[GAP Forum] "Cannot extend workspace" on alternative presentation for S_n

[Alexander Hulpke]

Dear Jerry Swan,
> 
> I'm interested in obtaining rewrite rules for the group of all pairwise transpositions of n elements, say 'Tn'. I know that in group-theoretic terms it contains a lot of redundancy - that's precisely why I want rewrite rules for it ;-)
> 
> Taking g := Tn( 65 ); works fine, but for n := 66, the code (attached) gives me 'cannot extend workspace' on two different machines (with maximal values for -o) running XP with 2GB and 4GB respectively. Since that this is so, might this perhaps be due to some internal limit rather than a memory isssue per se? I ask in order to avoid obtaining more memory.

The mailing list does not permit attachments. If you believe the attachment is of crucial importance please send it to me directly by private mail.

However I assume that you use a Knuth-Bendix on a presentation.
The implementation of the KB in GAP is rather basic and magnitudes beyond what KBMAG can do.
> Are there any suggestions for some alternative representation I might use (given that I'm interested in obtaining a rewriting system, it's not clear to me what my other representation options are, if any) . I'm familiar with KBMAG and MAF, but would prefer to be able to generate the group from within GAP if possible.
Your best bet is probably to load KBMAG as a GAP package (probably replacing the binaries with MAF) and then set
KB_REW:=KBMAG_REW;

This will have GAP use KBMAG in place of its own KB.

Best wishes,

   Alexander Hulpke
> 
>   Please do not use, copy or disclose the information contained in this message or in any attachment.
I have just done this by sending this mail.