Feature Status Updates
Friday March 2nd, 2001
Gervase Markham has polled various feature owners, and has written up a brief status report on some of mozilla's components. Click the 'Full Article' link to check them out.
On a side note, Gerv is looking for help in making this a weekly or monthly mozillaZine feature. If you are interested, let Gerv know.
Full Article...
thanks for putting this together, of course this means that we'll be expecting an update every week... :)
First, thanks Gerv!
Now this rocks. The status updates at mozilla.org <http://mozilla.org/status/> are a great resource but these overviews really help to clarify what's actually going on. I hope that someone can find the cycles to help make this a regular feature at mozillazine.
--Asa
I agree. This report was amazingly encouraging! It's nice to see that there will be significant improvements in just a little while.
What's the status on SSL? The current implementation is *SLOOOOOOOWWWWWW* for me on all sites I've tried. When I try a site with IE, first without SSL and then with SSL, the slowdown is not even noticeable. When I switch to Mozilla, the SSL version takes about 30 seconds to load while the normal non-encrypted version takes less than 1 second. My computer is a 500Mhz Athlon with 128 megs of ram, Windows 98 and my network connection is a 2Mbps ADSL line so that shouldn't be a problem. With IE, the SSL version of the site loads in less than 1 second so it's not the server.
I get similar results on ALL SSL sites I've tried. Any chance that someone is working on getting this fixed to usable levels? This is a serious hurdle for many people to switch to Mozilla / derivate products since they do their banking, buy stuff etc. online without it taking 50 times longer than it used to.
I have the same problem, Netscape 4.75 works just fine but no version of Mozilla works at all. The downloads are very long if they work at all.
SSL/PSM is an important feature to many end users. It should be addressed ASAP.
First, SSL/PSM work fine for me on my windows 95. No noticeable slowdown.
Second, if you look the PSM projects page (link on the mozilla.org/projects page), you will see that PSM2.0 is on its way, and it is supposed to be much better (in process instead of the current out of process). Some people already got PSM2.0 to compile..
Fabian.
yeah, it's in process but the communication is still with sockets and it's still a proxy.. that's hardly an optimal solution for speed...
I love the status update, i hope that we will see many more in the future. The mozilla.org updates seem to concentrate on the number of bugs fixed and left, whereas this concentrated on features and improvements we can look forward to. I like that!
It seems strange that so many modules are being rewritten and replaced late in beta. Doesn't that create a stability risk?
Five modules are being rewritten and replaced according to this report: image handling, view handling, caching, string classes, and view source.
With the exception of view source, all are components which could have major ripple effects throughout the entire system. Most could also have compatibility effects with respect to a majority of web pages in the world.
Am I really the only one who finds this disturbing?
I for one find the current imagelib and cache suck. In some cases, they're worse than Netscape 4.x. Try visiting a site with lots of animated GIFs, and you'll see how much Mozilla suffers from decoding the GIFs for every animation loop (especially on Linux). Sometimes Mozilla couldn't even cache the images used for javascript "onMouseOver" effects, and it looks plain awkward. Finally, many web developers have complained that "view source" doesn't always give you the source of the page that's displayed in your browser; rather, Mozilla tries to re-get the source from network. This will turn off many developers, and we'll have more and more web sites developed for IE.
Of course, I would like to thank Mozilla developers for recognizing these problems and fixing them. Wish them luck.
At least two of these large scale changes (imglib2 and the new cache) are addressing major performance problems - image loading bottlenecks and the lack of a decent cache. Addressing these issues should reduce the perceived difference between Mozilla's speed in loading and navigating between pages and IE's.
A lot of this work is being done on branches or not enabled, so that rather than landing things quickly and breaking all sorts of things, the individuals involved have time to test their changes thoroughly before landing their branches or enabling the feature.
Ultimately, these are things that have to be done before Mozilla can have a meaningful 1.0. If it takes a little longer to get it right, it's worthwhile (And this is not talking about taking time to perfect minute seldom-used features, this is talking about tuning page load and navigation performance, which is in many cases terrible).
I agr that these issues are significant and need to be addressed. My concern was not that they were the wrong changes to make but that they risk further slipping the delivery schedule by destabilizing the program. If there is a take-away from my message, it is just to underscore the need for caution in making such changes -- which it appears you are already aware of.
I think it's a matter of priority; do you want a fast, stable Mozilla 1.0 or a half-assed Mozilla with all sorts of bottle-necks and bugs that could've been fixed if the schedule was changed?
