<Glazblog/>

E pluribus unum

When I wrote my article The faster release process of Firefox, I did not expect so many reactions outside of the inner mozillian circles. I did not expect that Mike Kaply, who was thinking in the same terms and himself pondering an article, would jump on the bandwagon. I did not expect Asa's incredibly rough answers to his posts. In fact, I got many private messages thanking me for my article. I need to say it again here: I never said the faster release process is entirely bad. But that's another discussion.

I read things that really hurt the community and I think some conclusions have to be drawn, and some changes probably have to be applied.

  1. Mozilla is a lot of things. It's a Foundation, a Corporation, a large Community, millions of users, followers and evangelists. It's also a Manifesto, Products, Projects, Add-ons, Translations. For most of us in the Community, and for various reasons, Mozilla is also a very large part of our lives. We say "Mozilla should..." or "Mozilla should not..." because we do care about Mozilla, and we're not all nerds with only coding or localization abilities. We come from various landscapes, and if the coding skills of the community are used, the other skills are mostly ignored. The private comments on my blog post I received are very clear: some people did not want to express or delayed an opinion similar to mine in public (on a blog for instance) because they feared the reaction of Mozilla. Someone told me that "as soon as you start criticizing in public such high-profile decisions, you're tagged as a troll". I also got messages from add-on developers (XUL add-on developers I should say) saying they're fed up with the add-on update process, are afraid to have to do it now 10 times per year, and stating the new SDK supposed to resolve the issue does not offer them enough power for their extensions anyway. Like it or not, this is an indicator of a flakier relationship between the mothership and the community that has to be dealt with. Community does not need more goodies, it needs better integration and probably better support. Important detail: bugzilla and newsgroups are merely communication, and certainly not integration; and blogs sometimes, given the logorrhea of planet.mozilla.org, are a weak communication channel too.

  2. Freedom has many meanings, as we all know it. In the US, freedom of speech is almost total. Well, unless you start saying things that hurt your corporate employer for instance. In that case, freedom of speech often triggers the employer's freedom to let the employee go... Between the collapse of Netscape/the second birth of Mozilla and now, Mozillians (I mean employees here) have rarely, extremely rarely crossed the line. My old friend Tristan Nitot for instance always takes extreme attention to what he posts on his blog or even says in public because the distinction between perso and pro is never obvious. This is not easy to live like that when you work in a Free/OpenSource domain and I am the first one hit by the issue. My blog is my open diary, and I'm used to publish here exactly what's on my mind. But there's a big difference between my little, so little person and the Director of Firefox though: my visibility is low, very low. I don't represent officially an application used by hundreds of millions of people around the whole world and the whole world tech press is not all ears/eyes when I say or write something. Should the Director of Firefox be able to say in public, in the name of his personal freedom of speech, he/she finds brother product Thunderbird's UI offending and ugly and he/she stopped using it? Yes, he/she should be able to do it. But he/she should also be able to never forget again contributors are free to stop contributing. A balance of powers, clearly.

  3. Of course, given his visibility, his new position and the tone of his words, the whole world tech press has repeated ad nauseum Asa's words. Even if I think there's a serious strategic hole at the root of the problem, miscommunication did not help, to say the least. And miscommunication was created by misunderstanding and manichaeism. Nobody ever said Mozilla has to address all corporate needs, or has to provide ridiculous requests (Shaver said on his blog "one tab only", it's an excellent example) with a solution. As I said somewhere else, there is a large range of possibilities between "make corporations' lives easier" and "make corporations' lives harder". Given the world-wide reaction, I think it is now clear that the faster release process with fast EOLs causes the latter. Shaver and Sullivan said conversations with enterprises are ongoing. Better late than never. Don't forget governments, governmental or international orgs and military, please.

