Re: ConTeXt + asymptote

[Sanjoy Mahajan <sanjoy@mrao.cam.ac.uk>]

From:    Hans Hagen <pragma@wxs.nl>
> [asymptote eps] looks parsable

Those lines are only a bit of the output.  It's preceded by horrible
pstricks and font code to parse (about 800 lines in that eps file).
I'm not sure why asymptote decided to use pstricks.  I think it's for
figure inclusion, but there have to be other, less hacky ways.

>> +  if (texengine == "tex")
>> +    abort("No babel package in plain TeX.");
>>    texpreamble("\usepackage["+s+"]{babel}");

> btw, it should be:
>  if texengine == latex then texpreamble ...
> else you need an abort for each macro package

Yeah, I'm not proud of it.  By that point in the patching, I mostly
despaired of finding all the hardcoded latex spots and just added the
quickest hack to get something to work ("When the going gets tough,
the tough lower their standards.")  Not that it worked in the end.

From:    Taco Hoekwater <taco@elvenkind.com>
> It would not be hard to add the cmyk color space, but transparancy
> can not easily be done in PostScript output (it would require
> overlap removal code, and that is not present at all).

That's true, and the pdfmark hack added to postscript doesn't support
sending transparency information to a distiller (e.g. ps2pdf or
Acrobat distiller).  At least, it doesn't from what I could tell by
looking in the PDFmark reference manual from Adobe:
<http://partners.adobe.com/public/developer/en/acrobat/sdk/pdf/pdf_creation_apis_and_specs/pdfmarkReference.pdf>

Maybe it's time to abandon postscript and generate pdf directly?
About the only reason I sometimes run (non-pdf)tex is because I was
too lazy to convert metapost figures to pdf and hadn't added the scons
or Makefile rules to convert them automatically.

> I've been browsing a bit and I now know it has more of a C++ style
> of doing things, but personally I don't see that as a big advantage.

I agree.  Although perhaps related to the new syntax or the new
implementation: it's useful to have many mathematical functions ready
to use for graphing -- the asymptote authors keep adding functions
from the GNU scientific library -- instead of having to divide or
multiply by 256 (the metapost log and exp functions) or worry about
overrunning the integer range (and use the sarith package in MP).

> But more importantly, most of my illustrations depend heavily on the
> equation solving capabilities of MetaPost. I got the impression
> Asymptote cannot do that (but I may have missed something).

Right, I like the implicit equation solving in MP and find it very
intuitive and matched to how I think about figures.  Asymptote has
only explicit equation solving.  Laurence said (on the metapost list)
that his 3D package for MP doesn't have implicit solving either -- I
guess because the 3D objects are not native to MP so they don't
benefit from the built-in equation solver.

Maybe making 3d points native to MP is part of an answer to:

   What is actually needed to make it easy to do 3D in MP?

But I'm far from an expert in 3D drawings or MP, and maybe one of the
MP experts has more useful thoughts.

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.