1 |
On Wed, 25 Mar 2009 23:35:41 -0700 |
2 |
"Kevin O'Gorman" <kogorman@×××××.com> wrote: |
3 |
|
4 |
> > After a bit of Googling, it seems the accepted solution is to use |
5 |
> > HTML entities for those symbols and not try to use the raw |
6 |
> > characters as you are attempting to do. |
7 |
> > |
8 |
> > http://en.wikipedia.org/wiki/List_of_XML_and_HTML_character_entity_references |
9 |
> > |
10 |
> > Does that contain all of the symbols you need? If there are any |
11 |
> > others, you should be able to use the unicode versions. |
12 |
> |
13 |
> Sigh. My stuff is not for a mass audience. I can expect them to |
14 |
> install a font, |
15 |
> and I'd really not like to be fooling with entities that much -- |
16 |
> composition is laborious. It's really annoying to me to have a font |
17 |
> on my own system that is inacessable through browser features that |
18 |
> were apparently designed to allow just that. |
19 |
> |
20 |
> And Unicode is a complete mystery to me. I see stuff come in and |
21 |
> display as it should, but as an author it's just something I've never |
22 |
> used. How do you compose such stuff on a standard US-English |
23 |
> keyboard and system? |
24 |
> |
25 |
> I'll do what I have to do, but only when I'm convinced it's the best |
26 |
> alternative. |
27 |
|
28 |
An easier alternative (IMO) would be to serve the pages with |
29 |
charset="UTF-8" and compose them as Unicode pages. This way you can |
30 |
just use the raw UTF-8 characters without having to look up entity |
31 |
names. Example: <http://remarqs.net/misc/pi.htm>. To do it that way, |
32 |
you'll need to edit the pages with a UTF-8-capable editor and save |
33 |
them with that encoding, natch. |
34 |
|
35 |
The issue you were up against, trying to use a font specification to |
36 |
change one character into another one, just won't work with html. When |
37 |
a browser encounters a character, it should first check to see if it |
38 |
can use the specified font(s) to render it; if it can't be rendered |
39 |
using the specified font(s), the browser must look for a font that |
40 |
contains a glyph for the character and use that font instead. |
41 |
|
42 |
In the case where the the specified font is Symbol, there's no glyph |
43 |
for a 'p' in that font, so the browser must use some other font |
44 |
whenever it encounters a 'p'. What it must *not* do is change a 'p' to |
45 |
a 'π'. |
46 |
|
47 |
IE has always been broken in this way, and the old Netscape browsers |
48 |
were as well. At the start of the Mozilla project, they did away with |
49 |
this bug. IIRC, before Firefox 3, there was a hackaround to |
50 |
re-introduce the bug by editing some Firefox files, but it's no |
51 |
longer possible to trick Firefox into substituting one character for |
52 |
another. (I don't believe WebKit/KHTML browsers or Opera can be tricked |
53 |
into it either, but I'm not sure.) |
54 |
|
55 |
-- |
56 |
»Q« |
57 |
Kleeneness is next to Gödelness. |