Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!ogicse!littlei!omepd!omews35!colwell
From: colwell@omews35.intel.com (Robert Colwell)
Newsgroups: comp.arch
Subject: Re: Object oriented architectures
Message-ID: <6197@omepd.UUCP>
Date: 26 Nov 90 16:11:27 GMT
References: <2925@crdos1.crd.ge.COM> <3215@ns-mx.uiowa.edu>
Sender: news@omepd.UUCP
Reply-To: colwell@omews35.UUCP (Robert Colwell)
Organization: Intel Corp., Hillsboro, Oregon
Lines: 30

In article <3215@ns-mx.uiowa.edu> jones@pyrite.cs.uiowa.edu (Douglas W. Jones,201H MLH,3193350740,3193382879) writes:
>by davidsen@crdos1.crd.ge.COM (Wm E Davidsen Jr):
>> 
>> ... I believe that if you are looking at "production system which
>> broke new groupd," then you would have to include the Intel 432.
>
>The 432 only broke ground in the sense that it was a VLSI chipset.
>
>The first production object oriented architecture was, as far as I know,
>the Plessy System 250...

The second sentence excerpted above can be correct without validating the
first.  The 432 was not a VLSI adaptation of the Plessey 250, and I can't
recall any of the 432 papers ever even referring to it.  For that matter, I
doubt that Plessey would have wanted them to, given how the 432 turned out.

Nevertheless, the 432 deserves better than you're giving it.  Did the
Plessey machine treat the physical processors as abstract data types?  (I
hope the answer's no, or you'll screw up my rhetoric.  Don't have Levy's
book with me at the moment...)  The 432 did.  The 432 also had
self-dispatching processors, abstract dispatching port types, and other OS
hooks.  And Ada support for rendezvous.  One could argue that this
particular ground might have been better left unbroken, but I don't believe
you can make the case that they were following in somebody else's
footsteps.

Bob Colwell  mipon2!colwell@intel.com  503-696-4550
Intel Corp.  JF1-19
5200 NE Elam Young Parkway
Hillsboro, Oregon 97124