A practical solution to the Debian/Firefox problem?

While it’s interesting to debate who’s right and who’s wrong in the Debian/Firefox standoff, it’s doesn’t really get us any closer to a solution. While we were debating, though, “Peter” (sorry, I don’t have a last name) made a practical suggestion:

I think Debian’s stance should have been to distribute epiphany and move Firefox to the appropriate *verse repository.

That’s the best solution I’ve heard yet: Simply move Firefox to non-free (or drop it entirely) and make something else the default browser. That allows Debian to make its point without the destructive side effects of a fork (user confusion, extension incompatibility, fragmenting the userbase of the leading open source browser on the eve of IE 7, etc.). The end result is the same: Mozilla loses out on the additional users (at least until/unless it changes its policy), and Firefox is still there for those Debian users who want it. Best of all, it maximizes freedom—namely, the freedom of users to decide whether or not this is an issue worth switching over.

42 comments on “A practical solution to the Debian/Firefox problem?

  1. Joe Buck

    Certainly a switch to Epiphany is one option, but you say you fear fragmenting the userbase. Are you assuming that Debian users would choose not to run Epiphany and download Firefox instead? They can do the same with Iceweasel, otherwise the amount of fragmentation is the same.

    You are assuming that iceweasel would cause extension incompatibilty. I don’t see why that would happen: the initial version would have identical source code, replacing only the artwork and modifying the User-Agent string in a way that still advertises the browser as Firefox-compatible.

  2. wjl

    Ian,

    I – like others – suggested just the same a while ago at http://blog.thedebianuser.org/?p=161 and in the comments of Eric’s http://ze-dinosaur.livejournal.com/12083.html?view=31027#t31027 post.

    And I can understand why the echo on this wasn’t so great, because the licensing/branding thing is not the only issue with the Mozilla Corp.

    Far worse – at least for Debian or anyone claiming “Long Term Support” – is that the guys over at Mozilla want to approve every bug fix before it’s distributed. That alone would still be an understandable point, business-wise and from their point of view.

    But that the Mozilla Corp. drops support for older versions after about half a year or so *completely*, and suggest users to upgrade to newer versions, and suggest to Debian to “bend the DFSG a little” is beyond what could lead to a discussion which is positive for all sides.

    So if Sarge still has a 1.0.x version, and backporting all these security fixes seems hard enough, they (we) would still run into something which Mozilla wants to approve, but doesn’t support anymore themselves? That leaves the involved Debian guys in a Catch-22 IMHO.

  3. Matt Brubeck

    I’ve thought of that too. There’s at least one downside to that solution: If Debian ships a branded Firefox in non-free, then Debian isn’t free to patch it however they want. This means that users of Firefox in Debian miss out on potentially useful patches (in particular, integration patches). The “Iceweasel” solution has other major downsides, but it does allow Debian and its derivatives to ship the browser with customizations that they feel benefit their users.

    (Debian could ship non-free Firefox without patches AND free Iceweasel with patches, but then you’re confusing users once again.)

  4. Np237

    Epiphany is already the default browser in Debian. However, for some reason I don’t understand, the debian-installer people feel it is necessary to install *both* epiphany and firefox by default.

  5. asac

    … this has definitely been considered by mozilla maintainers. The mozilla maintainers apparently just don’t want spend their time getting patches signed off et al for a piece of software that is in “non-free”.

    However, anyone who wants to maintain such a package can try to do so (good luck).

    I somehow agree that there might be some marginal benefit from a firefox package in non-free for some purpose (like sending out a signal), but I don’t agree that stopping to distribute the debranded firefox in debian would help anyone. Its free software, so lets package it and ship it.

  6. Kevin Krammer

    “…Epiphany is already the default browser in Debian…”

    I guess you wanted to write “is already the default browser on GNOME” :)

    Because I am pretty sure the default browser on KDE is Konqueror

  7. Ubuntu Tutorials

    I don’t see how this would solve the user confusion. It seems like it’d really split up the user base. Some people using Epiphany, others using Firefox–no real standard to support or develop.

