Mozilla 0.9.6 Released
Tuesday November 20th, 2001
mozilla.org today made available for download binaries of the Mozilla 0.9.6 Milestone. The builds are available on the releases page, or you can get them directly from the ftp site.
#145 Re: Re: Re: Re: Re: Re: Re: Re: Re: What happened
Friday November 23rd, 2001 2:52 AM
You are replying to this message
" One expects some of these to slip, but the slip numbers are large, in the three figures per build"
What does "in the three figures per build" mean? So you're saying that fixing 1,600 or so bugs in a milestone and getting to the end of the 5 week cycle (these are _milestones_, checkpoints, oportunities for talkback data from a larger base, etc) with one hundred or even two hundred of the bugs targeted at that milestone unfixed means something? When nearly 200 individual developers set target milestones indicating that they'd like to fix some set of bugs for a particular milestone and don't get 100 percent of them fixed then our project is in jeopardy? If a developer says he's going to try to fix 15 bugs this week and we get to the end of the week and he's only fixed 13 or 14, you call that problem? I don't. I'd love for people to be a little more realistic when estimating what they can have fixed when but so long as the number of bugs actually fixed each milestone is predictable (and it is) then I'm not too concerned that some people are a little too ambitious. If we had targeted 1500 bugs at 0.9.6 and then fixed 1600 would we be better off than if we targeted 1700 and 0.9.6 and only fixed 1600? (the real story here is a bit more complex. we actually target about 2000 bugs at a milestone and fix about 750 of those plus 850 or so that weren't at the time (but may have been at one time) targeted at that milestone. nothing prevents someone from fixing a bug that hasn't been targeted yet or that is targeted at a later milestone. this could all be better and it would make my job managing the releases a bit easier if it was more accurate but it has very little to do with the actuall quality of the product and nothing to do with "the stability situation".)
Can you do me a favor? Go get M10, then M11, M12, M16, M18, 0.7, 0.8.1, 0.9.1, 0.9.3, 0.9.5 and 0.9.6 (feel free to get all the milestones but for my purposes this should be a good sample). Run M10 and use it for an hour or so. Then run the next for an hour or so. Then the next. When you get to 0.9.6 post a comment here telling me that we're not getting better with every milestone. If you can also include detailed explanations of how the product is worse now than it was in any of those previous milestones that would be helpful. I've been using Mozilla since before mozilla.org was providing binaries. I've been using it as my exclusive browser since M8 and as my exclusive mail client since M10. I use the browser and mail client about 8 of my 12 hours workday. I notice when things get better and I notice when they get worse. I don't usually test nightly builds on the weekends but I've used and tested the overwhelming majority of weekday nightly builds for the last 3 years. I admit to being a Mozilla advocate but I'm also a user and there aren't many (if any) users that have logged more hours on more builds than me and, while some critics aren't, I'm very comfortable saying that it's getting significantly better with every milestone. There aren't many that have logged more time reading and triaging bugs in Bugzilla either. You can throw all the numbers you like at me and you can point to mangelo's counts or the bugzilla graphs. You can talk about the rising lines and the different curves but you won't convince me that we're not getting better all the time. My experiential knowledge flies directly in the face of the numbers and lines you're so fond of mentioning here. I suspect that many others who respond to your posts here are coming from a similar place. We use the product and we _know_ it's getting better so when you and mangelo say that your graphs show it getting worse, we ain't buyin' it. I don't know about you but mangelo admits to not using the builds, not even the milestone builds he's 'reviewing' at his homepage. I do use the builds. I've used thousands of them on more platforms than most. Not only have I used mozilla builds but I've used mozilla based products coming from other companies making similar and completely different products. I've talked with developers, qa and users of other mozilla based products. I'm about as on top of the state of the code and the products coming out of it as anyone can be. You're not. Mangelo isn't either. There are a lot of people (millions of them) using mozilla and mozilla based products and the reviews and comments at mainstream journalism outlets like c|net, in newsgroups, bug databases, mailing lists, weblogs like mozillazine and slashdot, etc., all suggest that the product is getting better (even, suprise, "good"). I'm not denying that many of the lines are going up in the bug graphs (even "fixed") but I am disputing your statement that the product "stability situation is not under control". I've got the mtbf numbers, the user reviews and comments and the direct experience to back that up. You've got some counts of entries in a database made by over 16,000 different people. (did you know that the average number of people experiencing our 'topcrasher' bugs is going down at the same time as the number of people using our nightly and milestone builds is going up. figures like this would suggest that the product is getting more stable, not less.)
I use the builds. I feel confortable making judgements based on my formal and ad-hock testing of the builds. I spend a lot of time in the bug database and in the talkback database. I feel comfortable making judgements about stability of the builds. How much time do you spend analyzing bug data, talkback data and testing builds?