New mozilla.org Roadmap PostedSunday December 24th, 2000David Polberger writes: I'm 31337!! Or still up waiting for Santa. ;) To be on topic, the new roadmap still looks like it was drawn by a 2 year old, and I'd be happy to lend my talents to making one that doesn't but I think the goals displayed are achievable. Let's just hope the actual development path follows the lucky path instead of the unlucky eh? I think someone has a fetish for "." I've never seen more "." I cant believe it'll take an entire to reach version one. I would like to see mozilla reach version 2 in my lifetime. Thats all I ask. Thank you :-) I think someone has a fetish for "." I've never seen more "." I cant believe it'll take an entire to reach version one. I would like to see mozilla reach version 2 in my lifetime. Thats all I ask. Thank you :-) I just hope they go the lucky way. "th lucky way" has a lot to do with the desires of the Mozilla community as well as the pace of development. 1.0 will happen when the participants in this process decide that we've reached it. -Asa I had hoped that Netscape would first call their Moz release v5 instead of v6, or better yet called it a beta so they could upstage The Competition by saying that Netscape's beta version was better than The Competition's shipping version. It might be too late for that now, but this new road map looks like it included a chide to Netscape for taking a not-yet-beta product and releasing it as their final version release. Speaking of final version releases, I'm hoping that Mozilla drew up the plan based on its past experience, and not the stretched-out ambition of the Project's initial timeline that left all of us on the sideline disappointed. Stay tuned for Netscape 7, perhaps? Looking at the roadmap, it's sad to have to think that we pay dearly for perfection. So much time until an actual final 1.0 release, it's like an eternity. However, I do think I'll be alive for v2, as I'm 17 and if I extrapolate future plans from the current roadmap it should be ready by the time I'm 70. "it's sad to have to think that we pay dearly for perfection" What exactly are "we" paying here? And where does it say anything about perfection? How is this time like and eternity and what's it costing you to wait? Are you trying to ship a product based on Mozilla? Are you waiting for some API to stabilize so you can finish a time sensitive project you're working on? Did you invest money or development time with the expectation of some earlier ship date? Or are you just someone who likes to complain about how the workd isn't catering to your wants quite well enough. Your sarcasm ow whatever it is sure doesn't motivate me to make 1.0 happen sooner... Actually, I'd like to know what this post was supposed to accomplish? What was it that you hoped would be gained with these comments? Are you ill-informed? Are you trying to mislead or confuse others? Are you just rude? There are a lot of people working very hard on this project because it is important to them and your comments serve no useful purpose to the people actually doing something or those like yourself just quarterbacking from the sidelines. If you feel entitled to something here then you're just plain wrong. You want to see Mozilla 1.0 sooner (I'm still not sure why this is important to you) then step up and do something to make that happen. Insulting the people who are building a better browser for you is inconsiderate, unhelpful and just not cool. -Asa I totally and completely agree with you Asa... very good post --JED I have tried mailing Netscape direct several times, to no effect, So I'm hoping one of them is reading this. checkout this address: http://messenger.netscape.com/bookmark/4_5/images/112800_bottom_netscape6.gif that is the address of the Netscape 6 promotion image that I get every time I start my Netscape mail client. Trouble is they've spelt "convenient" wrong and its bloody embarrassing. Could someone kindly whisper in the direction of someone at Netscape that this is so? Thanks and sorry for going off topic. Netscape really should be more careful about typos, I wasn't the only one to notice that netscape.com/download were advising people to upgrade to "Netcsape 6" I am sorry... but i have to ask. Do you guys ever plan on finishing this project?????? What the &*ck. I am sure bible was written in less time. And it was a lot more stable that this piece of junk. I love netscape... and use their browser while it is about a million times slower then IE5.5. But... till when??? You all hate microsoft for putting out buggy products... but atleast they put out a product. I have said my prayers on mozilla. It will never complete. We may see a pre-release version sometime in the next century. Its all netscapes fault... thats what you get from free. Should paid and hired some decent people to program. i'm sure they don't have enough money to pay better programmers than the ones already working on it... You have said your prayers for Mozilla? I already consider Mozilla 0.7 as a "first release", because it is so stable and feature complete for me. Feel free to rename your browser to 6.0 if you want