmarklet to get a list of bugs mentioned in a given bug report, which is useful because many bug reports include links to related bugs.
ideally, the build will only break if two changes on the same night contradict each other
Another reason a change might break the build is if the software builds on multiple platforms. I think this is one reason Mozilla has dozens of computers continuously building it rather than one nightly build that may or may not compile the first time. This doesn't catch all checkins that make Mozilla unusable (we have a "smoketest" team that make sure each nightly works well enough to test it), but it does give everyone immediate feedback if a checkin stops the build from compiling or makes it fail some automatic tests.
Of course, it helps that compiling Mozilla takes less than an hour.
IE and Mozilla render tables incrementally -- they can show you part of a table before the entire table is loaded. That is a good thing because it effectively makes pages load more quickly. It would look strange if IE/Moz were to toss an already-displayed table because it reached the end of the document before seeing a tag closing the table. That would be kind of like displaying 90% of a porn image, one row of pixels at a time, and then suddenly replacing it with the text "This image cannot be displayed because it contains errors" rather than leaving it 90% displayed. (Mozilla currently does that, by the way.)
IE has lots of hacks that are only there to "help" webmasters who make coding mistakes and typos. Accepting tables without closing tags is not one of them.
Instead of boosting the pigeonrank of any page that doesn't validate, I think it makes more sense to list a bunch of competing sites and say how well each one validates. Here are some sites that I think do it right:
Search Engine HTML Validation Results - list of major search engines and web directories, showing how many HTML errors the W3C validator finds on both the front page and the results of a search for "mp3 rippers". The table was posted on June 29. One site (dmoz.org) made its front page validate by July 3.
Financial Institutions and Mozilla Operability - list of many banks, saying how well Mozilla works with each bank. By concentrating on the practical "Can I use this site with Mozilla" rather than the ideal "Does this site validate", this site is more useful to users trying to decide which competitor to choose. It is therefore more powerful for getting the sites to fix themselves.
and for contrast:
Free Web Hosts - list of free web hosts and whether the host makes an uploaded web page stop validating. Doesn't have enough data for a table yet, so there's not much pressure on hosts to change. Uploads XHTML test pages rather than HTML 4.01 test pages, which seems like an odd choice to me.
The list you linked to looks like a list of the most-visited sites on the web. Mozilla developers and evangelizers can use that kind of list in order to determine how important/urgent a given bug is. A list of the "most troublesome" sites would have msn.com and some banks at the top, not yahoo and x10.
The only thing you might have to redo in order to support a new browser is the web page coding. You also have to advertise, keep the prices for your products low, do graphic design for the web site, take phone calls from frustrated customers, design the backend for credit card processing, write and maintain a privacy policy, etc. I'm not a business guy, but my intuition says that web page coding does not account 30% of the money you use to lure customers.
They're in the business of filtering smart, after all.
Spoiler space
How did you get that past Slashdot's useful-comment filter? Impressive.
Enron, Worldcom, Microsoft?
(Discovered by Ksosez on #mozillazine.)
I could say the same thing about Google. A google search for "search" returns Google and Yahoo (which uses Google) at the top.
There is an HTML tag that defines what is the "next" page in a sequence, . Its proponents like to point out that it has been part of specs since HTML 2. Unfortunately, few sites use the tag. When the tag is not present, a browser could guess that a link with the text "next" was a link to the next page in a sequence, or it could guess that incrementing the last number in the url by 1 would go to the "next" page. (The increment bookmarklet uses the latter strategy and is very useful for surfing porn.)
My opinion is that Mozilla should guess based on the number in the URL (32627) when the link tag is not present. Guessing based on link text may not internationalize well, and many sites allow users to post links with text of their choice. I have suggested the keyboard shortcut Alt++ for next (102909 #c17). Pressing space (usually pgdn) at the bottom of a page might also go to the next page (59118).
Mycroft now has a Mozilla search plugin for the version of Google with keyboard shortcuts. And to be on topic, Mycroft also has a search plugin for Teoma.
If you're using a new version of Mozilla (newer than 1.1alpha), you can press Ctrl+Shift+F to go to your chosen search engine. If you install and select the Google Keys search plugin it's easy to do an entire search using only the keyboard. You can press Ctrl+Shift+F, type your query, press enter, and then press '2' to go to the second hit or 'KLK' to go to the second adwords ad.
The Teoma toolbar is missing a lot of features that the Google toolbar has.
- Teoma's toolbar does not have a "Search Site" button.
- Google adds "Google Search" to Internet Explorer's context menu for selected text. Teoma does not. (This feature is built into Mozilla, by the way.)
- Shift+Enter doesn't open the search results in a new window like it does in the Google Toolbar.
- Alt+Enter doesn't go straight to the first hit like it does in the Google Toolbar (I'm Feeling Lucky).
- Alt+D (focus address bar) and Alt+G (focus Google toolbar) do not work while the Teoma toolbar has focus. Furthermore, CLICKING on the Google toolbar does not work when the Teoma toolbar has focus.
It does, however, have Alt+S to focus the search field in the toolbar (Google uses Alt+G). This is good because the first thing you're going to do after focusing the search field is type a search phrase using the keyboard.The Teoma toolbar makes it easy to add and remove toolbar buttons. I like that.
The "Email this page to a friend" feature is useless. It creates a message that advertises the Teoma search bar and almost hides the information you were trying to send. Unlike the "Send Page" feature in IE and Mozilla, Teoma only sends a link to the page (like IE's and Mozilla's "Send Link"), but it doesn't tell you this.
Problems with BOTH toolbars:
Now...I have to say that Google doesn't appear to specifically target me with 'ads for Canadians', but there is nothing to stop them from doing so, and no way to tell if that is indeed what they are doing at some point in the future.
:)
Is that a bad thing? Google prides itself on serving relevant ads that users actually want to look at. For some ads, location matters when deciding whether an ad is relevant. When I put up my ads, I had a choice of languages and countries. I chose to advertise to everyone.
Btw, www.google.com probably gives you ads targeted to Canada if it can tell you're from Canada. And it can tell you're from Canada
Wow, I don't remember that. Was CNN angry?
They could have a penguin among the animals waving the American flag around the Google logo. I don't expect Google to do anything, though. It's too bad the Windependence Day people chose July 4, since they get mentally grouped with all the corporations abusing the American flag in ads and product labels since September.
What don't you like about www.google.ca? It's in English. It's closer to you, and therefore it should be faster.
You have it backwards. Google slashdots other sites. Have you ever seen a Geocities hit at the top of a search for song lyrics? Those links never work, because Geocities blocks links to sites that get too many hits.
You might be able to slashdot Google with a real-life event that everyone learns about at the same time, such as an earthquake or a big breaking-news story like 9/11. But even then, I'd expect Google to be able to handle it.
So how do you find those treasure troves?
1. Record URLs visited by Google Toolbar users. Google's Privacy Policy for their toolbar seems to allow this.
2. Get Earthlink to record all URLs visited by Earthlink users and put that database in a format that Googlebot understands.
3. Get Dreamhost to let Google index directories even where index.html exists. Let Dreamhost call the feature "[x] Submit all pages on www.squarefree.com to Google once a month" or "[x] Allow Google to find unlinked pages". Make it optional so that rare pages protected by "javascript passwords" can remain unindexed, but enable it by defalt.
4. Give Google users a place to upload their browser history files and say the information will only be used to find unindexed pages.
5. Don't worry about it, since unlinked pages will rank poorly anyway, and are likely to be old versions of pages that were renamed but not deleted.