This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: [gsl] gsl_matrix_size_t ?
- From: Brian Gough <bjg at network-theory dot co dot uk>
- To: James Bergstra <james dot bergstra at umontreal dot ca>
- Cc: gsl-discuss at sources dot redhat dot com
- Date: Mon, 11 Jul 2005 19:31:53 +0100
- Subject: Re: [gsl] gsl_matrix_size_t ?
- References: <8cb27cbf050630104165ed8c44@mail.gmail.com><17093.20601.669527.961525@network-theory.co.uk><20050708032134.GD3063@aphex.iro.umontreal.ca>
James Bergstra writes:
> Do you think it would be appropriate to add size_t to the list of types
> for which gsl_matrix and gsl_vector are defined?
>
> I find it convenient to use matrices of indices... and using any of
> uint, ulong causes frustration when switching between architectures
> because the native index type has a vary-ing size!
It is a good point, and a slight oversight in the current design that
it is not possible.
We probably want some type of general index-mapping object, which
could perhaps also handle sparse matrices. I will have to think about
it for a while.
> PS. Thank you for your link to MSL :) But, could you mention that it is
> also for signal processing and machine learning?
Thanks, I've added the extra information. The performance extensions
originally caught my eye... some nice ideas there.
--
Brian Gough
Network Theory Ltd,
Publishing Free Software Manuals --- http://www.network-theory.co.uk/