From: Hans Aberg (haberg@math.su.se)
Date: Sat Mar 11 2006 - 01:54:41 CST
On 10 Mar 2006, at 21:27, Keutgen, Walter wrote:
>
> Philippe is right. Because of the very formal mathematical
> language, it is just dangerous to translate 'measured by' by the
> common Hungarian verb for measuring.
I leave the Hungarian translation issue to others.
> I have searched the web. The only I have found is that in some
> 'programming language' U+225e is represented by '\measeq'.
Which programming language?
> In some texts the character is called – from the typesetter point
> of view of course – 'm over equal'. In order to keep the stress on
> the =, one could opt for 'm equals'. If in other languages one
> would need another letter, the symbol needs to be added to Unicode,
> otherwise one ends up with different fonts per language. Or should
> one then use a glyph variant selector?
All I have found, is that it is added by the STIX committee, which
somehow added because it appeared in some AMS/TeX file, it seems. I
have found no mathematical contexe where it is used. And with no such
math context, it is not possible to move further.
> If Gusztáv has time, he could contact the mathematics societies of
> Hungary and perhaps U.S.A.
I doubt it will help. One needs math contexts of is usage.
On 11 Mar 2006, at 05:15, Philippe Verdy wrote:
> So it seems to me that U+225E is just a particular precomposed
> symbol, chosen quite arbitrarily.
Possibly.
> It is not enough to represent all actual uses, and it may even be
> very english-centric for a very particular application or some
> documents produced in some US universities, or discovered in a book
> or article from a single author (or small group)...
The composing of symbols and symbol components into what in effect is
a new symbol is very common in pure math, and one cannot add them
all. Perhaps some such less common combination just slipped through.
> I think that mathematicians and physicians need a more complete set
> of notations, not reduced to the few ones that are encoded now as
> precomposed characters. For such notation, Unicode will not help,
> and it will be easier to use MathML or TeX, that allow more control
> on the final layout, and better respect of the semantics (lost by
> the existing precomposed characters as they are not decomposable)
This is another issue. I think one can add symbol components to a
universal character set, so that these things become expressible in a
text file, but the problem is that one needs to create a good theory
for doing it, and that there are more than one way to do it. For
example, one will have to decide between a rendering oriented
approach, as in TeX and MathML, or a more semantic, parsing oriented
approach, as that of compilers and computer language construction.
The latter approach would be to preferred, but requires a great deal
of more analysis of mathematical usage. I have a good idea of what
components might be involved in this latter approach, as I write on a
theorem prover, where those things appear naturally. But not even a
good idea will be enough for this approach to be carried out fully.
Hans Aberg
This archive was generated by hypermail 2.1.5 : Sat Mar 11 2006 - 01:56:57 CST