Why? I don't want to search for "Dashiel Hammett", I want to get back to that Wikipedia page *I already visited* on Dashiel Hammit. If I'm typing in the location bar, and it's going to autocomplete, I clearly want it to give me results from history.
I don't think he implies at all that Firefox is leading Safari in that case, just that the competition is forcing both to improve, which is a good thing. I think if you asked Gecko or WebKit developers they would admit there is competition happening that is making them work harder.
Not to burst your bubble, but Linux users only account for a tiny percentage of total users anyway, so I don't think it will make much of a difference.
I'm not going to argue aesthetics with you, just wanted to reply to your Google Desktop bit. AFAIK, Google Desktop does index your browser cache, so you can search your browsing history. I tried it out once and just didn't find it all that useful.
Re:How hinged is Firefox development to Gecko?
on
Firefox 3 Beta 5 Released
·
· Score: 2, Informative
I do know that Firefox nightlies DO NOT equal webkit nightlies. Firefox and Gecko are actually devoloped on separate branches and are only merged at intervals. Uh, this is completely wrong. Firefox and Gecko are developed on the same branch. Firefox isn't all of Gecko, but it's the biggest consumer by far, and as such the product cycles are heavily tied together.
The actual appearance got tweaked a bit in beta 5, the fonts are smaller and the colors more muted. As for your second point, are you arguing that because a full-text index of your browsing history is better, that nothing else is an improvement? I think that's bogus. The awesomebar is useful to me every day. Would Google Desktop be more useful? Maybe, but at that point I can just search Google itself and be done with it.
I just right click in the search field on the front page, and use "add a keyword for this search". Then I just type "wiki whatever" in my location bar.
Many of these aren't too bad, aside from being ugly (and reminding you that Firefox devs either don't know or don't care about Macs).
That's a bit harsh, I'd say. I think it's more along the lines of "writing good Mac apps is Hard, and doing so without using the native toolkit directly is doubly so." Firefox has not historically had a lot of love on Mac, but that has changed quite a bit in Firefox 3. Is it perfect? Of course not. Is it a lot better? Yes. There was a lot of work to be done, and unfortunately there was only so much time to do it in the development cycle. Some things like the 10.5-style menus were just too big of a change to take at that point in the release cycle. Keep in mind that by the time 10.5 was released, Firefox 3 was just about in beta, so it became hard to make big changes at that point.
In closing, Mac users are very picky about their UI (and that's ok), but it doesn't mean that Firefox developers don't care about Mac.
Chill, Acid3 is not the be-all end-all of web standards. Why should a release be rushed out just to get a perfect score on a test? I'm sure the next (non-security) release of Firefox after 3 will score higher on Acid3. Will it get 100/100? Who knows. Should developers spend their time on things that will gain points in Acid3, but not really improve any web developer or users's lives?
I've been using it for a while now, it's pretty nice. Encrypts your bookmarks before they leave your computer too, so your private data remains private.
Yeah, turns out spidermonkey (the JS engine) is faster at -Os than -O2. Probably due to code locality. Also, just FYI, if you're building your own Firefox, don't pass compiler options in --enable-optimize. Use CFLAGS/CXXFLAGS, otherwise you'll override our hand-picked per-module optimization flags.
Are you talking about audio coming from plugins? Because it's not the browser in that case, the browser has no control over what the plugin code does with your audio device. Yes, it sucks.
You can't compare sunspider results across systems. Obviously your system is faster than his. You can only compare different browsers/browser versions on the same hardware.
No, we're just profiling on browser startup/shutdown right now. I did do a build profiled on the benchmark, and it was pretty fast, but that's probably overkill. Mostly we just want to hit enough common code paths to make things faster. Turns out sunspider perf correlates pretty nicely to overall JS speed, since the benchmark is made up of real world code that people complained was slow.
Unless you've actually tried to do this, I would hold off on estimating. I mapped my whole town in OSM, driving around by myself taking audio notes while recording GPS tracks, and then doing the actual mapping at home with the notes for reference. Even though it's the town I live in (only for a year now, natch) it still took me probably 10 surveying trips to cover 90%+ of my town of 9,000 people. Trying to cover all of the roads, and get the names right is not easy. Even with a dedicated driver and surveyor, it would still be really hard to do. Of course, in the OSM model, labor is free, so you can just throw more people or more trips at the problem.:)
Yeah, except most people *want* to use Google, which is why it wound up as the default in the first place. The money came later. It's nice that it now pays good money, but it started out as the default because it's just the most useful tool. Maybe we can have this discussion again when there's a more useful search engine out there, when it's actually a concern.
Slashdot Mirror
Why? I don't want to search for "Dashiel Hammett", I want to get back to that Wikipedia page *I already visited* on Dashiel Hammit. If I'm typing in the location bar, and it's going to autocomplete, I clearly want it to give me results from history.