though, it's just one change in the chrome. By definition, Mozilla will probably never stop growing, and let's keep it that way. And maybe the Great Lizard heard your prayers, and will be kind enough to let you download Mozilla when it gets good enough for you. From what I understand, mozilla.org has no plans on "finishing this project." There won't be a end-all final release. Even after Mozilla 1.0, the trunk will still grow and take off into some fascinating directions. I'm sure that as long as Netscape is going to be around, the Mozilla project will continue to flourish. Exactly, since other open source projects are been in existence long before mozilla.org was founded, and they have never really finished (BSD, Linux). They are always on going, releasing stable releases with more bug fixes and new features periodically. Not to mention that Internet standards are always evolving, and new standards are being created, and Mozilla would eventually support them once those standards are finalized. Alex <:3)~~ Hmmm...finishing the project? No, we currently have no plans for that. Hint: the "1.0" signifies that the project is just beginning. Assuming you meant "when is 1.0 going to come out?", well, I assume you're capable of reading the roadmap like everybody else. Frankly, I don't understand what you're complaining about. Why are you so anxious to see a release? The difference between us and Microsoft is that we put out a release *every day*. They're called nightlies; try one. Hint #2: Mozilla 1.0 will be nothing more than a nightly with "Mozilla 1.0" on it. --Blake hello! i just installed beonex to take a look at mozilla again and i thought, it would be time to adress some of the issues i always wanted to post somewhere but never did... i'm not new to mozilla but the last weeks i used only khtml (or opera) and started work on a small and fast khtml browser but i still think mozilla will make it somehow. the technologie is impressive. it's bloated, yeah, but this is not much of a problem if you promise to work on it and release a really slick "navigator only" version someday ;) mozilla is one of the most usable browsers today and there are only small problems that are holding me away from it (besides memory usage). those are: - the gui. both classic and modern are great works, but i don't want to use them day by day. something clean and clear like classic combined with something new and fresh like modern would be great! at least you could use new buttons instead of the VERY old netscape ones.. :-/ also i would like to have it a bit smaller. the size of the netscape buttons is great. now if i could remove the text and make the throbber about half the size, the toolbar could be made a lot smaller. i also don't understand the h-space between the buttons. why is there so much wasted space? is it to make hitting the buttons easier? than there is a larger forward button and a smaller back button... i can understand that this is because of the text, but it should be the other way round, cause the back button should be used a lot more often. i know that this is just a matter of skinning but it's still important. and yes i already tried to do it myself and i failed ;) - a weird table rendering bug. the gecko rendering is the best i have ever seen. with ONE exception.... in tables with a small border, this border is almost everytime "buggy". sometimes it's not showing up at all or in doubled size. unfortunatly i mostly use the browser to access webforums and all those forums are rendered ugly by gecko. some examples are: http://forum.winner.de http://www.liebesforum.net http://forum.betachat.de (they are all german). i thought the bug would be killed soon but then i tested release after release and it was still the same.. the weird thing is, that it worked great in an early milestone (i can't remember the exact number) but then got screwed around M16. maybe it's fixed now, i don't use the latest mozilla nightlys so i'm not sure (the latest was from 6. december i think). yes this are not really big problems but i think they are important. if i would find a really great theme and this rendering bug would be fixed and i could download and load navigator only, i would be perfect happy with mozilla. finally a short note about rendering speed: i noticed that mozilla renders new incoming characters immidiatly unlike IE, netscape 4 or opera (which wait for a whole line to be transfered before they render it). this may be usefull, but the downside is that lines are not rendered immidiatly. i made a little testsuite to be sure. netscape4, opera and IE are rendering new incoming lines immidiatly while mozilla waits a while (a few seconds) until the new line is displayed. that's why netscape4, opera, etc seem to be a lot more "snappy" than mozilla. it would be great if you could somehow catch up and render new lines faster. btw, khtml even waits for a bunch of new lines before rendering them, so gecko seems to be faster than khtml but slower than opera, ie and netscape4 (if tables are not used). i hope some of the thoughts are usefull to you and i REALLY hope that some of them are already solved