Re: Introducing the idea of a "ROMAN VARIANT SELECTOR" (was: Re: Proposing Fraktur)

From: Kenneth Whistler (kenw@sybase.com)
Date: Thu Jan 31 2002 - 14:22:47 EST


Stefan suggested:

> > Maybe something like a "ROMAN VARIANT SELECTOR" would be appropriate:
>
> In any case, it'd be better to have *two* selectors, one to turn on Fraktur,
> and a different one to turn it off. Otherwise, you'd have to put the variant
> selector after *every* letter you want to be in antiqua, which would require
> quite a lot of space.

That sounds like a perfect example of where use of markup is appropriate.

Using Karl Pentzlin's Duden example:

Das sinkende Schiff sandte SOS-Rufe. (The sinking ship emitted SOS calls.)
fff Ffffffff ffffff Ffffff aaaffffff

<fraktur>Das sinkende Schiff sandte</fraktur> SOS<fraktur>-Rufe.</fraktur>

or conversely, perhaps better:

Das sinkende Schiff sandte <antiqua>SOS</antiqua>-Rufe.

And your rendering process then has all the context to decide whether
to apply the Fraktur rendering rules or the Antiqua rendering rules
to particular segments.

What makes you think that having these "selectors" encoded as characters,
rather than as markup, will improve the situation? You would need
other markup, anyway, to make actual font choices, for example.

And in general, the piling on of more stateful formatting
controls is damaging to the standard and damaging to the clear
relationship between the plain text content conveyed in Unicode
and the markup conveyed in markup languages such as HTML and XML.
(That relationship is already complicated enough at the edges, for
things like bidi -- and adding more instances of stateful transitions
that the purveyors of markup language believe belong in their
standards rather than in plain text doesn't help anything.)

There is a reason why similar stateful selectors such as U+206E NATIONAL
DIGIT SHAPES and U+206F NOMINAL DIGIT SHAPES are formally deprecated
in the Unicode Standard.

--Ken



This archive was generated by hypermail 2.1.2 : Thu Jan 31 2002 - 13:56:53 EST