MozillaZine


[.9] I'm sorry but I don't understand new releasing practice

Discussion about official Mozilla Firefox builds
daihard
Folder@Home

User avatar
 
Posts: 16633
Joined: November 17th, 2002, 6:27 pm
Location: Lynnwood, WA

Post Posted June 15th, 2004, 10:41 pm

[Unknown] wrote:What I mean is, if you can't have any features after point x.... you won't care if they are buggy, you'll get them in there! Part of the fun of programming is writing the new features, at least in my mind, so this makes it even worse.

It's very true, but that's really a double-edged sword. If you rush to get in the new feature that you enjoy writing, you'll end up spending far more time than you expected fixing problems with it later. Trust me, I've been there. ](*,)
Kubuntu 8.04 (kernel 2.6.24-25-generic) / KDE 3.5.10
CentOS 4.8 (kernel 2.6.9-78.0.22.ELsmp) / KDE 3.5.10
Mac OS X 10.6.1 (Snow Leopard) / iPhone 3GS (32GB black)

Kob
 
Posts: 33
Joined: December 4th, 2002, 12:56 am

Post Posted June 15th, 2004, 10:56 pm

Community, Dev and new users would be happy if a 2" high red letters would say on the download page: "This release is for testing only - not for production use. Please use 0.8 for production".
Community will test and report bugs, dev will get the required debugging support, and new users will not get burnt.

hariskar

User avatar
 
Posts: 498
Joined: November 2nd, 2003, 4:03 am
Location: Kavala, Greece

Post Posted June 15th, 2004, 11:00 pm

Kob wrote:Community, Dev and new users would be happy if a 2" high red letters would say on the download page: "This release is for testing only - not for production use. Please use 0.8 for production".
Community will test and report bugs, dev will get the required debugging support, and new users will not get burnt.


I agree..

fittysix
 
Posts: 131
Joined: November 30th, 2002, 9:54 pm

Post Posted June 15th, 2004, 11:25 pm

well, i think this is the first release to live up to the Beta Software title that it is, this is beta to the deffinition of the word
previous point releases have been far more production quality, but they are still beta none-the-less

I do agree that this release is a bit of a setback for the project in a media standpoint, however, this release was destined to be buggy; if it wasn't than there would be no reason for it not to be 1.0
i'm thinking they're going to release a 0.91 so that it doesn't look like they're intentionally releasing buggy software

Jugalator
 
Posts: 282
Joined: November 9th, 2002, 11:10 am
Location: Sweden

Post Posted June 15th, 2004, 11:27 pm

Yes, 0.9 can probably be a bit buggy, although I haven't personally stumbled upon any bugs listed by Peter(6) -- keep in mind that *all* Fx releases have had a large amount of known bugs scheduled for coming milestones, but that large amount is decreasing.

I don't really see a problem with this though. The important thing is that Fx is stable when it gets out of beta. Right now it's not even there -- the 1.0 beta is next.

So this isn't unusual with software under development, and I don't think we should expect the Firefox team to be somehow greater than all others.

Unfortunatelly a lot of people will be dissapointed and stop using it...

If "a lot of people" believe they could expect final release quality from pre-beta software, they're *really* misunderstanding what all these Firefox previews are and what the point of them is. Firefox is just in version 0.9, a pre-beta version, why is so many expecting so much from it? The program stability efforts have just begun (when 0.9 had been released) and will continue during the coming months until version 1.0 is released. Is that so hard to understand?
Last edited by Jugalator on June 15th, 2004, 11:33 pm, edited 3 times in total.

daihard
Folder@Home

User avatar
 
Posts: 16633
Joined: November 17th, 2002, 6:27 pm
Location: Lynnwood, WA

Post Posted June 15th, 2004, 11:28 pm

fittysix wrote:i'm thinking they're going to release a 0.91 so that it doesn't look like they're intentionally releasing buggy software

