Posted by Alan J. Flavell on 07/08/06 13:46

On Sat, 8 Jul 2006, Jukka K. Korpela wrote:
....
> In that sense, I don't think U+24C2
> CIRCLED LATIN CAPITAL LETTER M is the right answer.

I wasn't enthusiastic about it in practical terms, but I wouldn't rate
it as actually wrong. Still, it's a point of view, I won't argue
about it.

> On the practical side, if you use U+24C2 (no matter how you enter it
> in an HTML document), then it will probably be seen correctly only
> by users who have Microsoft Office installed so that they have Arial
> Unicode MS and by a small number of users who have installed
> Code2000 or some of the rare fonts that support this character,

MSIE's behaviour is more complex than that. I'm seeing the circle-M
in MSIE when Lucida Sans Unicode is selected. Even though SIL
Viewglyph says that the font does not contain the character. I
suspect (as on previous encounters with this kind of behaviour) that
IE is supplementing the font's repertoire by using glyphs from the
far-eastern (CJK) fonts which it installs when one enables support for
"Far Eastern" languages (in Win/XP), or, say, Japanese in Win/2K.

> see
> http://www.fileformat.info/info/unicode/char/24c2/fontsupport.htm

Interesting resource, thanks, although it does say:
"This only includes fonts installed on this server."

I'm still missing a utility which could tell me "which of my installed
fonts contain a glyph for the character U+xxxx" ? Doing the converse
is easy, there are numerous utilities which will list the contents of
a given font, but I haven't seen one which will search the fonts for a
requested character.

> Besides, people using IE would not see the character right even if
> they have Arial Unicode MS, unless the browser has been set to use
> that font by default (not a common or generally advisable choice by
> a user) or your document explicitly suggests that font.

MSIE continues to show the glyph to me, even if I choose Tahoma.
Again, I think it's extending the selected font by using a CJK font.

For users, I really can recommend enabling CJK support, even though I
can't read any of the CJK languages. For developers, obviously it
could be a bit of a problem, as it then suggests that characters are
available when, for the default installation settings, they would not
be.

> For completeness, I mention that an encircled "M" could also be
> written in Unicode as letter "M" followed by U+20DD COMBINING
> ENCLOSING CIRCLE, but this would be an even more problematic
> approach in practice.

Agreed, which was why I deliberately didn't mention it :-}

> > > It is _not_ the logo for the Paris Metro.
> >
> > Right. If it *was* specifically their logo, then you probably
> > wouldn't be allowed to use it without their permission.
>
> In that case, the symbol

Well, U+24c2 is a bona fide Unicode character, but it isn't the
Metro logo as such. I tried to make that distinction before. So it
hinges on what you mean by "symbol".

> most probably would not have been included into Unicode.

There's at least one precedent as a Private Use Area character.

/
# NOTE: The graphic image associated with the Apple logo character is
# not authorized for use without permission of Apple, and unauthorized
# use might constitute trademark infringement.
0xF8FF # Apple logo # Roman-0xF0, Symbol-0xF0, Croatian-0xD8
\

cheers

[Back to original message]