    I think I prefer the IceWeasel solution. It’d still leave all of the users with one of the most popular browsers (underneath anyway), maintain compatibility with extensions, and continue to follow the FF development. I mean, FF will continue to be open so the developers can continue to take the builds, re-package with non-patented graphics & we’re all on the same page.

    I think the whole thing sounds like a big pissing match, but I agree with Debian. We need to stick to our principles and maintain our freedom.

  8. Np237

    Kevin, sure. This means that only for XFCE the default is firefox. GNOME and KDE users wouldn’t lose much by not having it installed by default.

    “Ubuntu Tutorials”, Epiphany uses Gecko underneath and is compatible with firefox extensions.

  9. Np237

    I should have been clearer; epiphany isn’t compatible with all firefox extensions, but it is with firefox plugins, and most popular extensions have their match in epiphany-extensions.

  10. Chris Cunningham

    As pointed out, shipping in non-free means accepting the trademark licence, which means that the current Firefox package _still_ can’t be shipped in Etch for technical, rather than legal reasons. And to be honest, I don’t think that spite (which is all that this would be) really accomplishes much, beyond making for harder work for those Debian derivatives that want to ship Firefox as the default browser (which is basically all the GNOME ones).

    – Chris

  11. Carlos

    MMhhh. It’s a pretty good idea, and now I remember somebody mentioning this on Eric’s blog.

    But I guess the problem with the security fixes remain. Talk about his opening Pandora’s Box.

    Would Mozilla waive the requirement of reviewing patches for a version of Firefox they are no longer mantaining?

  12. chele

    I like this solution. I think firefox.deb can sit in non-free, and simply be a script which downloads and installs the latest official firefox.tgz from mozilla.com

    No need to patch or maintain it seperatly. Everyone knows how secure and usable the tar ball distributed by mozilla.com is. The only bugs Debian needs to worry about are in the installation script. Everything else gets forwarded directly to Mozilla.

  13. luna

    Both soluion are not exclusive. You can have an unsupported Mozilla Firefox in non-free and IceWeasel in main.

  14. Andre

    Chele,

    Only 83% of Debian users are on the i386 architecture [1]. At the moment Mozilla upstream doesn’t ship a native Firefox binary for the other 17% of us. I don’t see how your non-free downloading script helps.

    [1] popcon.debian.org

  15. Michael M.

    A regular theme on the Epiphany mailing list is some user requesting this or that feature Epiphany lacks, and invariably the devs defer to Gnome settings to provide said feature. Which means that if you try to use Epiphany outside of Gnome, you’re likely to find it lacking. I think that makes sense from the POV of Epiphany/Gnome development, and I also think Epiphany is a excellent choice if you’re a Gnome user. But it doesn’t make Epiphany a very good substitute for Firefox outside of the Gnome desktop environment.

    As long as IceWeasel maintains compatitibility with Firefox extensions (many of which will not work with Ephipany), I don’t see a problem with switching. I guess everybody has their own take on the situation, but for me (an ordinary and appreciative user of Debian), I don’t care what the browser is called or what the logo looks like, I just want the same functionality I can get with Firefox on Windows and OS X. If I can’t get that, it’s time to look for a new distro. If I can, IceWeasel it is!

  16. Marjorie Roome

    The point to be made with switching to Epiphany is that it doesn’t solve the legal problem.

    Epiphany depends on Firefox, since it uses the functionality of the Gecko engine (and I suspect, other features too). When they did the recent 1.5.0.7 security update to Firefox it broke the nspluginwrapper bridge from Flash 7 to my 64-bit Epiphany.

    Of ocurse I’m pretty indifferent whether my background package is called Firefox or IceWeasel, but I still depend on it being there for Epiphany to work.

  17. Ian Murdock Post author

    Certainly a switch to Epiphany is one option, but you say you fear fragmenting the userbase. Are you assuming that Debian users would choose not to run Epiphany and download Firefox instead? They can do the same with Iceweasel, otherwise the amount of fragmentation is the same.

    I was actually referring to fragmenting the Firefox userbase. Whether you like Mozilla’s trademark policy or not, it is one of the leading (if not the leading) mass market example of open source/free software. On the eve of the closed source world’s response to Firefox’s success, the release of IE 7, it would be a shame to see the Firefox userbase splinter in two. -ian

  18. Ian Murdock Post author

    You are assuming that iceweasel would cause extension incompatibilty. I don’t see why that would happen: the initial version would have identical source code, replacing only the artwork and modifying the User-Agent string in a way that still advertises the browser as Firefox-compatible.

