Re: tx fonts [Re: [Texmacs-dev] txexa font]

[Joris van der Hoeven]

On 13 May 2003, Nix N. Nix wrote:
> > To get the proper rendering you *really* have to create bitmaps for
> > the composite characters. Of course, that is not going to work well if
> > we want a continous range of braces (not a discrete one), but anyway I
> > do not think the font model would support this anyway.
>
> I wonder though:  We could treat the \bracext character as a special
> case, because it's so simple.  We could stretch its bitmap in only the x
> direction using a simple blt stretch, because however the antialiasing
> algorithm decides to vertically render a single column of pixels must be
> the same for all columns of pixels.  So, we can stretch the bitmap ad
> infinitum and continue to claim that it is correctly rendered.  It's
> like those Web site backgrounds that consist of a long, pixel-thin,
> vertical gradient which, when tiled, appears as a vertical gradient
> across the entire page.

I also thought about this. In fact, you do not even need the txexa
character to do that: the brace extremities suffice. In fact,
such a mechanism might also be useful for creating more general
extensible characters, like arrows (needed for MathML).
Notice that I still recommend to let the length be an integer
multiple of a small fixed physical length like 0.1fn.