  4. The Community does not make any kind of difference between corporate and individual users. Their involvement in Mozilla is for all users, wherever they are, at work or not, in all countries, whatever the religion, race, colour, gender or opinions. It has to be understood that decisions that leave a given class of legitimate users on the side of the road - whatever the reason - are often seen as destructive by the Community. Hey, that's the weak side of OSS: you benefit from the good bits and have to live with the bad bits.

  5. The whole story reminds me of the "Get rid of the JS Console" frenzy a few years ago. The Community reacted loud and clear to that, and that decision was overturned. Now, a few years later, Firefox nightlies offers more and more developer tools and even more are coming. Firebug and the Developer panels of Chrome have shown that early adopters need web development environments that are not add-ons to the browser, maintained outside of the main stream, but built inside the main stream. Firebug is not distributed with Firefox probably because it induces footprint/performance hits but otherwise, it should be in. Please note this is why xulrunner-based app developers cry so much about Venkman: Chromebug has many issues in that environment and Venkman is most of the time broken, not really supported, and its feature set and severe bugs' list have not changed in years. Anyway, I wanted to mention the "Get rid of the JS Console" history because it seems to me (and it could be related to item 1 above and how the Community feels) that MoCo is more and more reacting as a microcosm. My discussions with the rest of community show an increasing number of "they're not listening, they say they are Mozilla and know" signals. MoCo, please, even if reading this you think this is a wrong reaction from the community, this is now far more than a few "trolls", it is factual and something must be done.

  6. There is a major difference of perception between Users/Community and MoCo about what is Mozilla and what should be its strategy. This is absolutely normal. MoCo has to deal with daily operations of a large and visible organization having a lot (yes, a lot, even if it's small compared to other big names of the industry) of employees distributed all around the planet. It has to secure revenue streams and make strategic choices. Even sometimes painful strategic choices. It's fighting for survival in a highly competitive landscape. Users and Community see Mozilla only as the browsing alternative, the freedom of browsing choice, the ability to get rid of IE handcuffs in corporate environments, and the cool kid around the block allowing a well-customized browser on all platforms. For them, Mozilla is the expression of its Manifesto, period.

  7. That feeling does not touch only the users or the Community. It apparently touches MoCo too. Reading #developers' monday IRC logs, the first word that came to my mind was "mutiny"...

  8. Users are versatile. Show them something cool and they'll adopt it. Both at work and at home. IE10 will be cool. Fast and reliable. Given the specs Microsoft is pushing inside the CSS Working Group (grid layout, templates, flexing, etc.), it is clear that standards-based apps and add-ons are in sight. Fear the revenge of the blue e... Those who think corporations will never have any incentive to go back to IE if they left it are in my opinion plain wrong. Those who think the only competitors to Firefox are now WebKit-based browsers are also plain wrong.

It's an old couple needing just a bit of couple therapy. Nothing more but nothing less. They still love each other, no worries.

Now, back to code.

Comments

1. On Wednesday 29 June 2011, 18:11 by iacchi

Man, you should do this more often. One thing I unterstood: you're defenitely better than me in explaining this matter, and we need to talk about this right now.

2. On Wednesday 29 June 2011, 19:01 by Paul

Great post. Here are some more thoughts from the community.
http://arstechnica.com/business/new...
http://www.reddit.com/r/technology/...
http://news.slashdot.org/story/11/0...
http://news.slashdot.org/story/11/0...
http://news.ycombinator.com/item?id...

3. On Wednesday 29 June 2011, 19:42 by Tim

Well put.

I think splitting Firefox users into enterprise or home has a lot of destructive potential.

4. On Wednesday 29 June 2011, 20:22 by Alan Gresley

Well stated Daniel. I do note that my Aurora has auto updated to the future Firefox 6 and Firefox 7 nightlies are available for testing. Very exciting.

5. On Wednesday 29 June 2011, 22:05 by Kees Grinwis

In my opinion Mozilla should offer long time support (LTS) versions of Mozilla Firefox (and Thunderbird), support for those versions should be about 2 years (the same as the old release cycle), this simplifies the task of the IT department. When a company wants a longer LTS cycle for a certain version, then they should be able to pay for this extended support.