    Fair enough. Still, it should be a consideration, and if this was a simple matter of rebranding, I wouldn’t be as concerned. Given that the IceWeasel people are talking about adding features, though, it’s pretty easy to get off track from a compatibility point of view. -ian

  19. Ian Murdock Post author

    Far worse – at least for Debian or anyone claiming ‘Long Term Support’ – is that the guys over at Mozilla want to approve every bug fix before it’s distributed. That alone would still be an understandable point, business-wise and from their point of view.

    But that the Mozilla Corp. drops support for older versions after about half a year or so *completely*, and suggest users to upgrade to newer versions, and suggest to Debian to “bend the DFSG a little” is beyond what could lead to a discussion which is positive for all sides.

    I agree—the patch review requirement is unwieldy, and possibly unworkable. Still, isn’t this a problem that Red Hat and others have to face too? Why can’t Debian adopt the same solution? If it’s a matter of simply using the backported Red Hat patches, or of simply using what Mozilla gives you, then so be it. As long as it’s in “non-free”, I don’t see a problem, short of the problems the Mozilla policy imposes on you, and you can hardly take the blame for those.

    The thing that must be avoided at all costs, though, is this IceWeasel fork. I don’t think you are thinking broadly enough about how damaging this could potentially be to the larger effort. Have you ever heard the expression “cutting off your nose to spite your face” or the term “Pyrrhic victory“?

    -ian

  20. Ian Murdock Post author

    I’ve thought of that too. There’s at least one downside to that solution: If Debian ships a branded Firefox in non-free, then Debian isn’t free to patch it however they want. This means that users of Firefox in Debian miss out on potentially useful patches (in particular, integration patches). The “Iceweasel” solution has other major downsides, but it does allow Debian and its derivatives to ship the browser with customizations that they feel benefit their users.

    Fair enough. In my opinion, though, the downsides of the fork to the larger community (which are numerous) far outweigh the single benefit. -ian

  21. Ian Murdock Post author

    As pointed out, shipping in non-free means accepting the trademark licence, which means that the current Firefox package _still_ can’t be shipped in Etch for technical, rather than legal reasons. And to be honest, I don’t think that spite (which is all that this would be) really accomplishes much, beyond making for harder work for those Debian derivatives that want to ship Firefox as the default browser (which is basically all the GNOME ones).

    Go ahead and accept the trademark license and change the Firefox package to be compliant. I don’t understand why the result can’t go in non-free. Non-free packages are exempt from the usual licensing rules so long as they can still be distributed legally, which is certainly the case here. Or, at least, that’s how it worked when I created it. It’s not about spite, it’s about making your point without being destructive (you’re the ones who want to make a point, not me), and about giving users the choice of whether to use “non-free” software. -ian

  22. Ian Murdock Post author

    Epiphany depends on Firefox, since it uses the functionality of the Gecko engine (and I suspect, other features too).

    This presumably isn’t a problem, because there presumably isn’t a problem with having a package called “gecko” in main. -ian

  23. Np237

    Marjorie: “Epiphany depends on Firefox, since it uses the functionality of the Gecko engine (and I suspect, other features too).”

    Epiphany in Debian is built on top of xulrunner, not firefox. So far, MozCorp hasn’t asked anything wrt. patches that are applied to xulrunner, while I know for sure they are strongly disagreeing with these patches. Which means we can apply the set of patches we want to xulrunner.

    Ian: “I was actually referring to fragmenting the Firefox userbase.”

    Ian, I still don’t get your point. You are asking that we don’t fragment the firefox userbase, while proposing the use of epiphany which basically has the same consequence. Although more sophisticated, you can see it the same way as Iceweasel: a gecko-based browser with a different branding. I still don’t understand who gets hurt in this story, apart from MozCorp’s ego.

  24. Pingback: meneame.net

  25. Baptiste

    I’m sorry, but as a Debian end user, I wish to use the software I’m accustomed to, and in main if at all possible. I understand that as a business user, you want to avoid bad PR “at all costs”, but I don’t see why I should be sacrificed either.

