Branding guidelines: replace letter with unicode lookalike, screen readers and search engines?

Doing a website for a fairly major client.

Their brand guidelines specify that for certain trademarks, a unicode symbol
which looks like a stylized 'i' must be used instead of the letter 'i'
whenever it appears in the navigation, text, headers etc.

The website is content managed using a custom framework (super strict
semantic website builder basically, convention over configuration etc.) that
we've developed so from a technical standpoint making sure that this happens
is trivial.

However, I have major misgivings about doing this because although the words
are obvious when looking at them, it's going to be horrible for screen
readers, search engines etc. etc. which won't think that the character is
actually an 'i'

Clearly I can't use an image with alt text as a) I like my text to be text
thankyouvery much and b) it needs to occur anywhere that the trademarked
word appears, it would look inconsistent mixed in with real text, image
won't resize nicely with text-zoom etc. etc.

For the same reason I can't really use sifr or any similar things

Would I get away with doing a search/replacement via JavaScript? I'm pretty
sure that Google et al. will then just see the original unmodified text, but
how about screen readers? Gracefully degrades to just being the real word
without JS.

Could I also use JS to wrap them in acronym elements or something similar
for this reason? Misuse for sure, but trying to be pragmatic here.

If worse comes to worse I'll just tell them that I'm terribly sorry but
there are usibility issues around this. They are very keen on accessibility
so I can talk them around, but a win/win solution would be nice

On 23 July, 08:35, "Nik Coughlin" <nrkn.... (AT) gmail (DOT) com> wrote:

Their brand guidelines specify that for certain trademarks, a unicode symbol
which looks like a stylized 'i' must be used instead of the letter 'i'

Unicode doesn't have "symbols", it has characters (OK, codepoints to
be picky). Even Unicode alone isn't enough to specify the shape of the
glyph drawn to represent these.

So is this stylized "i" representably and communicable with just
Unicode such that I can see it correctly with the typeface "Arial
unicode MS" on my desktop, or would I _also_ need to be using the
client's own hand-drawn magic font with the special squiggles in it?
If it needs the font too, then I suspect you know the answer...

In article
<e85772c8-7f1e-417d-9043-faa7a3363da8 (AT) a26g2000yqn (DOT) googlegroups.com>,
Andy Dingley <dingbat (AT) codesmiths (DOT) com> wrote:

On 23 July, 08:35, "Nik Coughlin" <nrkn.... (AT) gmail (DOT) com> wrote:

Their brand guidelines specify that for certain trademarks, a unicode symbol
which looks like a stylized 'i' must be used instead of the letter 'i'

Unicode doesn't have "symbols", it has characters (OK, codepoints to
be picky). Even Unicode alone isn't enough to specify the shape of the
glyph drawn to represent these.

It wouldn't be one of unicode characters 236-239 (Arial) by any chance?
If it is, you could use the appropriate one of &iacute; &icirc; &igrave;
and &iuml;

Although I'm not sure what a text-to-speech reader would make of the
result e.g. &iacute;Phone ...

Nik Coughlin wrote:

Doing a website for a fairly major client.

Their brand guidelines specify that for certain trademarks, a unicode
symbol which looks like a stylized 'i' must be used instead of the
letter 'i' whenever it appears in the navigation, text, headers etc.

To expand on what Andy said, not only does Unicode not define the
appearance of any character, but there isn't any reason why a font
wouldn't exist in which the specific style of "i" that your client has
in mind would be the way "i" is rendered, with other characters having
similar style characteristics.

Nik Coughlin wrote:

Their brand guidelines specify that for certain trademarks, a unicode
symbol which looks like a stylized 'i' must be used instead of the
letter 'i' whenever it appears in the navigation, text, headers etc.

That's a rather misguided position, but exposing the error in its entirety
would require writing a book. Well, a chapter of a book. What? Oh, would you
believe a major part of chapter? Like "Characters and markup" in chapter 9
of my "Unicode Explained"...

The short story is that characters of that kind have very specific scope of
usage and using them outside that scope, though not illegal or sinful, is
just pointless and tends to cause much more problems than it might possibly
solve.

Specifically, which character would you use for a "stylicized" i? For
example, for an italicized i, you could conceivably, but misguidedly, use
MATHEMATICAL ITALIC SMALL I (U+1D456). However, as the name suggests, it's a
mathematical symbol, not any letter i in italics, and it was introduced into
Unicode for the sake of mathematical usage, in a context where italics carry
a change in meaning, not just stylistics. On the practical side, this
character is not supported in any mainstream font, just specialities like
Code2001 (a great font for its scope of use, but nothing that Joe Q. Public
can be expected to have in its computer) and DejaVu Serif (ditto). Thus, if
you try to use mathematical italic small i on a web page, the great majority
of visitors will see a box, a question mark, or something more enigmatic.

However, I have major misgivings about doing this because although
the words are obvious when looking at them, it's going to be horrible
for screen readers, search engines etc. etc. which won't think that
the character is actually an 'i'

That's really not the biggest problem, though a real drawback too.
Technically, MATHEMATICAL ITALIC SMALL I, for example, is compatibility
equivalent to normal "i", so any _advanced_ rendering system could be
expected to use "i" as a fallback - but anyone who thinks web browsers are
anywhere near to advanced character rendering tools should re-calibrate
their reality sensors.

Would I get away with doing a search/replacement via JavaScript?

No, because it cannot render a character if it does not exist in any of the
available fotns.

--
Yucca, http://www.cs.tut.fi/~jkorpela/