I prefer the former. I have all the time in the world, as long as 1.0 gets good I'm happy. After all, 1.0 will just be "yet another" nightly with a different package and name AFAIK
So it's not like 1.0 will be much different than any nightlies from those dates.
I think 1.0 won't be "yet another nightly" since it will get a longer branch, so it will have a longer stabilizing time than 0.8, 0.9 and as a result it will be more optimized than a nightly build.
What happenned to the win32 talkback .zip builds? Are the submitted bug reports useless (compared w/bugzilla) so talkback is now disabled?
This is exactly what I\'ve been wanting! Now I am more excited about Mozilla than ever before. I hope there will be more of these in the future.
Gerv, you rock! This kind of info is exactly what I've been wanting to see. This really gives a birds-view of what's going on with Mozilla, and that's just so much nicer than just looking at individual bugs.
I hope to see more of these in the future!
LOL, LOL, LOL
nglayout.debug.enable_scary_view_manager
LOL, LOL, LOL
Yes, this was goood! Things like this would have helped me figure out that libimg2==libPr0n much sooner than I had :-)
Also, anyone know the status of:
A) GPL licencing?
B) Newsgroup Renaming?
How's TRANSFORMIIX doing? (for next time, maybe?) Is that non-1.0?
check out <http://www.mozilla.org/projects/xslt/>, we target 0.9. Most issues are in bugzilla, the project page has a query link that get's you going.
Axel
Thanks a lot! Unfortunately, I keep crashing with a Necko error when I try to load the package. I'm using 2001030704. I'll search bugzilla later and file if it's not there.
[Axel's link has a stray comma in it; click here
<http://www.mozilla.org/projects/xslt/>
if you're too lazy to fix it yourself]
One day, I hope the status reports on mozilla.org are like this. Based around features, rather than engineers. No mention of how good particular antibiotics are, or that someone is having problems with their laptop, or other irrelevant stuff. And written in complete sentences!
Well done, Gerv ... Now, stop wasting time writing stuff for Mozillazine, and get back to your study! ;-)
-- mpt
Just a short addition, hyatt landed outliner yesterday. Happy filing ;-)
Great report Gerv, this should keep people who aren't "in" enough to understand the Mozilla.org reports on-board. Have you sent a copy for posting on Mozilla.org?
One project I'm really keen to know the status of is the support for XSL/XSLT in Mozilla. This is a key feature, which is currently reinforcing IE's dominance.
(My current project has just had to standardise on IE due to the need to deliver training material using XML & XSLT).
Read Pike's response to the "Thanks a bundle" comment in this article for the answer to your question. :-)
And do not standardize on Microsoft's older implementation of XML; make sure you use MSXML 3! earlier Microsoft implementations are non-standard. (though there's an encoding declaration bug in MSXML 3).
I've been following Mozilla and testing out builds and using it here and there, but reading this update helped clarify why the browser seemed so slow sometimes: the cache issue, or the image decoding overhead, etc. This is the update for testers/non-developers I've been waiting for, so thanks for putting it together. Even if you don't do another one, this has brought me up to speed at this late day in the game.
I've been following Mozilla and testing out builds and using it here and there, but reading this update helped clarify why the browser seemed so slow sometimes: the cache issue, or the image decoding overhead, etc. This is the update for testers/non-developers I've been waiting for, so thanks for putting it together. Even if you don't do another one, this has brought me up to speed at this late day in the game.
To be better, i hope there will be better documents rolled into the binary (nightly) builds, giving guidences to users for testing different compontents like XSLT, view manager (and anything else) not enabled by the default install.
as most of us are not hard core Mozilla coders, even don't know how to tweak prefs.js to turn some features on... having these kind of 'get-me-started' testing guidelines will help a lot, enabling more people to help testing.
It will be good, IMHO.
thanks a lot, Gerz! this is the perfect kind of summary for the vast majority of us out here, i'd imagine.. hope to see more, but only when you have the time, of course.. thanks again!
oops sorry for the typo... Gerv! 8)
Are they really going to keep on calling it libpr0n? :)
Guessing from some of the responses I've been reading here and there, why is this name supposed to be funny, or something? What am I missing? :)
Do a google search on pr0n and you'll get it!
pr0n = p0rn = pornography
BTW what is Hyatt\'s outliner meant to be had a look at the bug info. but I\'m still a bit lost
It's a replacement for the tree widget that's supposed to scale much better (tests with 2-million-item trees have acceptable performance, I am told). First being implemented in places like mailnews, where large trees happen fairly often.
when will it be used, e.g. for the mail folders and usenet trees?
According to the article, it says that it won't be activated until after Mozilla 0.8.1
|