That'd be a good idea, especially when the 0.9 release contains some new features that have never been tested before. That said, anyone who's used Microsoft products should know that "Release" vs. "Beta" doesn't mean much. :p :p
Kubuntu 8.04 (kernel 2.6.24-25-generic) / KDE 3.5.10
CentOS 4.8 (kernel 2.6.9-78.0.22.ELsmp) / KDE 3.5.10
Mac OS X 10.6.1 (Snow Leopard) / iPhone 3GS (32GB black)

Hybrid

User avatar
 
Posts: 210
Joined: December 3rd, 2003, 12:29 am
Location: /usr/src/linux

Post Posted June 15th, 2004, 11:40 pm

1.0 has been held back long enough. 0.9 needs to get out there and the project needs to push forward. Bugs or not, this release has been long overdue. It would have been great to fix many of the long trailing bugs that will no doubt flood this board with post after post asking for assistance or complaints.

Wether or not anyone agrees that 0.9 should have been made final, I think many people can agree that development needs to more forward and its been way to long since the last milestone which was far from perfect I might add. I do believe that aside from last minute changes, the 0.9 is on par or just slighty worse then 0.8 milestone, I mean lets not get ahead of ourselves. 0.8 wasnt rock solid as everyone likes to make it seem. It had its share of problems. The trunk builds have been the most stable which is a reverse way of developing but thats how its happened.

Lets move forward and forget about 0.9. I would love to see focus on the trunk and making 1.0 a true release. As for Ben, I am sure he was thinking along the same lines and made an executive decision on what was good for the project and not the milestone. Right or wrong, its done, so lets move forward and quit arguing over symantecs.

daihard
Folder@Home

User avatar
 
Posts: 16633
Joined: November 17th, 2002, 6:27 pm
Location: Lynnwood, WA

Post Posted June 15th, 2004, 11:45 pm

Hybrid wrote:Lets move forward and forget about 0.9. I would love to see focus on the trunk and making 1.0 a true release. As for Ben, I am sure he was thinking along the same lines and made an executive decision on what was good for the project and not the milestone. Right or wrong, its done, so lets move forward and quit arguing over symantecs.

From what I've learned so far, the 1.0 release will be based upon the same BRANCH that produced 0.9. Someone please correct me if I am wrong.
Kubuntu 8.04 (kernel 2.6.24-25-generic) / KDE 3.5.10
CentOS 4.8 (kernel 2.6.9-78.0.22.ELsmp) / KDE 3.5.10
Mac OS X 10.6.1 (Snow Leopard) / iPhone 3GS (32GB black)

[Unknown]
 
Posts: 130
Joined: October 7th, 2003, 12:32 am
Location: Los Angeles

Post Posted June 15th, 2004, 11:55 pm

Yes, I have read at least eight separate times that the branch is for 0.9 and 1.0... probably just a typo.

And, yes, I realize it's a double edged sword. Not only have I been there, but the next release of my own software is going to be feature complete :/. I decided to make an interim beta release to try to weed down some of the bugs, and there were some I mayn't ever have noticed... so that's good.

Either way, while I think some things will be a major set back - the bugs in theme management, the new default theme too early.... let's be realistic. It's a month. One freaking month. (or two?) There are programs I don't even realize had new versions for a month or even two at a time!

It is ENTIRELY POSSIBLE that most people will never even notice the 0.9, or even the 0.8, release happened. They'll just see 1.0... or even 1.1 or something... and upgrade. They won't even know a buggy release happened. Especially if they come from IE.

-[Unknown]

daihard
Folder@Home

User avatar
 
Posts: 16633
Joined: November 17th, 2002, 6:27 pm
Location: Lynnwood, WA

Post Posted June 16th, 2004, 12:00 am

[Unknown] wrote:It is ENTIRELY POSSIBLE that most people will never even notice the 0.9, or even the 0.8, release happened. They'll just see 1.0... or even 1.1 or something... and upgrade. They won't even know a buggy release happened. Especially if they come from IE.

