Precedent: We removed 'Windows; ' already, and there's bug on removing 'Macintosh; '. Knowing that the user is on X11 isn't especially helpful given that you know it's Linux already. The downside here is breaking sniffers that search for the entire 'X11; Linux' string. Need UA's for other browsers here.

To carry over some discussion from bug 581783, sniffers may just look for X11 to identify any *nix dialect (i.e, all UNIX variants use this token). Thus, it may serve as a secondary identification for similar platforms.

Fair point. What purpose do you think they have in doing so? If it's just for providing appropriate download binaries, 'X11' isn't terribly useful. One needs to know the specific arch. Support might be a different story, but I don't think that's a convincing enough case. Can you think of others?

Line-endings? I did something similar with a custom Gopher server in the early 1990s, including character-set selection in those pre-Unicode days. I think we stopped doing it in 1996, ahem.

Do any other browsers on Linux and Unixy OSes include X11 in their UA strings? /be

Mozilla/5.0 (X11; U; Linux i686; en-US) AppleWebKit/534.3 (KHTML, like Gecko) Chrome/6.0.462.0 Safari/534.3 Mozilla/5.0 (X11; U; Linux x86_64; en-US) AppleWebKit/533.4 (KHTML, like Gecko) Chrome/5.0.375.99 Safari/533.4 Opera/9.80 (X11; Linux i686; U; en) Presto/2.5.27 Version/10.60 Opera/9.62 (X11; Linux i686; U; en) Presto/2.1.1 Opera/8.54 (X11; Linux i686; U; de) Opera/7.54 (X11; Linux i686; U) [en] Mozilla/5.0 (compatible; Konqueror/4.4; Linux) KHTML/4.4.1 (like Gecko) Fedora/4.4.1-1.fc12 Everyone else does. Doesn't tell us what would break by removing it, though.

Ironically, removing it will allow us to be fingerprinted as Firefox. (I'll shut up, now)

(In reply to comment #5) > Everyone else does. Doesn't tell us what would break by removing it, though. (Er, Konqueror doesn't.)

(In reply to comment #6) > Ironically, removing it will allow us to be fingerprinted as Firefox. Several other aspects of our UA string will allow us to be fingerprinted as Firefox.

(In reply to comment #2) > If it's just for providing appropriate download binaries, 'X11' isn't terribly > useful. One needs to know the specific arch. Support might be a different > story, but I don't think that's a convincing enough case. For binaries one indeed would need to know the exact target platform. I can only think of distributing non-binaries like scripts, where the OS family is relevant without needing the exact OS variant to serve the correct version. The X11 would have only supplementary value for support, whoever is providing help should recognize the main OS. On the other hand, it may help when some rather uncommon OS is used to recognize it as a *nix variant.

Opera/9.24 (X11; SunOS i86pc; U; en) Mozilla/5.0 (X11; U; SunOS sun4u; en-US; rv:1.9b5) Gecko/2008032620 Firefox/3.0b5

I should mention that, for Fennec, Android shows up as '(Android; Linux ...; rv:...)'. The 'Android' bit is definitely useful there; it'll be different for N900 too. I think this is WONTFIX.

(In reply to comment #8) > (In reply to comment #6) > > Ironically, removing it will allow us to be fingerprinted as Firefox. > > Several other aspects of our UA string will allow us to be fingerprinted as > Firefox. Namely the word Firefox. :) (In reply to comment #11) > I should mention that, for Fennec, Android shows up as '(Android; Linux ...; > rv:...)'. > > The 'Android' bit is definitely useful there; it'll be different for N900 too. > > I think this is WONTFIX. What's wrong with ditching the "X11" but keeping the leading "Android" as needed?

More concerted proposal in bug 589353.