·······@yahoo.com (c hore) writes:
> What caused their demise? Why did Xerox not match the
> ascent of Symbolics and TI Explorer?
Mostly because they perpetuated INTERLISP-D on their lispms for some
time, well after everyone else had started to make the change to
Common Lisp. INTERLISP was not particularly compatible with CL,
unlike Zetalisp or MIT LispM Lisp. And it didn't have nearly as great
an effect on CL as ZL did, other than the idea of the all-Lisp development
environment (which INTERLISP users rejoiced in from quite early on).
Also because most of the Xerox LispMs had unutterably terrible
performance. I believe it was deep rather than shallow binding, with
no hardware hacks to accelerate the search. This made symbol
references very slow. And the version of INTERLISP running on the
D-machines was compatible with INTERLISP-10 from the PDP-10s, already
considered a dying environment, and not with INTERLISP-VAX (was that
the name?), which was gaining in popularity.
I could be wrong, however, not having ever used a D-machine.
ISTR having read that the Dorado's performance was actually quite a
lot better than the others, and that it would have made an excellent
LispM other than the fact that Xerox didn't make it available for sale
to the public. Steele and White's paper on the history of Lisp
describes the D-machines and their influence on the LispM scene.
A few other odd LispMs existed other than Xerox's, including one made
by a Japanese company (was it Fujitsu?). And there was the PERQ,
which ran Spice Lisp from CMU, the predecessor to CMU Common Lisp.
CMUCL still bears some evidence of PERQ ancestry in its code, if you
dig deep enough. The IBM RT cruft is more obvious in most places,
though.
The LispM marketed by Siemens was just a repackaged Xerox machine, the
Dandelion maybe?
'james
--
James A. Crippen <·····@unlambda.com> ,-./-. Anchorage, Alaska,
Lambda Unlimited: Recursion 'R' Us | |/ | USA, 61.20939N, -149.767W
Y = \f.(\x.f(xx)) (\x.f(xx)) | |\ | Earth, Sol System,
Y(F) = F(Y(F)) \_,-_/ Milky Way.
"James A. Crippen" wrote:
> ISTR having read that the Dorado's performance was actually quite a
> lot better than the others, and that it would have made an excellent
> LispM other than the fact that Xerox didn't make it available for sale
> to the public.
We had a Dorado at Schlumberger.
Paul
* James A Crippen wrote:
> Also because most of the Xerox LispMs had unutterably terrible
> performance. I believe it was deep rather than shallow binding, with
> no hardware hacks to accelerate the search. This made symbol
> references very slow.
They were pretty slow, but I suspect they were also massively cheaper
than a symbolics, since they sold the later boxes (1186) as office
workstations, running a different (not lisp-based) environment. The
hardware was a lot less industrial than Symbolics' (it would be hard
to be *more* industrial than Symbolics' HW, but it was less industrial
than Sun HW of the same era, say).
> And the version of INTERLISP running on the D-machines was
> compatible with INTERLISP-10 from the PDP-10s, already considered a
> dying environment, and not with INTERLISP-VAX (was that the name?),
> which was gaining in popularity.
I don't think that VAX InterLisp was popular (but I may be wrong...).
--tim