I like your sense of humour about IE... wait, it's reality. :D

In any case, it is very important that the 1.0 release be rock solid in order to appeal to the broad audience (and talk a lot of them into switching from IE). I believe using the same branch for the 1.0 release will be helpful in that respect.
Kubuntu 8.04 (kernel 2.6.24-25-generic) / KDE 3.5.10
CentOS 4.8 (kernel 2.6.9-78.0.22.ELsmp) / KDE 3.5.10
Mac OS X 10.6.1 (Snow Leopard) / iPhone 3GS (32GB black)

bytemaster

User avatar
 
Posts: 1264
Joined: August 18th, 2003, 11:34 pm

Post Posted June 16th, 2004, 12:34 am

I don't believe I have seen anyone list any solid major bugs. Many of the "major" bugs on Peter(6)'s posts have been arround for a long time, and are minor, IMHO, compared to all of the bugs that have been fixed between 0.8 and 0.9.

Granted, that the profile-migration problems are pretty bad, but we didn't even have a migration tool from 0.7 to 0.8 and people had to wipe out their Chrome folder or profile manually then to get the browser working.

This build seems to be very stable, and much more production ready than 0.8, with the exception of the extension manager. The extension manager has problems, but we have yet to have anyone nail down the problems to know if it lies in the manager or the extensions themselves.

Over all, this is a very good build, usage wise, and I think it works well.

The only complaint I would have is removing Extensions and themes from the Options menu. I understand putting them under tools, but I belive then they should be in both places.

Now, if anyone has a specific problem with 0.9, that is a different story, but to just say "problems" isn't really helpful. Once you clean up your old profile, everything seems to go smoothly.
Brainbench Most Valuable Professional (MVP) for Computer Technical Support
www.brainbench.com - <a href="http://www.brainbench.com/xml/bb/transcript/public/viewtranscript.xml?pid=742832">view my transcript</a>

Jubijub
 
Posts: 135
Joined: April 11th, 2003, 10:25 am
Location: Lyon, France

Post Posted June 16th, 2004, 1:02 am

djl47 wrote:
Twenty odd years ago when I worked at an aerospace contractor in southern California we ridiculed a decision made by a VP instructing the data center to no longer limit the number of users who could sign on to the mainframe. It seemed absurd at the time to let 500 or so users sign on to a system that was barely able to service the load created by 250 users. The load made the system glacial. Many users staggered their hours so they could get their work done. In hindsight it was an outstanding decision. When *every* user in the company from the data processing staff to the aircraft and missle engineers bitched about poor response time the company coughed up the funds to increase capacity.

Ben's decision to include an unpopular theme and lightly tested new features in v.9 may be one of the best software development decisions ever made. The new theme and extension managers had to be included in the v.9 release, there is just no practical way these features could have been robustly tested if introduced after v.9. Prior to the introduction of the new theme there was very little feedback in these forums on the new extension and theme managers and precious few themes or extensions that had been made complaint with the new managers. Ben's decision stired up the pot, cause lots of whiners to have kittens, and inspired some users like myself to look at themes and extensions for the first time since v.5. I no longer care how the decision was made or who still has their panties in a knot about it. It was the right decision and I am going to support it. When the new theme was released the theme authors were tripping over themselves making alternatives compliant with the new theme manager. Extensions were updated by third parties when the extension author could not be reached. The result is that far more users tested the new managers with a more diverse selection of themes and extensions than would have occurred if Ben had chosen the status quo. All this bitching and moaning tells me that Ben whacked a complacent user community upside the head with an oversized LART and got our attention. That may not have been Ben's intent but it has in my not so humble opinion been a positive development and Firefox is going to be a better browser because of it.

Going forward. The devisive complaining needs to stop. It hurts the Firefox user community. The developers currently need our eyeballs to find problems in the new release and the future daily builds. Ben and the other developers can fix the bugs provided we do an adequate job of finding and documenting them.

Now we just have to stop providing zip builds and force everyone to use the installer. Test the product by using it the same way that the great unwashed masses will use it. And that means using the installer.


