[Clipart] Metadata

[Jonadab the Unsightly One]

>>One thing that would add convenience would be if I could have
>>it parse straight from a scalar (taken from form input), rather
>>than from a file.  
>
> Easy enough to add; I'll put it on the todo list and try to get 
 > to it in the very near future.

Actually, it turns out that CGI::Lite writes file uploads into
a file anyway, so I don't need this after all.  Though it would
still probably be a useful addition to the API in general.

> Ahh, gotcha, yes that's the same approach I use:  Get something 
 > that does some piece of the job in some crude fashion asap,
 > then as time permits, refactor it into a more proper form.

I think Perl lends itself very well to this sort of development,
because of a couple of things about the language.  First, it's
a large language (read: provides lots of features), so if you
know the language pretty well it's fairly easy to get something
working quickly, especially since the CPAN does half your work
for you most of the time.  (Parrot is going to bring this latter
benefit to a lot of other languages, too, which is a Good Thing,
IMO.)  Third, Perl code tends to be fairly concise, and so it's
easy to find things in it and therefore (relatively) easy to
maintain, IMO.  Also, the short change/compile/run cycle makes
debugging consume less time, so refactoring (which tends to
introduce a lot of easy-to-fix bugs) is less painful.

But I don't want to turn this into a language advocacy thread.
I was only observing that the start-rapidly-and-refine model of
development is a good fit for the Perl language; whereas, some
languages lend themselves better toward a more thoroughgoing
approach.