On 11/18/08 12:56 PM, "Jeff Savit" <[EMAIL PROTECTED]> wrote:

>> Very heavy, as well. Takes two *strong* people to move one. Takes a serious
>> cart to move one very far.
> So, be certain of its placement when you put it in the attractive,
> modern livingroom of the future!

Right next to the plastic bowl chairs (thank you, Saarinen!)

>>>> Anyhow, language based processors seem to be a seductive design pattern
>>>> that hasn't really gained momentum for general purpose computing.
>> Don't forget the Xerox D-machines and the Symbolics, and the LMI machines as
>> well. Symbolics still has a lot of applications in DOD space, and there are
>> still hundreds of D-machines and D-machine descendants in the space program.
> I didn't know  the Dorado and Dolphin had p-code like instruction sets;
> I thought they were conventional instruction set architectures along the
> lines of a Nova..  I would be amazed if D-* were still in use after 30+
> years. The real innovation there was in the WYSIWYG graphical view it
> provided - a breathtaking leap forward.

No, the D-machines had several sets of loadable language-specific
pseudoinstruction sets (Lisp, Prolog were commercially available, and there
were a few special purpose language sets only used inside Xerox and Stanford
AI). The core processor had a very p-System like virtual microcode gadget
that you could do neat things with. I have a fairly complete set of native
Lisp microcode worlds for my Dandetiger, but I don't think my copy of the
native Prolog world is still usable. 8-(

(The CHAOSnet implementation in Prolog was a genuinely inspiring piece of
work -- a complete networking implementation written only in logical
assertions. Not for the ordinary mortal to attempt, that's for certain. )

Re: D-machines still in use -- Goddard SFC still has about 20 that I know
of. JSC has a bunch more, plus some of the 608x systems that are Son-of-D.
Most of the shuttle mission checklist documents were developed back in the
late 80s/early 90s, and if they change platforms, they'd have to recertify
all the docs again, so it's cheaper to keep the D-machines running (over
real original Ethernet coax with D connectors and real DEC orange-stripe
DELNIs) than recertify. I also know of at least two Symbolics 36x0s still in
use in various bits of the DOD, but that's about all I can say about them.

>> Many, many hours with both. And p-System FORTRAN, and the p-Writing Tools
>> package developed at UCSD, which was a brilliant set of writing assistant
>> gadgets for teaching technical writing, and an ancestor of Google Apps by
>> marriage.  Modula, less time, but contributed a fair amount of doc kvetching
>> to the CMS Modula port.
>>
> You must mean the _other_ Modula-2 port to CMS, not mine. I had Dick
> Karpinski to harass me :-) after I presented it at MODUS and USUS. Ah,
> memories! If anyone wants my compiler and runtime system I'm sure I can
> dig it up. I added libraries to access CP and CMS functions for systems
> programming; I even did a VMCF interface. But REXX made that use-case
> superfluous.

Gee, I didn't know you did one. I was working with one that was done in the
Netherlands around 1989 -- I seem to remember Andy Tannenbaum having
something to do with it. Probably one of his students or post-docs. If you
want, I can put it up for FTP or other downloads -- I'd like to have a look.

-- db

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to