Won’t Someone Think of the Add-on Developers?

I have been biting my tongue watching all the changes that add-on developers are being required to make for Firefox 4, but the various theme changes that are going in which are going to cause toolbar icons to be scaled up and/or down has put me over the edge. (See bug 583231). A lot of work is going to be required by add-on developers (and artists and others) to work around these and other issues. It doesn’t help that these Firefox “betas” are no where near beta quality, and involve major changes with each beta update, which means add-on developers are having to do rework with every “beta.”

I emailed folks from the AMO team about this and they indicated that their role is only to communicate what add-on developers need to fix, not to work with the Firefox team in any way to represent add-on developers.

Update:That sentence above did not properly represent the email conversation. I stand by my assertion though – the AMO team does not proactively work with the Firefox team to look out for the needs of addon-developers – they communicate information after the fact. They are not on the front line saying “How will this affect add-on developers?”

So there is no one involved in the Firefox 4 process that is advocating for Firefox add-on developers. We’re just expected to “roll with the punches.” I understand that we’re just consumers of the platform, but add-ons have built up a great deal of brand equity for Firefox.

I find it quite short sighted of the Firefox team as a whole to alienate add-on developers this much. I’m aware that this a major release, but there should be more and better messaging around these changes. And they could certainly be batched better instead of spreading them across every beta.

So please, think of the add-on developers…

Please note: I reserve the right to delete comments that are offensive or off-topic.

Leave a Reply

Your email address will not be published. Required fields are marked *

21 thoughts on “Won’t Someone Think of the Add-on Developers?

  1. And with the add-on developers, the users as well. I am probably not the only one seeing the huge number of arbitrary version incompatibilities and how it harms the Firefox experience (when you rely upon add-ons…).

    You speak wise words, and I hope someone else realizes and does something about it, after all add-ons are one of the superpowers of Firefox! (and partly what makes it stand its ground against the fierce competition)

  2. Well, as always, you should wait til the code freeze to start developing for Firefox 4. Doing it on the go will only warrant constant rewrites, it’s to be expected. There’s really no need to develop for the first betas, I think.

    I feel for you, but changes have to be made. The fact that they will come all in Firefox 4 is a good thing, not a bad thing. Imagine if they’d have came in Firefox 4.0, then a few others in 4.1, then a few others again in 4.2, that’d be much worse.

  3. > Well, as always, you should wait til the code freeze to start developing for Firefox 4. Doing it on the go will only warrant constant rewrites, it’s to be expected. There’s really no need to develop for the first betas, I think.

    What code freeze? There is a “code freeze” for each beta. The final code freeze is way off, but we are still being told to update our add-ons now, even though they break from beta to beta. AMO requires add-ons to work with the latest Firefox beta in order to be recommended add-ons.

  4. For years it has been known and planned that Firefox 4 / Gecko 2 was going to have massive API/ABI breaking changes. So nothing unexpected there.

    Mozilla also has not changed the way they handle the beta releases. Major features have always been added/modified between beta releases, and theme changes often don’t come until the mid – to later -beta releases. So again, nothing new there.

    This is often why extension developers wait to update their extensions until the RC phase.

    JetPack is meant to alleviate all of this, by abstracting all such changes behind a stable API. So it’s not that they aren’t thinking about the developers, there just slow in getting the ‘fix’ out.

    > AMO requires add-ons to work with the latest Firefox beta in order to be recommended add-ons.

    That’s an AMO policy and should be taken up with them, not the developers. I agree that it’s kind of stupid.

    Though I suppose this goes back to the age old argument of what’s better: incrementally keeping things up to date, or making such changes all at once.

  5. Ah, I didn’t know about that.

    If your business depends on your add-on being recommended, then it’s too serious a business for me, and I really have nothing to say. Otherwise, have you considered not caring about the “recommended” state? 😛 I know it’s a bit… wild, but well, if it’s like climbing a mountain with a rock only to throw it back down again (like that Greek dude of yester), I think it’s best just to leave the rock at the bottom of the slope.

