Re: Font changes

Hello all.

On the subject of remaining compatible with HTML, Patrick points out:

>(Robert) must know the difficulties that might ensue by choosing an, in
>principle, adequately general but badly fitting structural design.  I
>assume that he will watch for that.

Since Dave says more up to date information on HTML font machinery
will be forthcoming soon, I am inclined to put fonts on the back
burner until then.  This thread of discussion has already given me
enough to go on to put in some place holders in the code...

However, I do want to verify my assumptions one last time.  Patrick
writes:

>If this display list is to have single arguments for font changes then
>it is good that it will be only an internal data format.  Is it the
>case then that this is only entertained for deviations from the
>standard display modes corresponding to display, text, script and
>scriptscript in TeX-style composition?  So that, say, the display list
>will have to have font information like this for such things as are
>commonly put in bold (vectors, dyadics, some bundles, etc.),
>blackboard bold (reals, complexes, integers, etc), script (Schwartz
>distribution spaces, Hilbert space names, etc.), fraktur (Lie
>algebras, ideals, local rings, etc.)  or whatever?
>
>Furthermore is, in conventional output, the sizing of special characters
>considered as clearly varying along conventional lines, e.g. 10-7-5 point?

My assumptions are:

	1) the display list *will* have to contain font information 
	   for things like bold for vectors, fraktur for Lie algebras,
	   etc.  However, to whatever degree is feasible, this
           information will *not* have to be put in the HTML, and will
	   be added by the parser according to heuristic rules.  The 
	   author will, of course, be *able* to override the default
	   processing by adding HTML mark up.
	
	2) As far as the 10-7-5 sizing, and other conventions of
           typesetting, that is up to the renderer to get right, and
	   doesn't appear either in the HTML or the display list.

	3) The commonest use of the (mfont ) specification will be to 
	   deviate from the base font, but it can also be used to
           specify complete font changes, for instance, to fraktur.
	   However, since the display list is generated internally
	   by the parser, it is okay for such changes to be on the
	   verbose and technical side.

Robert