UTF-8 banned from being default in Chrome, Firefox

* UTF-8 banned from being default in Chrome, Firefox
@ 2017-08-19  9:10 Nominal Animal
  2017-08-19 21:29 ` Andy Oram
  2017-08-19 23:05 ` J.B. Nicholson
  0 siblings, 2 replies; 6+ messages in thread
From: Nominal Animal @ 2017-08-19  9:10 UTC (permalink / raw)
  To: libreplanet-discuss

We've lost another freedom, choosing the default character set encoding
in Firefox and Chrome/Chromium browsers.

I only recently noticed that Firefox has banned UTF-8 as a default
character set encoding about four years ago.
Chrome followed suit this year.

(By banning, I mean you cannot set it as the default, not via the UI,
or by directly editing the user preferences: the setting is disallowed
in the sources. The only way to remove the ban is to recompile the
browsers from sources, applying a patch that removes the ban.)

This affects those web pages that do not declare a specific character 
set, like many mailing list archives; and local plain text files.

(list.gnu.org provides a Content-Type header with charset=UTF-8,
  which I do warmly commend the list admin! -- but for example
  marc.info does not.)

Both Chrome and Firefox developers insist that the *proper* default
character set is none other than Windows-1252, or some other
encoding (other than UTF-8) for non-Western users.

The system locale used by the user is ignored in all cases.

The underlying logic for these developers, as far as I could
ascertain from the bug reports and changelogs, was that
only "legacy" charsets should be selectable as a default,
and UTF-8 is the (only) non-legacy character set encoding.

In my opinion, that is a non-sequitur: those settings have
nothing to do with "legacy", and everything to do with "default".

I tried to raise this question with the developers involved
(particularly Henri Sivonen for Firefox), but I was blocked;
apparently these issues are already covered in the relevant
bugzillas (marked WONTFIX).

(Also, particularly Sivonen, likes to claim that choosing
  UTF-8 as the character set encoding if no encoding is specified
  is "unreasonable". On systems like GNU/Linux or GNU/Hurd, or
  even BSD variants, there often are no "legacy" content without
  character set definition, other than in UTF-8 -- this is particularly
  true in my case, including the servers I maintain --, so
  "unreasonable" seems to me be a disguise for something else.)

I was wondering whether this kind of removal of user choice
is something the LibrePlanet folks could help with?

I suspect publicity is the best disinfectant against such encroachment.

Best regards,
    Nominal Animal

    (I use a pseudonym in order to handle criticism better, as
     everything directed to "Nominal Animal" is directed at my
     expression/advice/opinions, not at my "social person".
     My physical identity is not a secret, and I can provide it
     if needed, but as mentioned, I function better under this
     pseudonym.)

_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss

^ permalink raw reply	[flat|nested] 6+ messages in thread