262394 - Guidelines tell people to use ISO-8859-1 instead of UTF-8

Status: RESOLVED INVALID

(www.mozilla.org :: General, defect)

Description

[Dave Miller [:justdave]]

<quote>
Use the ISO-8859-1 aka. Latin-1 character encoding. Represent characters that
are outside the ISO-8859-1 repertoire (eg. typographically correct dashes and
quotes) as decimal numeric character references. For example, encode the em dash
(—) as &#8212;.
</quote>

Shouldn't we be telling people to use UTF-8?

Comment 1

[fantasai]

Attachment: patch

Comment 2

[fantasai]

justdave, do these changes look good to you?

Comment 3

[Dave Miller [:justdave]]

> justdave, do these changes look good to you?

No, these instructions imply that ISO-8859-1 pages are UTF-8 compatible.  They
aren't.  Pages which are completely ASCII are UTF-8 compatible.  ISO-8859-1
numerical entities are UTF-8 compatible, because the ISO-8859-1 character codes
are a subset of Unicode (which is what we were discussing on IRC).  However, the
ISO-8859-1 characters with codes above 127 are encoded differently in ISO-8859-1
and UTF-8.  All ISO-8859-1 characters can be converted to UTF-16 by adding a
zero byte in front of each character.  However, UTF-8 only allows 7 bits in the
first byte of a code (reserving the 8th bit for a continuation flag which
indicates that more bytes follow).  This means an ISO-8859-1 character with a
code above 127 requires 2 bytes to express in UTF-8.

Comment 4

[Anne (:annevk)]

When bug 154570 is fixed, we have to fix this. Before that, it would be quite
impossible. Although we could advice people to use US-ASCII as charset and use
character references for all characters that don't fit...

Comment 5

[Christian :Biesinger (don't email me, ping me on IRC)]

> ISO-8859-1
> numerical entities are UTF-8 compatible, because the ISO-8859-1 character codes
> are a subset of Unicode

aren't numerical character references always a unicode reference?

Comment 6

[Anne (:annevk)]

Yes. Every character reference is encoding independent.

Eventually we should just switch to UTF-8 for a day or so and see what breaks
and fix that.

Comment 7

[Gordon P. Hemsley [:GPHemsley]]

Given the age of this bug and that page, I think we can probably just archive that page and move its replacement contents to one of many wikis, if necessary.

Comment 8

[David Boswell]

(In reply to comment #7)
> Given the age of this bug and that page, I think we can probably just
> archive that page and move its replacement contents to one of many wikis, if
> necessary.

Makes sense.  Looks like there's a bug open to archive that already.  Will address soon.

Comment 9

[do not use]

Closing as part of our bug triage process.  This page now directs people to the style guide.  http://www.mozilla.org/en-US/styleguide/

For more about the bug triage process, please see https://blog.mozilla.org/websites/2012/11/15/mozilla-org-bug-triage-process/