Re: OT - Curl & patents

[address@hidden (Neelakantan Krishnaswami)]

Christopher Barber <cbarber@curl.com> wrote:
> 
> Yes, these are definitely perceptions that we have to overcome, but
> we feel that for a client-side language to succeed, it's integrity
> must be sustained.  [...] A strictly enforced standard would be
> another approach to this problem, but I cannot think of any examples
> of this working very well in the domain of programming languages.

FYI, this has worked for Standard ML. There is a formal operational
semantics for the core language and module system, and a well-defined
standard library (the SML Basis) on top of that. 100 KLOC SML programs
can be compiled on multiple compiler implementations, with the porting
effort primarily restricted to feeding the different compilers in the
way they expect.

The key is to both define the language unambiguously, *and* to provide
a rich enough standard library that people can use it to write useful
programs. (Scheme hits the first target but fails on the second....)

Anyway, good luck with Curl. I have no opinion on whether you will be
a success or not; my last attempt to predict the success of a
technology was in 1993, when I looked at NCSA Mosaic and decided that
this World Wide Web thing would never amount to anything. :)


Neel