    As Debian has little room to compromise without harming its users, I suggest you turn to the Mozilla Corp. After all, they are the ones responsible for Firefox PR, and they could solve the all thing in a minute: simply return to the compromise that was brokered by Gervase Markham less than one year ago.

    While the IceWeasel pun may not be the best choice for a new name, it has been floating around for a long time, so the Mozilla guys should have seen this coming (and they could have quietly asked Eric Dorland not to choose it). I guess they were so sure that Debian would have to accept their conditions, that they didn’t even think of the consequences if their plan didn’t work out…

  26. Forest

    Why not move Firefox to non-free but still include IceWeasel? I personally don’t like the Epiphany web browser, but I would still like to use a browser that is truly free, and I do think that Mozilla’s abuse of their trademark makes their product non-free. One problem I do see with putting Firefox in non-free is that somebody still has to maintain it, and that’s a lot of work for a program that we are basically suggesting people should not use.

  27. Carlos

    “Why not move Firefox to non-free but still include IceWeasel?”

    It seems to me this would unfairly burden the Firefox/IceWeasel Debian packagers, as they would have to deal with duplicate bugs and in practice have to work a lot more than necessary.

    Ian is quite right to worry about the fragmentation of the Firefox user base. But it seems to me that Debian and Ubuntu will only be the beggining of MoCo’s struggle to protect their trademark. It’s not too suprising that the only distributions that seem to be going along with their policy are big commercial companies. There’s nothing wrong with that, of course, but I wonder how many of the community-only distros like Slackware, Gentoo will respond to MoCo’s requests.

    I suspect that even without the DFSG’s issues, the requirement to drop support for a given version of Firefox/Thuderbird after six months (or whatever the official support of mozilla is) is a show stopper for many distributions. If Mozilla were to meet Debian mid-way on this, maybe moving Firefox to non-free, would look way more attractive.

  28. Luis

    I might have missed it, but I didn’t hear anything about the money that Firefox produces and I was wondering if it might have something to do with it all.

    As far as I know, Mozilla gets most of its money from Google for including the google searchbox in the browser. So when IceWeasel is released, where will the money from IceWeasel go, to Mozilla or to Debian? And what if they (Debian) decide to change the default google searchbox to a yahoo searchbox because yahoo makes them a good offer? Is this possible?

  29. Rune Baggetun

    If IceWeasel is what is needed for Debian to keep its freedom, then IceWeasel is welcomed. I am saying this as a Debian user. But IceWeasel should not be incompatible with Firefox (extensions, plugins etc).

    I see IceWeasel as an attempt to really make a good gecko based Linux browser (as I understand it most Firefox developers today are Windows fans).

    When it comes to the trademark issues. This is a Mozilla made problem and not a Debian made problem. Address Mozilla about this and not Debian.

    (Why do you need my e-mail in order for me to reply to your posting ?)

  30. Pingback: ProgProg : Ian on Debian vs Mozilla

  31. Pingback: defmay { Debian y Mozilla }

  32. Pingback: Le blog de LostInBrittany

  33. Bernd

    >I agree—the patch review requirement is unwieldy, and possibly unworkable.

    I can’t understand why debian does not use the opportunity to get a real review on its patches. Real review increases quality. Or are you afraid to get a r- bzbarsky? As http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=354622;msg=122 indicates the patches that are applied have quality issues and would definitely benefit from a review. More than this, I have seen Alexander Sack trying to back port patches that I committed and he had enormous difficulties as the code base did change a lot in the mean time and I could only guide him a little bit. I would say there are only a few people who are really capable to review these changes. My guess is that for core gecko changes the chances are > 90% that they have Mozilla CVS write permission.

  34. Glanz

    Matt Brubeck: “This means that users of Firefox in Debian miss out on potentially useful patches (in particular, integration patches).”

    I know this isn’t in the recent Debian culture, but have you thought of getting user input before deciding? If I were a negative person, I would interpret that statement to be a euphemism for, “This means that users of Firefox in Debian miss out on potentially glitchy patches.”

