On Friday 16 April 2004 11:19, Valery Pipin wrote:
just in addition to my prevous message
| On Friday 16 April 2004 10:06, James Amundson wrote:
| | I am forwarding this on Viktor Toth's behalf. Is anyone *currently*
| | working on the tensor package?
|
| This is a good news!
| I'm working with it in everyday work. I consider the cvs version as
| workable (in most cases). I also wrote some extension to manage the
| Levi-Chewitta symbols and these packages were posted to the list an year
| ago. I belive that cvs version gives correct results on examples given in
| the manual as well. These exaples are reproduced in itensor.dem. I guess
| Viktor should check the cvs firstly.
|
| best rgds,
| Valery
|
|snipped |
| | One issue that affects both of these packages is that prompting (e.g., in
| | ENTERMATRIX) doesn't work correctly: prompts are not displayed presumably
| | because output buffers are not flushed. Is this something you could
| | address in an upcoming release?
Do you mean entertensor()?
This works for me on the maxima cvs in (gcl,cmucl,sbcl) implementation
| |
| | And... unrelated but whereas in the past, I always compiled MAXIMA with
| | GCL, it appears that recent versions do not work correctly if compiled
| | with it. CMUCL works fine. I presume this problem is known.
It seems to me the cvs version works equally well (on my programs) with all of
above lisps.
As to me the most important issues are:
1)components(..), remcon() are'nt working
2)loading ctensr after itensr fails
3)the better integration between itensr and ctensr is needed
In ctensr the metric variables are in lower case while itensor function
generate() produces them are in upper case
4) simplification could be better
5) there is no tool to manage indicially the Levy-Chevitta symbols
Finally,
before uploading changes to itensor/ctensr
Please be sure that the following demos runs ok
"itensor.dem", "ademo.dem", "shwartz.dem"
best rgds,
Valery