Help us defend against VMS!

[Barry Shein]

Doug Gwyn responding to Rahul Desi
>>But BLISS is a machine-dependent language.
>
>I agree with most of your article, but BLISS isn't particularly
>machine-dependent.  It is comparable to C in many ways.  I have
>seen implementations for three substantially different architectures,
>and there may be more.

I think the confusion here is that although Bliss is definitely a
potentially machine independant language I doubt much VMS useage of
Bliss is portable. It's much easier to work in all sorts of machine
dependancies as you have to provide all sorts of user definitions for
fundamental operations for each program you write, there are Bliss
libraries but they tend to be weak on the portability issue. Also,
there was never the tradition of portability in the Bliss community,
programmers in practice seem to rarely consider it as a goal.

I've programmed a fair amount in Bliss on a TOPS-20 system and was
even thinking of portability but it was very hard to maintain (things
like direct linkage to assembler routines to do system calls
necessarily find their way into your code, almost unavoidably.) I
would say the rewrite would approach a 50% re-effort even with the
best of intentions, mostly because the non-portable pieces will
also tend to be the most technically dense.

So it remains more of a theory than practice, and I don't think Bliss
will become the wave of the future (or was ever even the wave of the
past, mostly due to DEC charging a huge amount of money for a Bliss
compiler and the community that would use it not usually being in
a position to buy such products.) It's too bad, it's an interesting
language, albeit a bit dated.

	-Barry Shein, Boston University