adam's grump of the day: icons in fonts (was Re: web-assets-httpd stuck in limbo?)

Nicolas Mailhot nicolas.mailhot at
Thu Jan 16 17:37:09 UTC 2014

Le Jeu 16 janvier 2014 18:13, Adam Williamson a écrit :
> On Thu, 2014-01-16 at 10:52 -0500, Matthias Clasen wrote:
>> On Wed, 2014-01-15 at 22:31 -0800, Adam Williamson wrote:
>> > On Wed, 2014-01-15 at 22:59 -0700, T.C. Hollingsworth wrote:
>> >
>> > > * Another individual thought that all web authors are stupid for
>> > > wanting to use fancy fonts and that I am wasting my time.  (He might
>> > > be right about that last bit... :-P)
>> >
>> > While we're doing asides, that one *does* get right on my nerves.
>> >
>> > If anyone overrides font choices in their browser config and wonders
>> why
>> > an increasing number of sites - inc. github, and the wordpress admin
>> > interface - seem to display weird hieroglyphs all over the place, it's
>> > because of this "clever trick": web designers have decided that it's a
>> > really good idea to abuse font rendering engines as a way to render
>> > icons, and starting shipping icons as made-up Unicode codepoints in
>> > their sites' custom fonts. If you override their font choice, then of
>> > course these icons wind up as garbage, because your font does not have
>> > them, because ICONS AREN'T FUCKING TEXT CHARACTERS, web designers.
>> It makes a lot of sense, actually. At least the symbolic icons that have
>> become prevalent in our uis share a lot of characteristics with text,
>> and can benefit from getting the same treatment as glyphs.
> Then get the Unicode people sign off on treating your special special
> characters as text, otherwise it's just an abuse of the system.

And, actually, Microsoft did exactly that : its special UI symbols (in
Segoe UI) have all standard codepoints, you can use the
Microsoft font with other apps and you can make apps written around the
Microsoft font work with FLOSS fonts that implement those codepoints.

Not-unicode is only needed for vanity symbols like company logos would never accept.

Nicolas Mailhot

More information about the devel mailing list