    And yes, I do mean the final freeze. It’s scheduled to September the 1st, and some people think it won’t happen so soon, but as soon as bug 130078 lands, things should speed up a lot in terms of features landing, so I’m thinking it will be around that date that the final freeze happens. Betas are being released every two weeks, so that’s another reason not to care about early betas.

    Didn’t the same thing happen with 3.0? I know it wasn’t as drastic as 4.0, but there ware like 3 or 4 betas that were not feature complete too.

  6. I don’t really understand why you feel that incremental beta updates create more work for you. It’s good that you try to keep your addons compatible with trunk but I don’t think there’s a need to make your addon look perfectly in every beta. In fact, as you’ve seen, spending time on it isn’t really worth it; the theme is going to see quite a few updates on the way to Firefox 4 final and in some cases (like with the iconsize issue) we don’t even know yet what the final solution is going to look like.

  7. @admin: We’re not requiring any add-ons to be compatible with Firefox 4 betas in order to keep their recommended status (or apply for it). We understand that Firefox 4 betas are still very malleable, but we want developers to start paying attention to them.

  8. Mike, are you paraphrasing when you wrote this?

    “I emailed folks from the AMO team about this and they indicated that their role is only to communicate what add-on developers need to fix, not to work with the Firefox team in any way to represent add-on developers. ”

    It just doesn’t sound like something that the AMO team would say.

  9. A lot of developers for some reason don’t understand that for minor updates (not major versions like 4), that you can should mark your compatibility as 3.5.* or 3.6.* for example and then your extension won’t break for every tiny security patch.

  10. @rey:

    Totally paraphrasing. And probably got it wrong. From – http://blog.mozilla.com/addons/2010/08/03/add-ons-review-update-14/#comment-113802 – where I reposted what I asked:

    @Jorge,

    I did not interpret your response as an “I don’t care,” I interpreted it as “no.” I asked this question:

    > I’m just wondering if folks on your team are advocating at all for
    addon developers or just letting the changes happen?

    And your answer was more of a “letting the changes happen.”

    What I want to know is if the AMO team participates in Firefox meetings and says “How will that impact add-on developers?”

    Is anyone advocating for add-on developers, or are we on our own?

  11. Maybe we just need to adjust our expectations. FF 4.0b3 is going to ship with Bug 579621, a bug so severe it should stop even an alpha release. So we just need to chill out, and wait until a real beta comes along.

    One positive difference over previous releases is there seems to be little pressure from AMO or users to ship an addon compatible with the beta. Which is good because we can’t create them.

    Plus, this gives us some time to work on extensions for Google Chrome.

  12. Your sentiment that Firefox developers don’t take into account how changes are going to affect add-ons is wrong. This probably goes back to the misunderstanding that we couldn’t/wouldn’t take changes after realizing that they would affect add-ons.

  13. Michael, I don’t have the impression that the Firefox team “doesn’t care” about add-ons – they do but they simply have to make breaking changes in this release. For most part adapting the extension is easy. That toolbar icon issue is very annoying however, while the “old” icons will continue to work the scaling makes them look really ugly. And fixing this while keeping compatibility with old Firefox versions or other applications (SeaMonkey? Songbird?) is everything but trivial. I have been increasingly voicing my concerns here as well. But the theme/UI isn’t final yet and supposedly a solution is in the works. Maybe in Beta 4?

  14. The Firefox guys even paid only scant attention to the needs of other Mozilla products, so what would you expect? Around July 1, a change in the XPCOM registration process broke all other apps, many (but far from all) extensions, and made the addons manager ugly in all themes other than Default (and impractical in Default, but that’s to be expected considering the kiddie-toy look favoured by Firefox devs as opposed to the no-nonsense look of SeaMonkey including the Modern theme; but that’s just my own preference, maybe at almost 60 I’m already an old-fashioned dinosaur). Thunderbird and SeaMonkey took a week or two to recover, but Lightning (an extension, but produced by Mozilla itself) hasn’t had a single successful trunk build since June 29, despite automatically starting a new build attempt every hour or so. So, if even the “second-party” guys (non-Firefox Mozilla, I mean) are having a hard time “rolling with the punches”, what would you think can be expected in favour of third-party extensions?

