Re: [Axiom-developer] Re: Axiom next release

[root]

Can't imagine how Latex was involved in that error.
Usually that error is because the file is missing.
INTERP.EXPOSED is a list of "exposed constructors".
The interpreter does not see every possible constructor
because it does not need to. INTERP.EXPOSED is one of the
databases I build. You'll notice several copies of it as
it is in the process of moving directories and I haven't
had the time to kill all possible old references. NAG built
a "share" directory and did not follow the rule that all
shipped files live under "mnt". I'm reworking it so that
will again be true.

I'm glad to see that you can do arithmetic. More is coming
but I'm blocked by a bug which I have not yet found. Yet 2+2
exercises a fair amount of the machinery and is an encouraging
sign.

For documentation purposes I'm using a program called "dia" to
build a graph of the lattice. It has helped me find a bug already.
I'm snowed in at home today so progress will be made :-)

The most useful thing we need is to try to port the system to
another platform. It is a huge job in general as some of the
later parts are in C. This part should work provided you can
get latex, noweb, and gcl to compile and have various tools
like make, patch, ar, cp, tar, etc available. The MKS toolkit
is ideal for windows but rumor has it that various tools have
been successfully ported. Please take a few notes about porting
as, hopefully, you won't be the only one to try to move the
system around. If you find that we need various packages from
the web let me know.

On the porting front HP has a deal where you can sign up to use
their computing farm for porting. I wrote about it in another
email in this group. I'm signed up and itching to play with the
64-bit version but I fear that isn't going to happen until my
mother has one in her toaster. You might want to check into it
if you want a real challenge.

If you want a simple place to start you can look at the issue of
documenting the algebra. It needs work at two levels. First, if
you look at the .spad.pamphlet files you'll see that they are TeX
documents. Each domain (classes with representation structures),
package (classes without representation) and category requires
a long explanation of its mathematics or computer science (for
data structures), its details (such as cross references to other
domains with \cite, specifics of the algorithm, references to
the literature, etc). Look at dhmatrix.spad.pamphlet for an example.

Second, if you look at the algebra compiles you'll see a raft-load of
messages about functions that are not documented or have improper
words in their documentation. There is an automated build of online
documentation that occurs from the so-called "plus-plus" (lines
beginning with ++) documentation (lines beginning with -- are
ignored). The ++ comments require a certain style using special
keywords which, of course, themselves are not documented. It appears
that these were never maintained once the system left my hands.

Or choose what interests you to work on. There is work needed
everywhere.

Tim