ALT image for text?

[phil_gg04@treefic.com]

Dear All,

I've recently been looking at some of the symbols defined by Unicode.
Quite a lot of geometric shapes, arrows and so on are defined and they
make a good alternative to small .gif/.png images e.g. for buttons.
(Have a look at http://www.unicode.org/charts/).

But they will only display correctly if the user has a font that
includes them. And as far as I can tell there is no way that a page
can tell if a symbols has been displayed as intended or not.

This is actually the opposite of the ALT attribute for images: that
defines text to use if an image can't be dispayed, but I'd like to
display an image if a character can't be displayed. Can anyone think
of a way to make this work?

Failing that, does anyone have any experience to suggest a pragmatic
subset of Unicode that can reasonably be expected to be supported by
most browsers?

Regards

--Phil.

[C A Upsdell]

Re: ALT image for text?

phil_gg04@treef ic.com wrote:[color=blue]
> Dear All,
>
> I've recently been looking at some of the symbols defined by Unicode.
> Quite a lot of geometric shapes, arrows and so on are defined and they
> make a good alternative to small .gif/.png images e.g. for buttons.
> (Have a look at http://www.unicode.org/charts/).
>
> But they will only display correctly if the user has a font that
> includes them. And as far as I can tell there is no way that a page
> can tell if a symbols has been displayed as intended or not.
>
> This is actually the opposite of the ALT attribute for images: that
> defines text to use if an image can't be dispayed, but I'd like to
> display an image if a character can't be displayed. Can anyone think
> of a way to make this work?
>
> Failing that, does anyone have any experience to suggest a pragmatic
> subset of Unicode that can reasonably be expected to be supported by
> most browsers?[/color]

No more than the standard character entities ... and even then some
characters will not appear. See
http://www.upsdell.com/BrowserNews/res_entities .

[Jukka K. Korpela]

Re: ALT image for text?

phil_gg04@treef ic.com wrote:
[color=blue]
> I've recently been looking at some of the symbols defined by Unicode.
> Quite a lot of geometric shapes, arrows and so on are defined and they
> make a good alternative to small .gif/.png images e.g. for buttons.[/color]

In principle yes, but in practice images work better for buttons, if you
are not satisfied with buttons containing just normal text. Buttons aren't
"real text". In text (paragraphs) where symbols should adapt to the font
size, you could use Unicode characters even though this implies some risks.
The simplest arrows, for example, are relatively safe.
[color=blue]
> But they will only display correctly if the user has a font that
> includes them.[/color]

And there's more - IE (generally) won't show the character unless the
_current_ font includes it.
[color=blue]
> And as far as I can tell there is no way that a page
> can tell if a symbols has been displayed as intended or not.[/color]

That's sad indeed. It would be nice if we could specify alternatives to
special characters; but we can't.
[color=blue]
> This is actually the opposite of the ALT attribute for images: that
> defines text to use if an image can't be dispayed, but I'd like to
> display an image if a character can't be displayed. Can anyone think
> of a way to make this work?[/color]

I don't think so. But it's not really the opposite of ALT; rather, the same
general idea: specify something as content but also specify a "fallback"
for it, in cases the primary content cannot be presented. (Quite often, the
ultimate fallback would consist of ASCII characters, e.g. "-->" for an
arrow symbol, or maybe - especially thinking about how things work in aural
browsing - "approaches " or "affects" or "changes to" or whatever the arrow
_means_ in a particular context.)
[color=blue]
> Failing that, does anyone have any experience to suggest a pragmatic
> subset of Unicode that can reasonably be expected to be supported by
> most browsers?[/color]

In most browsing situations in Western countries, Windows Latin 1 works,
and all the rest is more or less unsafe. But it's _relatively_ safe to use
characters from the Windows Glyph List 4, see

http://www.eu.microsoft.com/typography/otspec/wgl4.htm

or (more conveniently)

http://www.alanwood.net/demos/wgl4.html

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[Alan J. Flavell]

Re: ALT image for text?

On Mon, 28 Feb 2005 phil_gg04@treef ic.com wrote:
[color=blue]
> (Have a look at http://www.unicode.org/charts/).
>
> But they will only display correctly if the user has a font that
> includes them.[/color]

As already mentioned: it's worse with IE. It doesn't go hunting
characters, it only looks for whether the font claims to cover a whole
group of characters.
[color=blue]
> And as far as I can tell there is no way that a page
> can tell if a symbols has been displayed as intended or not.[/color]

Quite right. The web doesn't have that level of feedback. IE should
have taken advice from Lynx, and provided analogous fallback if they
couldn't find a glyph.
[color=blue]
> This is actually the opposite of the ALT attribute for images: that
> defines text to use if an image can't be dispayed, but I'd like to
> display an image if a character can't be displayed. Can anyone think
> of a way to make this work?[/color]

Well, <span title="whotsit" >&#bignumber; </span> would stand a fair
chance of producing a popup that said "whotsit", *if* the reader had
the presence of mind to hover over the point in question. I'm not
saying it's ideal, but at least it's better than nothing.
[color=blue]
> Failing that, does anyone have any experience to suggest a pragmatic
> subset of Unicode that can reasonably be expected to be supported by
> most browsers?[/color]

There's no single answer to that, but you might find some relevant
discussion in

http://ppewww.ph.gla.ac.uk/~flavell/charset/browsers-fonts.html

You'd have to assess your potential audience, is what it comes down
to.

