[Axiom-developer] Re: postscript files from Axiom graphics

[root]

Bill,

re: make doing more work than required.

yes, this needs to be cleaned up. in general the makefiles are
designed to do the absolute minimum amount of work.

the technique i use is the same one i use on noweb. make has the
property that if you have a stanza with no dependencies you can
use this as a "marker". so for noweb you'll see that i have a
stanza:


noweb:
        ....... 
        echo "mark that we built noweb" >noweb

which creates a file called noweb. as long as that file exists
this stanza will never be rerun. removing the noweb file indicates
that this stanza needs to be rerun.

this technique was used to keep interpsys from being
rebuilt when not required. interpsys loads a subset of the
algebra. i determined this subset by a frequency analysis of
the file loads. the top N files were chosen to be preloaded
into the image because they are used so often.

so there should be an interpsys marker file which guards against
unnecessary rebuilds. however the stanzas for files that get 
preloaded need to erase this marker file in order to force the
rebuild. this is work yet to be done.

re: Gfun.c.pamphlet

please diff -Naur Gfun.c.pamphlet and Gfun.c.pamphlet.new 
and send me the diff. 

re: color

there are a lot of options possible while creating graphs, none
of which i remember so we'll have to drag them out of the code
and document them both in the code.pamphlet files and in 
bookvol1.pamphlet (the tutorial volume in book--main--1).

as an aside, you'll note that the book is now in 3 volumes
 bookvol1 -- tutorial
 bookvol2 -- programming
 bookvol3 -- reference

re: batch generation

we did this when we made the axiom images in the book.
i believe some of the code lives in appendix F of the book.
however, it is possible we had X around.

the issue is two-fold. first, axiom needs to write the data
from the formulas. second, there is a program called viewAlone
which can display that data. Automating the first step should
be reasonably easy. I believe the functions to do it exist.

In order to display the data there is a program
[[src/graph/viewAlone]] which gets built by the make. If you look at
the file [[src/graph/fileformats.pamphlet]] you'll see that i've
documented the file formats used by viewAlone, at least in the 2D case.  
you can run this program on the file [[int/graph/parabola]]

i don't remember how the postscript generation is done but i know
we wrote a bunch of postscript macros. however the ps
generation occurs it is likely that it does not depend on X.
we may have to modify the code to give a "-batch" fastpath.

t