> the major groups i see (and some notes on each) are: > > execution > syntax > regexes I would like to rename this 'pattern matching' to accommodate pluggability even here, and to leave door open for fuzzy matching and other more esoteric matching schema. > data structures Even SVs should be pluggable enough so that I can have bigints, sparse strings (implemented as trees), complex numbers. True multidimensional arrays? Sparse arrays? Sorry, I guess I'm advocating my own ideas :-) > objects > extensions > > new xs and modules Native interfaces, in general. If there's an interface in the system, in C/C++(/Java?), I want to be able to use it. > byte code This includes 'packaging' (of scripts/modules). Time for a '.par'? > microperl > cleanup How about 'portability'? I think we are more portable than Java and I like to keep it that way. 64-bitness, byteorder, filenames, trying to generalize system APIs so that the C/UNIX libc is not necessarily the first interface ffered to newbies (raw APIs like signals should of course be available where available, I'm not suggesting hiding them away) Or is the subject too cross-subject to warrant a wg, instead have portability as a guiding principle of all the wgs? -- $jhi++; # http://www.iki.fi/jhi/ # There is this special biologist word we use for 'stable'. # It is 'dead'. -- Jack CohenThread Previous | Thread Next