On Thu, 14 Dec 2000 11:45:44 +0900, Ben Tilly <ben_tilly / hotmail.com> wrote: >By contrast Guido has tried to maintain more control over >Python's licensing. You're incorrect, and I'd like to stamp out this FUD. Before version 1.6, the Python license was the same as the MIT X11 license. With version 1.6, CNRI, GvR's former employer and my current one, demanded a licensing change, very much against the will of GvR and the Python community. The 1.6/2.0 license, visible at http://hdl.handle.net/1895.22/1012, is much longer, but most of the terms have little bearing on what you can do with the software; only terms 2 and 3 matter, and according to them you're still perfectly free to use and distribute modified versions of the software. Commercial products can embed Python and don't have to ship any source code, Infoseek and CADscript being two examples of projects that do that. Nor do you have to hew to any language definition; the Alice project made serious language-level modifications (case-insensitive identifiers, and changing the definition of the / operator) and redistributed the resulting binaries. You're simply incorrect in this statement. >If this restriction appears likely to cause problems, then >it may be worthwhile for someone to take some time out to >come up with less restrictive replacements for the key >sections. PCRE might be worth contemplating as a BSD-licensed regular expression library, since it provides Perl-compatibility and is fairly readable. The pcre_compile() function is pretty hairy, though, but I'm sure PCRE's author would accept improvements; he was quite friendly when I was adapting PCRE to Python. --amk