axiom-developer
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Axiom-developer] Re: silver edition plot failure


From: M. Edward (Ed) Borasky
Subject: [Axiom-developer] Re: silver edition plot failure
Date: Sun, 12 Aug 2007 14:58:14 -0700
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8.1.6) Gecko/20070807 SeaMonkey/1.1.4

address@hidden wrote:
> Ed,
> 
>> (Silver Edition, downloaded with "git" last night, successfully 
>> compiled, but crashed trying to do a plot):
> 
> I'm concerned about this crash. Can you send me the exact line you
> typed that caused the error?
> 
> I run a test suite on hyperdoc and graphics by hand to ensure that
> they both work before I do a commit. I have the last 5 branch builds
> available here and all the tests I run work. Clearly I'm missing
> something. Sorry about that.
> 
> Tim
> 
> 
> 

What I did was:

1. Download and build Axiom.
2. Start Axiom.
3. In Hyperdoc, go to the Draw menu, pick "A Parametrically Defined
Tube", "Continue", "Do It".

It crashed on some kind of permission problem -- it wasn't able to write
"gazonk0" dot something in /tmp. I just tried to reproduce it, and it
looks like it's working now. So it was probably some kind of environment
problem (Gentoo Linux, testing level, amd64 dual core with 4 GB.)

I've been having a number of other problems on this system recently -- I
got ambitious and tried to bring it up to GCC 4.2 and enough stuff broke
that I had to spend two days rebuilding everything back to GCC 4.1.2.

So I'm on the air now. I used to have a few things in Hyperdoc that
broke with permissions issues, but I haven't had a chance to do much
with Axiom until this weekend, and I assume they've been fixed. I pretty
much only use Hyperdoc at the moment.

I bought the paper book and have worked through some simple cases but
haven't had a chance to really use Axiom for real work, which in my case
is continuous and discrete time Markov, semi-Markov and generalized
stochastic Markov processes associated with computer and communications
systems modeling. Once I get my other big project going -- profiling the
Ruby interpreter -- I'll be getting back to the Markov stuff.

Most people are doing these things numerically, and almost all of the
usable codes are "academic licensed" -- freely interchangeable amongst
universities but expensive as all get-out in a commercial environment.
So I can't, as I noted on (I think) the Sage list, get something working
at home and take it into my workplace.

But this is really a mixed symbolic and numerical application, involving
matrices (sparse and huge at times), Laplace transforms, Kronecker
products, data structures, partial differential and integral equations,
and a few other things. To do it right, you need a very-high-level
mathematical language like Axiom, and you need "literate programming" or
"reproducible research" for the documentation.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]