[Stan Brown]

Re: ALT image for text?

"C A Upsdell" wrote in comp.infosystem s.www.authoring.html:[color=blue]
>No more than the standard character entities ... and even then some
>characters will not appear. See
>http://www.upsdell.com/BrowserNews/res_entities .[/color]

HTTP/1.1 404 Object Not Found

--

Stan Brown, Oak Road Systems, Tompkins County, New York, USA

http://OakRoadSystems.com/

[Jukka K. Korpela]

Re: ALT image for text?

Stan Brown <the_stan_brown @fastmail.fm> wrote:
[color=blue][color=green]
>>http://www.upsdell.com/BrowserNews/res_entities .[/color]
>
> HTTP/1.1 404 Object Not Found[/color]

The correct URL seems to be

http://www.upsdell.com/BrowserNews/res_entities.htm

(I think the URL that was posted worked for me yesterday. Maybe something,
like automagic selection of suffix, was changed on the server.)

The information there looks correct (unlike in so many "references "),
but I don't see much added value in comparison with e.g. the
well-known tables at http://www.htmlhelp.com/

I find the final statement on the page somewhat amusing:
"This page was updated 8. tammikuuta 2005."
(A major task in localization is to understand what should not be localized
at all and what the locale should depend on.)

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[phil_gg04@treefic.com]

Re: ALT image for text?

I asked:[color=blue]
> [unicode symbols] will only display correctly if the user has a font[/color]
that[color=blue]
> includes them. And as far as I can tell there is no way that a page
> can tell if a symbol has been displayed as intended or not.[/color]

Thanks for the feedback everyone - even if there is no "magic solution"
that I had been hoping for.

If you're interested, have a look at for example

http://treefic.com/preview/treefic/royal92?a=tree_page&root_id=10273&zoom=2&ancg=2&descg=2

I'm using double-shafted arrows at the top and bottom of the diagram
and a triangle shape to form the "resize handle" for the dialogs.
Elsewhere I'm using the tick (aka "check mark") symbol &#x2713;. I'd be
interested to know if these work for you.

(I'm aware that there are other issues with these pages; they are only
a preview but constructive feedback would be appreciated.)

--Phil.

[C A Upsdell]

Re: ALT image for text?

Stan Brown wrote:[color=blue]
> "C A Upsdell" wrote in comp.infosystem s.www.authoring.html:
>[color=green]
>>No more than the standard character entities ... and even then some
>>characters will not appear. See
>>http://www.upsdell.com/BrowserNews/res_entities .[/color]
>
>
> HTTP/1.1 404 Object Not Found[/color]

Sorry about that. Dropped the .htm. Should be

http://www.upsdell.com/BrowserNews/res_entities.htm

[Stan Brown]

Re: ALT image for text?

"" wrote in comp.infosystem s.www.authoring.html:[color=blue]
>If you're interested, have a look at for example
>
>http://treefic.com/preview/treefic/r...ancg=2&descg=2
>
>I'm using double-shafted arrows at the top and bottom of the diagram
>and a triangle shape to form the "resize handle" for the dialogs.
>Elsewhere I'm using the tick (aka "check mark") symbol &#x2713;. I'd be
>interested to know if these work for you.[/color]

The double-shafted arrows look fine in Win XP Mozilla 1.7 -- no
surprise, as Mozilla is known for going out and finding Unicode
characters in any available font. I didn't find the check mark,
perhaps because I wasn't looking in the right place.

I did notice that the large fixed-position View and Find boxes
covered up part of the diagram.

--

Stan Brown, Oak Road Systems, Tompkins County, New York, USA

http://OakRoadSystems.com/

[phil_gg04@treefic.com]

Re: ALT image for text?

>>http://treefic.com/preview/treefic/r...ancg=2&descg=2[color=blue]
> I did notice that the large fixed-position View and Find boxes
> covered up part of the diagram.[/color]

Thanks for looking Stan. Interesting. Did you notice that these boxes
can be minimised to get them out of the way, and moved around? Maybe
you have javscript off; even so I think that it tries to have a long
enough vertical scrollbar that it should be possible to see all of the
diagram with scrolling. I'm curious because there is a difficult
balance here between having enough screen area to see a large diagram
and making features like Find sufficiently prominent.

--Phil.