The Firefox review deficit

Another week, another batch of review requests, another few that I didn’t get to added to the steadily growing pile of rotting patches.

For all of the rhetoric about building a community of hackers around a strong core group, we really suck as a group at bringing in the community involvement. Adding more peers isn’t the answer, since I can’t think of anyone waiting in the wings that’s sufficiently in sync enough to be brought in as a peer. But we’re in a situation where we need to address the review process, or we’ll just end up frustrating and alienating the people we’re supposed to be attracting.

Seemingly, I’m the most responsive peer for Firefox reviews, which presents a problem for me. Every hour I spend on Firefox instead of my contract work costs me money. My time is of course constrained by family responsibilities and the desire to have some replica of time not spent thinking about code or application UI (and of course, health concerns) . So I’m finding that I’m stuck between actually reviewing code, or making money. Usually, work wins out, but at this stage, I don’t know how much longer I’ll be able to justify making less money to keep stuff moving into the tree.

This isn’t to say that everyone else is slacking. Ben just landed the massive prefwindow V changes, plus whatever he’s really doing at Google. Blake’s in school and starting a company. Vlad and bryner are occupied by their employers’ respective Mozilla priorities. But the absence of someone with the time and patience to weed through and be responsive to review requests, along with possibly having the time to fix more things as reported (just on a day-to-day polish level) seems to be a growing problem. No one really has time to focus on the details, on the niggling little things that are so important to the polished feel of an app. Sooner or later, it will bite us on the ass.


  1. Doron says:

    I hate asking review from you for that reason 🙂

    I wonder, if Seamonkey moved to toolkit, would xpfe peers be added to toolkit peers, such as Neil? That would help.

  2. Yusuf Goolamabbas says:

    Would something similar to Poul Henning Kemp’s fund raising drive help ?
    Should Mozilla Foundation consider hiring you for a specific period with the sole purpose of working on reviews ?

  3. michaell says:

    Previous comment looks to me to have hit the nail on the head.

    If extra peers are not likely to be a feasible solution, at least in the short term, then either some funding needs to come from somewhere else, or the Foundation needs to pay you and/or Blake to do some more of this stuff, or Ben needs to spend less time on big stuff and more time on reviews and polishing (which is probably not the best thing as it means redirecting existing time rather than increasing available time…)

  4. IMHO we need more reviewers. If that means putting mconnor under the MoFo umbrella… that’s great.

    We need people like Ben doing big stuff (since he’s one of the few who have the time/resources to undertake it… not many others can).

    Perhaps it’s time to rank reviews. For example, trivial patches don’t need as much experience to check over. Perhaps that way some people can sift and remove the stupid stuff… leaving those with much more experience to review the larger patches.

    Either way, the only way to continue development… is to keep the developers developing.

    The other option is to create a ratio rule:

    For every new feature/enhancement/big thing, you must first:
    1. Fix 3 bugs
    2. Review 5 patches

    That way there’s a little bit of balancing done. Might not hurt. I’ve heard of companies doing “50% of patches are bug fixes” and increasing that requirement as the product closes in on release.

    Perhaps that’s a better way to negotiate resources?

    Just some sudo-random thoughts

  5. Mike says:


    Its a good idea, but Its tough to dictate what people work on in the OSS world if they’re working for free. I do tend to adhere to “the devil is in the details” as applied to software, so I try to keep up with the correctness-type fixes, along with broken/substandard UI bits. But most unpaid OSS hackers seem to prefer “cool” things.

    It’d be interesting if MoFo hired someone to help maintain Firefox. We have a big picture/big feature guy in Ben, who always finds ways to innovate, and I don’t believe his time is best spent doing maintenance-type fixes. Given that MoFo has four or five paid QA people testing builds daily, it seems an odd concept that there’s no one to take/fix regressions in one of the premier apps on a day-to-day basis. At the same time, I can’t imagine pulling dbaron or jst or brendan into the Firefox UI hacking world. Their time, as with ben’s, is better spent elsewhere. I don’t know who the right person for the job is, either. Maybe Josh Aas will be pitching in on this, but no one’s said whether he’s going to be working on Firefox/Thunderbird, and I haven’t had enough interaction with him to even know how his UI instincts are as compared to the current Firefox/toolkit peers.

  6. Mike says:

    And fwiw, I’d love to work on Firefox full-time, but I’m not expecting or asking for it. I’m largely just putting the concern out there, in the hopes that someone comes up with a solution.

    We’re the leading edge of an OSS push into the mainstream, and we need to match Microsoft et al on pure quality of the product. We missed far too many polish bugs for 1.0, simply because we didn’t have the resources. Now they’re seeming even more scarce, and that’s concerning me enough to go public.

  7. alanjstr says:

    Rule number 1: cover your own ass.
    If MoFo isn’t willing to pay you, then you need to take care of yourself, your health, and your family.

