I've been trying to get solar physicists to use PDL for years now,
including handing out hundreds of "complete source installation" CDs
with all of the ancillary libraries on them. Only now are people
starting to ask about how to use PDL. In fact today someone asked me
about PDL after having tried (and rejected) *Java* of all things...
Installation has been an issue on older or weird systems, but almost
all the interest I've had comes from people running various up-to-
date flavors of Linux, which are straightforward to install. It is a
problem that CPAN doesn't install properly (the 2.4.2 CPAN module
asserts that it failed even though it often succeeds).
Several folks have recently jumped on the GDL bandwagon, though why
one would want a bug-for-bug compatible IDL replacement I don't
know. NumPy is slightly more popular, which is too bad as it is
beginner-friendly and expert-hostile (IMHO). Folks who have adopted
other solutions and have talked to me tend to complain about the
perceived formidability of Perl -- until you're used to seeing '$'
and '@' everywhere, it's a bit overwhelming...
A tutorial would be just the right thing. Several have been started,
but as far as I know none have been finished. I'd be happy to co-
author one if anyone feels like cobbling together a book -- this
probably requires getting 2-3 people to hammer out a structure and
then divide up responsibility for individual sections. The
PDL_BEGGINING_BOOK CVS tree is an interesting start but now
unfortunately somewhat out of date...
My guess is that we can't get a book written with the same
freewheeling contribution style that works for the software in
general -- the reason is that it takes more effort and discipline to
write English than to write small modules of code.
Cheers,
Craig
On Jun 26, 2006, at 1:58 PM, David Whysong wrote:
On 6/26/06, Karl Glazebrook <[EMAIL PROTECTED]> wrote:
I often look amazed at how inefficiently some people try to do things
in packages like IRAF (which only those in astro will understand the
reference).
IRAF has moved to Python (pyraf), as has AIPS++ (casapy). In a few
years, python will probably be the standard interface language for
astronomical data reduction... at least, for those people who don't
use IDL.
As much as I like PDL, it's capabilities and completeness are lacking
in some areas compared to Python - notably, the graphical plotting
library used in casapy, which is very impressive.
That said I WOULD like to see wider adoption rather than feel like
some archmage. A linux binary distro would help a lot.
I'm not convinced that a special-purpose linux distro would be useful
for most potential users. It's better to have a good binary .rpm or
similar package. A lot of people probably don't want to build from
source, and a custom distro is only useful to people who want a
computer dedicated to running PDL.
--
David Whysong
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
