[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Can Java be one-upped?

   This message came out of the blue from a Unix/Linux guru - one
I respect very much - who went on to greener pastures a year ago.
Note the title and tone as well as the request for information.
It fits in with "transputer-class" hardware and the recent
popularity of transputers on Slashdot.
   Larry Dickson

>Date: Fri, 10 Sep 1999 10:11:02 -0700
>To: Lawrence Dickson <ldickson@xxxxxxxxxx>
>From: David Bullock <david.bullock@xxxxxxxxxxxxxxxx>
>Subject: Re: Can Java be one-upped?
>At 10:28 AM 9/10/1999 , you wrote:
>>At 10:11 AM 9/10/1999 , you wrote:
>>>How goes it?
>>>Still doing the Occam thing?
>>Yes, both indirectly (design of a NAS based on occam, therefore
>>using many small independent programs connected by pipes); and
>>directly (but this is in abeyance because I have no time).
>>    My opinions have not changed, in fact have been confirmed by
>>experience of this year. There will be no escape from the
>>software morass except by genuine decentralization (including
>>knowledge decentralization) and flat, i.e. occam, coding
>>techniques. I'd say we put in a month's worth of effort for
>>each week's worth of value because of butting our heads against
>>these problems.
>>    occam-com@xxxxxxxxx is still going strong on the academic
>>side, but the commercial world still seems committed to Moore's
>>Law based supercentralization. At some point it will collapse
>>of its own weight.
>>    Larry
>That's starting to be seen as the practical limits of Moore's laws are
>starting to appear on the horizon (at least for microprocessor
>miniturization).  I've always been a huge fan of asymmetric
>multiprocessing, which is one of the reasons I loved the disk arrays so
>much.  The computing world had it right with the old FEP/mainframe design,
>and even the old commodore 64's, Atari's and Amigas were VERY decentralized
>processing-wise.  They were much more efficient and featureful than what
>Intel is hell bent on doing (putting everything on the CPU).
>Are there any good quality compilers/interpreters for Occam for Linux?  Or
>a good website?  I'm kind of interested in Python, but the Occam principle
>is one that resonates well with me.
>David Bullock
>Director of Information Systems
>SilkRoad, Inc., 9707 Waples St., San Diego, CA 92121
>Phone: 619-457-6767 x1165
>FAX: 619-457-6751