or in progress :) would like to hear some comments. but don't flame me cause i don't do all the things myself. i don't expect anything from the mozilla developers, i just want to give my impressions as an unbiased user and try to point out some (IMHO major) usabilty issues. keep up the good work. free software simply rocks. i looked at the webforums (nightly build 2000122420, win32) and couldn't see any ugly rendering results. could you describe more exactly what bugs do you see with mozilla. I don't know what you see, but I see missing horizontal lines in his tables (each forum entry should be separated by a horizontal line). This is covered in bug 48827 http://bugzilla.mozilla.org/show_bug.cgi?id=48827 I think. There's not really been any recent activity to fix this. exactly! i made a screenshot to show this here: http://home.wtal.de/borgmann/tablebug.png it's really destroying all the beauty of gecko rendering :( also notice the different HR lines! this should never happen... i downloaded the latest mozilla nightly and the bug is still there. I happen to like the way mozilla treats character/line rendering. That same piecemeal rendering that can appear slow on fast sites can also appear faster on slow sites. One of my biggest gripes with IE is that when a site is slow, it can take forever for something I need to be displayed--when I am looking for a specific link, graphic, or whatever, I don\'t always need to wait for the entire page or line to be rendered before a part of it is useful to me. Perhaps I don\'t know what I\'m talking about. All I do know is that I have made similar observations about rendering in the past, and that I appreciate the way mozilla renders pages. Maybe what I like is something completely different. Just a thought. Not to add things, but how difficult would it be to add character vs. line rendering as an option? Perhaps not worth the wait. I do not see any clear criteria for an acceptance test for the milestone releases. I am afraid that the whole PDT massacre like \\\' I will not hold the release for this\\\' will repeat and mozilla will ship with well known crashers. In order to prevent these long living crash bugs I would propose to make the checkin policies more strict, requiring the assignees of repeatable crashers to fix them first before they submit any other peace of source code. The lack of criteria will probably make it very difficult to assign the target milestone. How many html4 , css1 and css2 errors will we tolerate? Can you imagine to ship mozilla1.0 with a known crasher, as it has been done with NS6.0? How much memory we will give mozilla? I think we should add some criteria extending the dates. Otherwise all decision\\\'s will be based on the wisdom of the drivers and theire private opinion instead of facts. Read the last line of the roadmap: "We'll publish a "Mozilla 1.0" draft definition document shortly..." I have red the sentence, and this is my "other input from the community." While another keypoint of the posting was the crash prevention and if we do not get more traction on the crash problem we will face the same problems as during the PDT phase. hi *, sorry for this completely off topic question. Does anyone know of a good resource on how to make tables so Mozilla or Netscape eats it fast? I am currently working on a big table, and it takes NN4.7 41secs and Moziall 0.6 33secs to render it. (nope, i' am not using a 486, I am on a PentiumIII, 265 MB RAM). IE, Opera and Konqueror render the page in about 5 secs, right as the Codes comes in. Table width are all set. Any Ideas? Any links or so that might help? thanx ! phisch mmm... haven't had any problems like that. Moz since day one has always kicked IE and NS's butt with tables.. any URL for me to look at? would help. thanks Do the tables in question use the bgcolor attribute? I seem to remember something about this, but can't find it in bugzilla. I have a similar problem with tables at netflix.com. Maybe I'll file a bug report. Hello vondo, yes, they do have bgcolor. I use the table to make a calendar with all shifts of our workers (that is, why I currently cant hand out my table, it contains to much sensitive data). It\\\\\\\'s like lightblue background is nightshift1, dark blue is night shift 2 etc. the table has 32 cols and 112 rows. uh. OK every one, I have a week off now :)) ********************************* Happy New Year to everyone on Mozilla.org and Mozillazine.org! ********************************* phisch Since the number of outstanding bugs just continues to go up and up, why should we think that Mozilla will _ever_ be a professional quality product? The only way to predict when a product will be shippable with adequate quality is to extrapolate from the bug curve, and the constant increase in the defect line over the months predicts that it will never reach a low level of defects. There appear to be basic architectural flaws in the Mozilla code base which lead to excessive regressions and a low bug fix rate. If these aren't addressed, then we might as well write off this code base. I think that a couple of the reasons the bug count may be increasing is that