🚀 go-pugleaf

>>>>> "Joe" == Joe Armstrong <joe@erix.ericsson.se> writes:

> Also note that the largest Erlang application now weighs in at
> 460,000 lines of code (which if our factor of 7 is correct for large
> programs represents 3.2 Million lines of C++)

> 	(is free ... details of how to get it deleted ...)

> But is this really unfortunate? - Universities etc. get Erlang free.
> But that's not much good if you want to make products. For product
> development "Erlang Systems" sells fully supported commercial
> versions of Erlang. The commercial version of Erlang is
> significantly faster than the academic version and full support is
> offered.

I've not really looked at the technical details of Erlang, but this
restriction on distribution really is a problem for this particular
application (GNU extension language).  FSF is based on the free
availability of source (that is, I must be able to get the source,
modify it, and give it to my friends, who must all have the same
rights).  That appears to rule out Erlang.

Scheme is much more attractive from that point of view: write a scheme
interpreter/compiler (which is probably easier than an Erlang system),
and then release it with a modified GPL, which would allow its
addition to commercial apps without "infecting" them.

Scheme is also closer to emacs lisp, which is an important
consideration, since presumably emacs will want to use this extension
language eventually, and it's important that existing packages can be
made to work easily.

>    Change of thread to Telescript ....

[deleted: it would be nice to see some details of Telescript]

> I *have* seen things which in Telescript which I don't fully
> understand.  To be secure Telescript needs some strong form of
> encription -- if I'm going to pay my bills electronically I don't
> want weak encrytion methods being used.

> Thus telescript must have some form of strong encryption - but (the
> sting in the tail) this will cause export problems from the US and
> make it very difficult for Telescript to become an international
> standard (which the General Magic people want).

This could be a problem.  Now, if only they released source: then we
can link it with our own RSA/DES codes!

>     Joe Armstrong

-- 
Bruce                   Institute of Advanced Scientific Computation
bruce@liverpool.ac.uk   University of Liverpool