mailing-list for TeXmacs Users

[Joris van der Hoeven <address@hidden>]

On Fri, 23 May 2003, David Allouche wrote:
> On Thu, May 22, 2003 at 11:33:56PM +0200, Norbert Nemec wrote:
> > On Donnerstag, 22. Mai 2003 20:24, Joris van der Hoeven wrote:
> > > > Maybe I am wrong, that is OK with me, I am just wondering 'can it be
> > > > done?'  Can I have two letters in a row in math mode where both 
> > > > letters
> > > > are italics and there is no extra space between them?
> > >
> > > Yes; it suffices to insert a dummy expression like an empty group.
> > > But I am afraid that this is not what you should do here.
> >
> > Similar situation I had problems with: multiple indices like f^{a b c}
> > (stucture constants) The spaces just look ugly within a formula. What I am
> > doing at the moment is f^{a^{}b^{}c} - looks fine, but internally, it is, 
> > of
> > course completely ugly. Any ideas for better solutions?
>
> Probably texmacs needs some conterpart to MathML ⁣.

Yes, that is the solution I wanted to propose too.

> But this contradicts a design goal of a WYSIWYG math editor which is:
> avoid multiple invisible symbols with different semantics.

There may be additional tricks in order to make such distinctions
apparent. One might for instance render the characters slightly
differently on the screen (e.g. using pastel colors), but the same
on paper. One also does notice the difference when placing
the cursor right to the symbol.

> This issue is probably related to the larger (mathml related) issue of
> math grammar. For example, in your document it would be useful to
> identify logically those "structure constants" and have the typesetter
> automatically render commas in superscripts as InvisbleComma.

That is the other way, but this is really a matter of
mathematical and physical macros, not of grammar.
We already do have "binop"; likewise, one might have
"structure-constant", "tensor", "Lie-bracket",
"ket", "hypergeometric-function", etc.
This is indeed a direction which I recommend,
because it adds structure to documents.

> Also it would be nice to be able to customize the math fonts on a
> per-document basis. For example, some typesetter customization should
> allow me to say: "use this glyph for this symbol". That would make
> happy those who insist on italics derivatives.

I rather intended to regard that as a font property, but you are right.

> That italics derivatives are right or wrong is of no concern here.
> TeXmacs provides sensible defaults, but some people want it otherwise
> (maybe someone with more hierarchic authority is insisting on it). So
> it should be possible to do it otherwise.

Of course, but a first thing which had to be done was to introduce
special symbols for the derivative and remarkable constants.

> This again is also related to the still larger issue of allowing
> customization of syntax. For example, different people may use
> different spacing conventions with punctuations. Currently one can get
> the desired layout with customized keybinding which insert rigid and
> elastic spaces before punctuation. But this is error prone, does
> not rely and logical structure and not customizable by style.

This has also been foreseen (texmath.syx),
but further customizability has still to be implemented.

> Any thoughts on this issue?

Yes, but progress on these issues really depend on *your* progress on
the MathML conversion stuff. When you will get to content markup,
then these issues will progressively be dealt with.