The iPhone 3G is quite usable for normal web browsing. Maybe it's purely connection speed?
I don't think he implies at all that Firefox is leading Safari in that case, just that the competition is forcing both to improve, which is a good thing. I think if you asked Gecko or WebKit developers they would admit there is competition happening that is making them work harder.
Not to burst your bubble, but Linux users only account for a tiny percentage of total users anyway, so I don't think it will make much of a difference.
I'm not going to argue aesthetics with you, just wanted to reply to your Google Desktop bit. AFAIK, Google Desktop does index your browser cache, so you can search your browsing history. I tried it out once and just didn't find it all that useful.
The actual appearance got tweaked a bit in beta 5, the fonts are smaller and the colors more muted. As for your second point, are you arguing that because a full-text index of your browsing history is better, that nothing else is an improvement? I think that's bogus. The awesomebar is useful to me every day. Would Google Desktop be more useful? Maybe, but at that point I can just search Google itself and be done with it.
I just right click in the search field on the front page, and use "add a keyword for this search". Then I just type "wiki whatever" in my location bar.
It's just an informal name, not the actual name for the location bar. Are we not allowed to nickname things?
That's a bit harsh, I'd say. I think it's more along the lines of "writing good Mac apps is Hard, and doing so without using the native toolkit directly is doubly so." Firefox has not historically had a lot of love on Mac, but that has changed quite a bit in Firefox 3. Is it perfect? Of course not. Is it a lot better? Yes. There was a lot of work to be done, and unfortunately there was only so much time to do it in the development cycle. Some things like the 10.5-style menus were just too big of a change to take at that point in the release cycle. Keep in mind that by the time 10.5 was released, Firefox 3 was just about in beta, so it became hard to make big changes at that point.
In closing, Mac users are very picky about their UI (and that's ok), but it doesn't mean that Firefox developers don't care about Mac.
Chill, Acid3 is not the be-all end-all of web standards. Why should a release be rushed out just to get a perfect score on a test? I'm sure the next (non-security) release of Firefox after 3 will score higher on Acid3. Will it get 100/100? Who knows. Should developers spend their time on things that will gain points in Acid3, but not really improve any web developer or users's lives?
Or, you know, maybe education doesn't work.
And yes, it is faster and has less crashes and uses less memory. Did you think that was a binary choice?
Give Weave a shot: http://labs.mozilla.com/2007/12/introducing-weave/
I've been using it for a while now, it's pretty nice. Encrypts your bookmarks before they leave your computer too, so your private data remains private.
I don't think we turned it on on OSX (I think Michael was mistaken), but it's on for WIndows and Linux.
Yep, we actually contracted Jason Evans to help us massage JEmalloc into the shape that we needed. He's a nice guy.
We got full page zoom in Firefox 3.
Firefox: fulfilling your fantasies.
Yeah, turns out spidermonkey (the JS engine) is faster at -Os than -O2. Probably due to code locality. Also, just FYI, if you're building your own Firefox, don't pass compiler options in --enable-optimize. Use CFLAGS/CXXFLAGS, otherwise you'll override our hand-picked per-module optimization flags.
Are you talking about audio coming from plugins? Because it's not the browser in that case, the browser has no control over what the plugin code does with your audio device. Yes, it sucks.
You can't compare sunspider results across systems. Obviously your system is faster than his. You can only compare different browsers/browser versions on the same hardware.
No, we're just profiling on browser startup/shutdown right now. I did do a build profiled on the benchmark, and it was pretty fast, but that's probably overkill. Mostly we just want to hit enough common code paths to make things faster. Turns out sunspider perf correlates pretty nicely to overall JS speed, since the benchmark is made up of real world code that people complained was slow.
Unless you've actually tried to do this, I would hold off on estimating. I mapped my whole town in OSM, driving around by myself taking audio notes while recording GPS tracks, and then doing the actual mapping at home with the notes for reference. Even though it's the town I live in (only for a year now, natch) it still took me probably 10 surveying trips to cover 90%+ of my town of 9,000 people. Trying to cover all of the roads, and get the names right is not easy. Even with a dedicated driver and surveyor, it would still be really hard to do. Of course, in the OSM model, labor is free, so you can just throw more people or more trips at the problem. :)
Yeah, except most people *want* to use Google, which is why it wound up as the default in the first place. The money came later. It's nice that it now pays good money, but it started out as the default because it's just the most useful tool. Maybe we can have this discussion again when there's a more useful search engine out there, when it's actually a concern.
Hint: you can't fix crash bugs you don't know about (or can't reproduce)!
Why doesn't that one guy just use Lynx?
Let me know when you've got that regression test suite done that tests the entire internet. I'm sure we can get it checked into CVS. Thanks!