Re: SWI-Prolog 7

On 4/12/2013, at 11:27 PM, Michael Richter wrote:
> I was just wondering if we'd be able to set aside the pedantic standards
> bickering in the list for a while?

I think Ulrich has done a good job of explain that it's not
"pedantic"; the hugely practical issue involved is compatibility.

There was a time when there was a huge range of "culturally
compatible" but not *practically* compatible Lisp systems;
a time I remember well.  It required elaborate dialect conversion
tools if you wanted to use someone else's code.  My personal
experience covers Rutgers/UCI Lisp, "Portable Standard Lisp",
Interlisp-D, a version of Interlisp ported to the B6700, another
and rather odder Lisp on the B6700 where tags were in the pointers,
not the targets, and a couple of other pre-Common Lisps I have
managed to forget.  Oh yeah, T!  How could I forget T?  They al
had CAR, CDR, CONS, and so on, but there wasn't even a portable
way to define a function!  

The Lisp community got together and developed Common Lisp, and
CLtL1 came out.  Then they got working to produce ANSI Common
Lisp, and found more quirks and gotchas than you could shake a
wand at in CLtL1.  And eventually, we got CLtL2 and ANSI Common
Lisp and the freely available Hyperspec, and people pulled their
implementations together.

Some of the older Lisp systems went under.  But today, I have
several Common Lisp implementations, and I am *astonished* if
code I have written in the light of the Hyperspec doesn't just
work without being touched in all of them.

THAT is as practical as it gets.  This is a *HUGE* change
from "Lisp is a family of culturally related languages" to
"Common Lisp is something you can rely on".

The development of ISO Prolog was a greatly prolonged process,
not least because a number of people all wanted to *change*
the language THEIR way, and F**K the existing users.  But the
standard did come out.

I hated it.  I poured enormous effort into it, and felt, and
still feel, abused.

BUT since the standard came out, work on fixing it and
extending it has been
 - steady
 - careful
 - open.

It's not the standard I wanted.
But it's the best standard we've got,
and it's getting better.

If you like the changes to SWI Prolog and the direction it's
going, fine.  But don't dismiss a serious concern for the
welfare of practical programmers as pedantic dick-waving.