    Now Debian developers keep citing “Software freedom”. True. Freedom for developers, not normal users.

  35. Vi

    Was researching Linux distros for W2K replacement. Debian is out of question thanks to the IceWeasel nonsense. Since when freedom is synonymous to stupidity?

  36. Mark Brown

    Bernd: The problem with the review isn’t a lack of willingness to get review done (AIUI the vast bulk of the current patches have either been submitted or are pulled from Mozilla). The problem is the requirement for signoff before they can be deployed. This means that if Mozilla has no interest in reviewing the patch Debian is dead in the water; without Mozilla nothing can be touched.

    Given that, for example, Debian supports a much wider range of architectures that MoCo, uses a different GCC version and isn’t enthusiastic about introducing a major new release of FireFox to a stable Debian release the signoff requirement is a big problem even ignoring any freeness issues that it entails.

  37. Bernd

    >This means that if Mozilla has no interest in reviewing the patch Debian is dead in the >water; without Mozilla nothing can be touched.
    This is a straw man. It opens a scenario of a bad Mozilla organization that tries to dry out debian. They have better things to do. In contrary MConnor asked for separate reviewable patches, debian does not deliver. Other vendors supply patches and get them reviewed. And that is not hypothetical. I think that fixes should be upstream.

  38. Jed

    Do people not realize that issues like Debian using a different GCC version and these un-reviewed patches cause HUGE headaches for Extension developers and Third-party developers like myself?

    I’m a debian user, but the last thing on my list of things to do will be to check and debug compatibility issues with iceweasle for my extensions and third-party plugins.

    Same issue applies now to debian’s current ‘firefox’, hence why MoCo is doing this.

    A shame really because finally a OS app is getting mainstream enough that loads of companies are now porting their services and plugins to it, effectively making people less and less dependent on windows.

  39. Erin

    > Was researching Linux distros for W2K replacement.
    > Debian is out of question thanks to the IceWeasel nonsense.
    > Since when freedom is synonymous to stupidity?

    Choosing ideals over convenience isn’t stupidity. It is a stand, and Debian is supposed to be about adhering to a set of principles based around the Open Source Definition.

    But I see both sides; Mozilla makes its money by assuring that you will have a certain kind of experience when you see software with its trademarked symbols. Debian was benefiting from Mozilla’s popularity, but also using it’s own popularity to pressure the legal boundaries that Mozilla plays by. When Mozilla decided it had the popularity to challenge Debian on it, Mozilla surely did it only when it could win — because it understood that Debian’s continued existence is based on it’s own proclamation to a different set of “laws.” There was only one choice Debian could make in keeping with it’s proclaimed commitments, so there is no need to be mad at Debian or Mozilla, in my mind.

    Debian will pay a price. But so what. In the big picture, the clash is between the two systems of law, one emphasizing the proprietary, one emphasizing the urge to keep things non-proprietary. Debian will just lose a little of the extra attraction of having a big famous recognizable piece of software on it’s platform. Question is: should it have been there in the first place? I think so. Debian was pushing the boundaries of both sets of laws, and it benefited from that by extending its user base. Now it is just paying a small price; I doubt that many people will actually leave Debian because of this issue. It has just lost one of it’s better selling points.

    To be fair, the Mozilla move was also just a little bit “Microsoftesque.” In the abstract sense that Mozilla, like any good profit-loving company, was using it’s popularity to exert control over something external that it felt was diminishing it’s profit.

    Not that that’s wrong, but if Debian can’t or won’t stick to it’s principles when directly challenged, what the hell is the point of it?

    I’m not worried that Mozilla will ever be a Microsoft. But then, aren’t we all hoping that one day Microsoft will be “just another Mozilla”? It will one day be, I think — and largely thanks to the tide of Open Software that Debian (among other things) rides on. In the software industry, that tide has proven a natural antidote to monopoly. Sort of like gravity, it appears weak in isolation — but has an uncanny cooperative ability that ultimately makes it very powerful.

    In the end, both sides — the get-rich incentive that proprietary offers and the informational “tidal power” that sharing offers — are needed for the balance that leads to ultimate efficiency.

  40. Pingback: Le blog de LostInBrittany

Comments are closed.