(I speak now as a user of Perl)
Perl is for making easy things easy and hard things possible. Perl
currently makes a lot of things easy:
- text manipulation
- filters
- reading and writing data from files
- making/deleting/reading directories and contents
- short programs to do minor tasks
I don't think we should sacrifice any of the current ease. That would
be a mistake. It'd alienate Perl's current userbase, e.g. me.
But there are some things that Perl doesn't make easy, but which it
should:
- binary data (pack and unpack are a sublanguage nastier than REs IMHO)
- fixed field structures (`use fields' kind of compile-time checking)
- compilation to portable bytecode for faster runtime and easier
distribution
- ability to serialize CVs
- managing events from different sources (Tk + network + serial)
- use less 'memory';
- pluggable data types ("this hash is a hash of integers, so give me
the speed and memory wins from knowing this")
I think it's perfectly possible to add these things to Perl, possibly
changing some of the existing language to better fit them in, without
losing the bits that make us use Perl now.
Nat