This makes sense to me...I must admit I took the problem more on a user side than on the developer side...
Actually I do follow the Eclipse project developments, and a similar thing happened : they changed the plugin API in a way all old extensions won't work anymore...
The fact is a lot of people released new plugins or updated plugins matching v3.0...

If you're right, I still reckon more communication should have been done on this, explaining the reasons....
I agree with somes posts here : this project gains a major boost thanks to the community (testing, bug report, validation, etc...)...
MCScott seems to be far more aware of this (it's very usual to see him asking for testing of a particular feature that requires testing, such as the spam filter for instance, and updating it from the testers comments...

On the other hand, ben seems to do it more alone...if he wanted very particular things to be tested, he could have asked for so...instead of releasing with no words a release basically saying "hello, I'm buggy, please test me, but find what's wrong first"...

johnleemk
 
Posts: 1464
Joined: October 29th, 2003, 6:19 am
Location: Malaysia

Post Posted June 16th, 2004, 3:58 am

Actions speak louder than words, Jubijub. Ben has asked people to test before, but apparently he decided that the scale wasn't big enough. Plus, he may have decided that bad publicity is better than none. Knowing what Firefox is and why it exists is better than not knowing about other browsers than IE at all.
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8b) Gecko/20050205 Firefox/1.0+

Schpyder

User avatar
 
Posts: 42
Joined: January 26th, 2003, 11:51 am

Post Posted June 16th, 2004, 4:06 am

Hybrid wrote:1.0 has been held back long enough. 0.9 needs to get out there and the project needs to push forward. Bugs or not, this release has been long overdue. It would have been great to fix many of the long trailing bugs that will no doubt flood this board with post after post asking for assistance or complaints.


Well, so much for being a bug-driven project. :/

Let's get a couple things straight. I don't think anyone's arguing that including the new features was a bad decision for 0.9. I (and I think others) are arguing that maybe a week or two more should have been taken to look at some of the glaring bugs that were most likely to bite people in the ass. Milestone releases are, by and large, your public face, so it behooves the developers to have them a bit more polished than your standard nightly. This was the whole point of the trunk/branch system. Remember, it's much harder to get rid of a bad impression than it is to lose a good one.

I'm using 0.9 right now, and I quite like it. I haven't run into any bugs yet, and all the extensions I use work fine. I don't use themes, so I don't have problems there. I even like the new theme quite a bit (except for the navigation buttons, but I don't ever have to see those with my setup). But I also hang out on quite a few other large forums with large Firefox followings, and it seems that 60%+ of the people trying 0.9 are having large problems, whether with migration, or extensions, or themes (ESPECIALLY themes). And a large chunk of those are throwing up their hands and saying "screw this, back to 0.8." Or back to their other browsers if that's where they're coming from. This is NOT the way to get more eyes reviewing your bugs! If you turn off your userbase enough, they just won't care about the whiz-bang features and you're no better off than with a regular old nightly, except for getting a lot more people less interested in your product.

It's become apparent now that Firefox has moved from a bug-driven project to a date-driven project. I can only hope this doesn't hurt it more than it already has (from a public perception point of view). I liked the milestones meaning something. They were safe bets. Not any more.
"More fun than a barrel of monkeys... with dynamite strapped to their backs!"

BenBasson
Moderator

User avatar
 
Posts: 13671
Joined: February 13th, 2004, 5:49 am
Location: London, UK

Post Posted June 16th, 2004, 4:58 am

All I can say is that if all the people who insisted on using the Trunk builds to avoid "problems" had downloaded and tested the Branch builds a bit more, instead of largely ignoring them, then 0.9 would probably be in a much better state. Pretty much all I read up to the release of 0.9RC and 0.9 was "Extensions work with the trunk, so I'm sticking with it" or "The trunk is more stable, so I'm testing that"... as if either points make any sense when nightlies are supposed to be for testing anyway.

Now there's much less of a problem getting test feedback and bug reports. Ben probably didn't push out 0.9 to that end, but it's what's happened.

Return to Firefox Builds


Who is online

Users browsing this forum: No registered users and 0 guests