On the other hand, as a individual user I'm happy with the new release cycle, new features are delivered faster so the long wait for new features is over.

Although I'm not (currently) using XulRunner: In my opinion XulRunner should be treated the same as 'enterprise' Firefox, so that embedders would know that their platform is supported for a certain amount of time. Ideally this would be longer than 2 years, however 2 years would be a nice base - again when embedders want longer support, they should be able to purchase this...

6. On Wednesday 29 June 2011, 22:33 by RNiK

Chapeaux!

7. On Wednesday 29 June 2011, 23:33 by Havvy

Can you give a link to Monday's #developer log?

8. On Thursday 30 June 2011, 00:17 by Danny Moules

"I also got messages from add-on developers (XUL add-on developers I should say) saying they're fed up with the add-on update process"

At the risk of sounding like a broken record on the AMO review process:

It's easier to get patches into Fx itself than it is to get a similarly complex add-on reviewed. It's security theatre since the people who are doing the reviewing aren't looking for real security flaws, they're reading off a checklist. How many are actually software security analysts? And if I try to put the code directly into Firefox as a patch, where's the security reviewer? Are add-on developers somehow less trustworthy than Firefox developers? I can't imagine how when they're the same people. Yet AMO decided they needed to protect 'brand image' and if that means being irrational and not consulting with anybody then so be it.

Of course, this doesn't even touch on the question of how anyone can communicate these concerns. No doubt someone will tell me I should take the complaints to #irc. Then someone from the forum will say it belongs in the forum. Or moz.Groups. Or Bugzilla. Or SUMO. Or private emails to AMO. Or privates emails to you. Or my blog. Or Planet. Or Slashdot. Or the Guardian. Or a Mozilla conference somewhere in Bulgaria. Or a sly comment somewhere in some Fx source code ('//This would never get past AMO review'). Or a bounty to a consultant to lobby on my behalf. Or all the above. Or the 'relevant subset' of the above.

I'd also like to suggest that certain high level 'vision' briefings, which shall go unnamed, only serve to dilute and detract from Mozilla's mission. That's the vision we should be working towards... not the flavour of the month ideas from Mozilla managers.

When I look at Wikipedia, another community I admire, I think of so many things it does perfectly which the Mozilla community haven't even got the most basic grasp of... and it makes me rather sad. Not that it's all sunshine and smiles over there of course - but there's plenty of scope for improvement in this community.

9. On Thursday 30 June 2011, 00:21 by Danny Moules

Hmm I like this quote: http://arstechnica.com/business/new...

"As a soldier, I absolutely can't allow any device on my network that hasn't been thoroughly tested. People's lives depend on security and 6 weeks isn't enough time for my shop to verify that FF is safe to use on government PCs."

Working full-time on medical software in a role that's heavily tipped towards security, I cannot agree more.

10. On Thursday 30 June 2011, 04:46 by Dan

Ass Dotzler just has too much power. He has tons of power in determining technical decisions but AFAIK he never studied CS or anything similar- how much programming experience does he even have? He's ecstatic about silly 1px cosmetic differences in FF5 ( http://weblogs.mozillazine.org/asa/... ) and ignores painful realities about FF internals. We have to remember that the guy is a designer, marketer, QA organizer, etc but not an engineer, a programmer, etc. His visions about what Mozilla and the Web should not bear anywhere near the weight of people like Glazman, gerv, Brendan, etc who have grappled with and worked to improve the Web and its standards.

The recent debacles with Thunderbird and enterprise show that he's a loose cannon who is doing Mozilla more harm than good.

