A recent clpmisc post (link @ GG) resulted in a largely obnoxious and long thread in which amongst other things people talked about the already oft discussed question re whether Perl should be standardized or not.
In a post (link @ GG) I just read today, Peter J. Holzer -in reply to a mention of which Larry Rosler's famous "ANSI Standard Perl?" interview- pretty much settles the down the question:
The argument about the "marketing value" of a standard has some merit. But a standard is neither necessary nor sufficient for success, even in corporate and governmental environments.
There is no Windows ISO standard, no Java ISO standard, and yet these are used very widely. OTOH there is an ISO (or at least ANSI) CPU standard (for a subset of the SPARC ABI, IIRC) which nobody cares about and I think the Pascal or BASIC ISO standards are similarily inconsequential.Successful standards (like C90 or POSIX) are developed after there have been multiple implementations and when the implementors actually want to converge on a single standard. In the case of an open source interpreted language like perl5 there is little value in a standard. There is only one implementation, a second implementation is unlikely to happen (porting is almost always easier than rewriting from scratch) if it happens compatibility is probably only a minor goal.