the product is being put in front of an increasingly larger audience, and being held to increasingly higher standards as it moves toward v1.0. Could you provide some more detail on what these "basic architectural flaws" are and how they lead to excessive regressions? An increasing bug count is not in itself sufficient evidence to support this conclusion. I use Mozilla nightly builds for some time now and they are getting better and better. I'm curious - how do you know that IE's bug registry isn't going up and up? Of course new versions of IE, or any significant software product will show the same "curve" Strauss is talking about. The alternative is that your new versions provide the same or less functionality at greater cost - who would buy that? Larger (more powerful, more complete) products have more bugs. Without a "Silver Bullet" breakthrough that will remain the case forever. I've worked on projects that used the "follow the curve" metric Strauss describes. They ship just as buggy as other projects, but people learn not to file bugs, because to do so delays ship. Great - now we don't even know where the bugs are! Why does that matter? The question is whether Mozilla is a failing project. That has nothing to do with IE. Anyone who has been associated with the repeated shipment of commercial software products, as I have for many years, knows that the defect curve is the primary determinant of software readiness to ship. Late stage development activities invariably consist of attempting to drive this defect curve down to an acceptable level under a feature freeze regimen. Since the project began, the defect curve for Mozilla has never gone down. (Though you can't see this fact any more, since bug trend graphing itself has been broken for months.) Even given feature freeze, code review before checkin, and strict bug ranking, the curve is still not going down. That is a description of a project which never becomes shippable. I wouldn't need to have this discussion with any trained software professional, but unfortunately Mozilla has bought into the open source community's set of assumptions -- primary among which is that flakey software is perfectly acceptable. I have yet to meet an advocate of the open source religion who had any familiarity with any software engineering process models from the 1960's to the present, and so it is possible in that community to simply ignore obvious bellwethers such as an ever-increasing defect curve and high regression rate. The people involved in open source have only worked on toy projects on their own time, for the most part, and do not understand the basics of software process or quality. Mozilla has not developed the project management structures that would be needed to put this project on a shipment track, instead promoting the untrained keyboard jockeys who've made the mess so far to project management positions for which they have no qualifications. Lee The increase in the bug curve has been going on since months before the Netscape 1.0 release. Inability to fix bugs without creating regressions is sufficient reason to believe in basic architectural failings. A mature software development process would then begin root-cause analysis to find the problems. Without performing that investigation, I don't know what the root causes are -- though I suspect the long compile-link-run turnaround cycle is a critical one. Lee An increasing bug number is a result of a bad planned test strategy, a lot of the bug\'s appear by randomly accessing the web. If you remember the PDT phase, I would also not call it a planned release strategy. A change in the bug curve would indicate that we are moving toward something that could be branded as 1.0 and this should be shown as one criteria in the roadmap. I did a query for open browser bugs its above 8000. Assuming that 30% are dupes, and that we know all bugs, it will require 5600 bugs to be fixed. If you distribute this to a year this means more than 15 bugs fixed every day, without any regressions, and taking a regression rate of 30%, it would require 20 bug fixes per day. It seems that mozilla needs more contributors otherwise all these roadmaps are sweet dreams. I am one of those crazy folks spending their time in Bugzilla, so I know what I'm talking about. There are more bugs reporting now that one year ago? Does it mean there are really more bugs in Mozilla? Or does it mean our user base is growing? For your information, and based on my experience with Bugzilla, half the bugs reported are duplicates or invalid, a quarter are RFE's (requests for enhancement) and the last quarter are either minor bugs (typos, etc), or profile problems. Again for your information, regressions are usually fixed within the day, within the week at most. Also we do not oblige you to use Mozilla and to read this Mozilla forum. If you want to flame us post a Slashdot article. Thanks, Fabian Guisset. > I am one of those crazy folks spending their time in Bugzilla Me to. (search for my name in the bugzilla comments) I don't flame. I am seriously concerned that we need more contributors. I am talking about 290 HTMLTable bugs