  15. I don’t think UI or code changes are that much of a big deal. Sure it’s a pain in the butt, and sure it means taking the time out to make the changes. But in some ways that is what makes the add-ons for Firefox some of the most powerful and amazing add-ons for any piece of software. If Mozilla wasn’t constantly changing things, many authors would be tempted to say “and it was good” and just walk away. Heck, there are many who do anyway. But the ones who stick at it, who take the updates as opportunities to make changes they’ve been meaning to, and not just add support for the next version, that’s the spirit that we love about using and we love about developing for Firefox.

    Sure, there are moments when you think, “wow, that was just plain inconsiderate programming, I can’t believe they didn’t realize this would cause issues for us over here”, but I don’t think most of that is purposeful. The fact that none of the default themes are cross-platform compatible, despite cross-platform add-ons being one of the tauted benefits of Firefox, seems strange and ultimately inconsiderate to add-on developers, as well as to users to me. But I wouldn’t call that purposefully malicious.

    The breaking of compatibility between Personas and Themes does however feel very malicious. Especially considering that at the same time the “Get Themes” link the Add-Ons Manager was redirected to the Get Personas home-page, and Themes were removed from the AMO Recommended Add-Ons list and the Add-Ons Manager Recommended Add-Ons as well. There also appears to be no interest internally to fix swapping Themes without a restart, despite this being one of the reasons for claiming that Personas are superior to Themes, a claim that is still made months and months after ShareBird put together the SwitchThemes extension as a proof of concept. Those are the things that make me wonder if Mozilla even *wants* add-on developers to contribute.

  16. Didn’t realize it had gotten that bad. As consumer of add-ons, greatly appreciate the value some of them add. Especially appreciate importance of keeping core app and all plugins up to date for security reasons, given how dangerous the online world has become, so seems to me that core app team should be working VERY closely with add-on developers/maintainers to ensure that whole ecosystem is as safe and sane to use and maintain as possible.

    Thank you, Mike, for great work you’ve been doing, and sorry it’s become such a burden. Please hang in there.

  17. There also appears to be no interest internally to fix swapping Themes without a restart, despite this being one of the reasons for claiming that Personas are superior to Themes, a claim that is still made months and months after ShareBird put together the SwitchThemes extension as a proof of concept. Those are the things that make me wonder if Mozilla even *wants* add-on developers to contribute.

    To me it looks like history repeats itself.
    Once upon a time there was the Netscape corporation (making Netscape 4.7) and the Mozilla Foundation (making what would become Netscape 6); and when Netscape 6 came out, there were quite a number (well maybe a couple of handfuls) of exciting themes for download on its very frontpage.
    Then the Corporation took over Netscape 6, made Netscape 7, and finding themes for the latter was much less obvious.
    When Mozilla (the Foundation, at the time) put out Firefox and Thunderbird, they had a new Extension Manager, and lots and lots of extensions and themes (OK, maybe a few tens, but to me it seemed incredibly many by comparison). From time to time I would check them all, ranking them alphabetically at addons.mozilla.org, to make sure none had escaped my notice.
    Today, Firefox comes from the Mozilla Corporation, Thunderbird from the MozillaMessaging branch of the Foundation, Netscape has died out, but its Suite has been reborn as SeaMonkey, produced by a bunch of slightly crazy volunteers (the word “crazy” is a compliment under my pen) who do it in their free time, under the Foundation’s benevolent aegis. Each of these teams is headed by different people, reacting differently in the same circumstances, and having slightly different goals and aims, though of course the Mozilla Mission of Promoting Free Choice is supposed to govern them all. Then there are addons for all of them jointly or severally: the number of these addons has literally explosed compared to what it was in Netscape 6 time, so that anyone may pick and choose: I believe that this wealth of possible choices is a GoodThing™. Even a few (a negligible few by the numbers, but some of them of outstanding quality) are produced under the Mozilla trademark itself and distributed with some of the applications; but when you compare the three .zip (or the three .tar.bz2 or the three .dmg) you notice that the number of included addons is not the same: another symptom, I think, of the different attitudes of the different teams producing them.