11. On Thursday 30 June 2011, 05:06 by James

I've followed the Mozilla project long enough to see quite a few of these kinds of public arguments. There was the time when the Mozilla higher ups decided to make Gecko 1.7 the long term branch for the Firefox 1.0 series without telling the platform team, the time they decided to discontinue the Seamonkey app suite before the 1.8 release (though its developers had been working on it), and the time they decided anything not needed for Firefox couldn't even be included in mozilla central (forcing the creation of the separate comm central and the constant porting of patches into it). And there were similar debates over the futures of Thunderbird and XULRunner. Too often, those in charge didn't seek community input and/or didn't publicize the decisions early or widely enough (e.g. "we've been talking about it on IRC for a while now"). They also didn't seem to understand why people would be so angry about those decisions. The unfortunate result was that many community members felt--not entirely wrongly--that Mozilla didn't care about anything other than Firefox or making it more flashy. It's a shame because you lose disillusioned community members, and it could have been largely avoided by communicating better, without much more effort on Mozilla's part.

12. On Thursday 30 June 2011, 06:15 by Mike Kaply

IRC log from Sunday:

http://people.mozilla.org/~mrbkap/e...

13. On Thursday 30 June 2011, 11:00 by Danny Moules

"16:09 < gavin> the mozilla community is far too large for you to suggest a single priority
16:10 < Asa> gavin: I disagree. I don't think we have the resources, even with our grand community, to focus on more than one thing. "

... I'll leave other people to pick this one apart.

14. On Thursday 30 June 2011, 12:28 by Wladimir Palant

@Danny Moules: Even though your comments on AMO are very much off topic, I will try to answer. I have been doing security reviews of the top add-ons on AMO in irregular intervals - with devastating results at first. However, it now got a lot harder to find significant security issues in the add-ons and it is clearly not because the add-on developers got a better idea about how to write add-ons (they didn't, on average the code quality is still horrible). But the review process clearly helps, it filters out submissions with the most blatant security vulnerabilities. In other words, whatever you might think about the review process, it definitely serves a purpose. It could be faster but AMO is unfortunately relying on community help here and is permanently shorthanded.

@Mike Kaply: Thank you for this link. It really start to look like Asa is the only one in support of "We don't want enterprises" solution while everybody else is busy looking for a real solution.

15. On Saturday 2 July 2011, 04:25 by skierpage

Mr. Kaply: So entertaining to watch senior engineers spending their Sunday trying to get Asa to be less of a complete energy-sapping, morale-destroying, undiplomatic, bull-headed, my-way-*and*-the-way-I-say-it dick! I thought it was the other way around in software companies.

16. On Monday 4 July 2011, 00:48 by Jorge Villalobos

@Danny Moules: Wladimir already addressed most of the points you raised about the review process, so I won't repeat them. I'd just like to point out that I find it very ironic that you complain there's now way to voice those arguments, while in the following paragraph you list the majority of possibilities, all of which I follow. Send me an email (jorge AT mozilla DOT com), post in the forum (https://forums.mozilla.org/addons/v...), look us up on irc (#amo-editors), send a message to our mailing list (amo-editors AT mozilla DOT org), or file a bug (Component: addons.mozilla.org). Any will work. Much more than commenting on a blog post that has little to no connection with your complaints.

Just to add something that is not off-topic: a large portion of add-on developers are fairly happy with the way we manage automatic compatibility bumps on AMO. A good chunk of our add-ons are upgraded automatically, so there's no additional effort involved for developers. Having said that, the (also large) portion of developers that need to upgrade their add-ons on every cycle have their work cut out for them. It's really unfortunate, and I'm sure that we will be losing many developers because of it. There aren't many willing to update their add-ons that often, specially not when it's not the main priority in their lives. On the other hand, most simple add-ons can be ported to the new SDK, so that will provide an avenue for those feeling the pressure of 6-week cycles.

While I'd like slower cycles (quarterly releases sounded better to me), it looks like the most compelling argument to support the enterprise is to have a LTS branch, which wouldn't help add-on developers much...