    That said, MoFo needs to stop doing things like completely rewriting the prefs panel and start fixing bugs and reviewing patches instead. While features are always great to have, we keep creating new bugs instead of taking care of the existing ones.

    This isn’t UMO, where a complete rewrite is what is needed. And UMO is small enough that it can handle it. Firefox is not small.

    We have community involvement in contributing patches, but if they just sit around and bitrot it does no good.

  8. Mike says:

    Note that I’m not asking MoFo to pay me. I just don’t figure I can keep doing this effectively, so I’m trying to figure out a viable means to keep at least the current trickle going.

    Second, MoFo didn’t rewrite the prefwindow, Ben did, and it was necessary for OS X/GNOME compliance to do so. Considering that the roadmap called for HIG compliance fixes in 1.1, it had to happen. The rework was an added bonus of making a new preferences system.

  9. bernd says:

    It seems to me that the firefox model doesn’t scale well with reduced available time resources. There have been enough developers that warned about the downside of the ff review scheme, cutting the supply of new people. In a university one calles that the unity of research and education. There are enough good people, one issue is probably that they got the earlier message and write extensions now.
    While I completely agree with a limit who works on the frontend, I believe that toolkit is a core technology and should go under core review rules. This would give you access to a couple of more sr’s who should be able to balance the load. This will/should anyway happen if the suite becomes dependent on the new toolkit as it would replace the xpfe/toolkit.

  10. jhermans says:

    In my day-to-day work, I have to make a distinction between bug-requests and change-requests. The latter requires a careful examination of the product-management team before I’m allowed to start working on them : that can sometimes take several months and similar requests are often denied (when no apparent advantage is found). Bug-requests are much easier – you can start on your own if you find the time (although a priority-list is also maintained).

    Maybe we can something different for Firefox? Mofo has to decide which changes are worthwhile to implement. We can’t stop someone to write patches on non-decided change-requests ofcourse, but they can’t be accepted without permission. That isn’t much different from the current proces.

    Real bug-requests don’t require that much administration, ands such be accepted easier (s/r reviews still apply ofcourse).

  11. jhermans says:

    >Maybe we can something different for Firefox?

    That should be: Maybe we can use something similar for Firefox?

  12. I’ve got a better idea (I call it a “solution”, but enough with the wordage)…


    By taking a small sample of your DNA, we can make another mconnor (we’ll call him mconnor2). By doing this, we can double the output.

  13. R.J. Keller says:

    ahh great blog post Mike! I definitely agree. One of the things I did in Help Systems (at least until I got caught doing it :)) was checkin patches without review if they were trivial and eventually if someone showed skill, they were given permission to review larger documentation changes (in the case of Firefox trivial changes -> small bug fixes) and only new features or large bug fixes needed review. I had this system running for about a half year and got way more productivity in help systems than when it was under m.o control.

    I’m so happy you decided to respond on this problem, because I’m sure a lot of people will listen to you more than me. Good luck!


  14. Anonymous Coward says:

    Though it may sound cliche, thanks for all the work you do on this fantastic product. I have been in tough situations like yours, and sometimes a giant thank you really helps out. It still doesn’t pay the bills, but just know you are appreciated by the masses

  15. Tef says:

    hello I’m french and I have seen a news on but do you think really that Firefox 2.0 will be released on 2006 because there is not enough “hacker” to secure it, i would bring my knowledges to help you but I’m just a developper and if a have time to look source code of firefox why not !

  16. ss says:

    firefox should enhace the bookmark which always delete all archived sites, be sure, make bakcup of ur brookmark

  17. Carlos Collinao says:

    Hola a Todos. A veces los buenos productos tienen sus dificultades para su desarrollo.
    Tienes mi apoyo. Grs

  18. Jose says:

    Mozilla FireFox is in simple words…
    THE BEST…..
    In long words….
    just the best browser i ever had, is simple, and very, very powerful, the developer tools and lots of other extentions are very good and help to improve my sites a lot…



  19. hyperboole says:

    This is very sad news as I thought that Firefox had a more lively and devoted community. I guess I’ll have a go at reviewing some bug patches and writing a few. I guess you could put the donations link in bigger text. That’d get you a few quid. You should probably make this public to teh community and hope that this will get them more involved,

  20. musicplay

    The Firefox review def…

  21. Anonymous says:

    sheridan hotel…

  22. Anonymous says:

    istanbul hotel…

  23. Anonymous says:

    st martin hotel…

  24. Anonymous says:

    atlanta ga hotel…

  25. Anonymous says:

    romania hotel…

  26. Anonymous says:

    kensington hotel…

  27. paris hotel says:

    paris hotel…

  28. delano hotel says:

    delano hotel…

  29. jerusalem hotel…

  30. hello, nice day…

  31. Annuncio incontri personali….

    Annuncio incontri personali….

Leave a Reply