assigned to karnaze, 190 layout bugs assigned to buster and the bug list of waterson also looks rather impressive. And I asked myself whether people spending their time in bugzilla would not be more effective by directly fixing the bugs. hello again, i hope the developers are aware that the table rendering quality becomes worse from nightly to nightly in the last days? i took this screenshot with a recent nightly (watch the titlebar): http://home.wtal.de/borgmann/tablecrap.png this is not even funny anymore. i hope someone can tell me, that this is just because you're currently working on the table problems :) hmm, no table checkins over christmas, the urls render fine under Win98. Should be an OS specific problem. Have fun while fixing them. :-) (Any help is wellcome). Yup they have problems with table borders and table sizing. And I use Windows 2000 with the latest nightly. I hope they fix it soon. In the meantime just vote for the bugs on bugzilla. There are lots of table related bugs reported in bugzilla. The page in question seems to have lots of nested tables and have border problems in mozilla. Why don't you try searching bugzilla for HTMLTables component bugs with border in the summary? ie http://bugzilla.mozilla.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&product=Browser&component=HTMLTables&cmdtype=doit&order=Reuse+same+sort+as+last+time I was wondering whether start up speed is going to be addressed. Right now, a nightly (with a exclamation point on it) on an AMD k6-2 300mhz with 64 mb of RAM it takes close to 15-20 seconds to load. On a PIII 600mhz with 0.5Gb of RAM it takes about 5-10 seconds. That is clearly unacceptable. This is when I fire them up the first time. Second time around is a bit better with 10-12 seconds on the AMD300 and 3-6 seconds on PIII-600. Still, it takes a while. I hate to compare it again IE, but it is the benchmark against which all the browsers are judged. Anyway, IE comes up almost instantly (within 1-3 seconds) on both PCs. I did similar tests, on a PII 333 and 128mb ram. Moz took about 11-12 sec. for the first start and about 5-6 for the second. IE started up in about 2 seconds. So then I timed how long it took to actually startup and completely render my home page, I was totally shocked when they turned out to take almost exactly the same amount of time. It seems that IE renders displays the menubar and button bar almost immediately while still not having fully loaded all the dll\'s it requires. Either IE starts displays UI when it's not ready to use it / render pages - or Mozilla is simply faster at rendering (at least at your home page). Couldn't that be true as well? This isn't intended as criticism; I know part of the perceived start up problem in comparison to IE is because so much of IE is loaded when Windows is loaded, etc. But I thought folks should know: at a conference I attended in December, at least 4 folks who identified themselves as long-time Netscape users told me that they had downloaded Netscape 6 and then uninstalled it; for the most part, they were driven away by the start up time. (None of them knew anything about Mozilla.) This in response to a recommendation on my part that they use Mozilla or Netscape 6 to read pages in Unicode. I don't think your tests were accurate. I just did a sample test on several pages and IE came out well ahead in starting up and rendering the page. I've tried www.zdnetonebox.com, www.cnn.com and www.news.com. The netted Mozilla 1 second advantage compared to the simple start up test. Regardless of the semantics, the startup time is hugely important. My work day consists of opening and closing the browser many times (i guess, I do other stuff as well), having to wait for moz to open while ie comes up instantly is ridiculous. I'm sure this has been discussed before, but I'm pretty new to Mozilla (using nightlies under Linux). What is the rationale for tying the browser and mail/news so closely? It seems that if they were separate projects each could proceed at its own pace. Also, I think the functions of browsing, mail, and news are all distinct. I like the browser quite a bit, but after suffering with the performance and feature quirks of the mail/news part, I switched to PAN. I have yet to find a mail program that does what I like. I'm sure this has been discussed before, but I'm pretty new to Mozilla (using nightlies under Linux). What is the rationale for tying the browser and mail/news so closely? It seems that if they were separate projects each could proceed at its own pace. Also, I think the functions of browsing, mail, and news are all distinct. I like the browser quite a bit, but after suffering with the performance and feature quirks of the mail/news part, I switched to PAN. I have yet to find a mail program that does what I like. When is it going to be released? My guess is that it will be released on Monday. Does anyone know when the next version will release? And what version it would be, like 6.01 or what? I think it's funny how no one comments about the releases at all. Roadmaps need at least a little bit of follow-through. definitely looking forward to 0.7 tho |