Hi John

The title of the Fox book sounds "hauntingly familiar", so that could be it. 
You 
definitely have the title of the Iliffe book. Iliffe and Barton were friends. 
When we started to think Moore's law might actually hold, one of the notions 
was 
that you could have a lot of base-bounds registers that would confine executing 
objects absolutely. This would allow objects to be both fine-grained and small, 
and to be safe.

The base-bounds idea later got convolved with MMUs for page swapping, but they 
are really different ideas. 


Cheers,

Alan




________________________________
From: John Zabroski <[email protected]>
To: Fundamentals of New Computing <[email protected]>
Sent: Tue, July 26, 2011 7:48:59 AM
Subject: Re: [fonc] Re: [squeak-dev] [ANN] Alan Kay to talk about "Next steps 
for qualitatively improving programming" at HPI in Potsdam




On Sun, Jul 24, 2011 at 5:37 AM, Alan Kay <[email protected]> wrote:

One of the things on his list was to read a book of papers that had been 
gathered by (Stanley?) Fox from the UK. I recall that some of the papers in 
this 
book were classics by Strachey and Landin on foundations of programming 
languages, lambda calculus equivalents of Algol, etc. I recall that most of the 
papers in this slim book were very good.
>


"Advances in programming and non-numeric computation" by Leslie Fox?
 


Another book on his list was by Iliffe (who among other things was the designer 
of the Rice University  computer, an early one using "keywords" which were 
called "Descriptors" on the -- even earlier -- B5000). This book was about a 
new 
machine design of his.
>


I have mentioned this book before on this list [1].  It's called Iliffe's 
"Basic 
Machine Principles", and his hardware design was so-called the Basic Machine.  
There are two versions of the book.  Incidentally, for an interesting 
historical 
perspective on the Basic Machine and why it failed, see [2] for the memories of 
one of Iliffe's colleague.

Basic Machine Principles is basically an attack on the von Neumann Architecture 
and the use of base and limit registers to govern address space segmentation.  
Iliffe thought that programs themselves should know how to access their 
resources, and that they should only be allowed to access other programs 
resources through explicit communication, and that such communication the 
"obvious" solution to the von Neumann bottleneck.


[1] http://www.mail-archive.com/[email protected]/msg01329.html
[2] Scarrott, Gordon G. "From Torsional Mode Delay Lines to DAP".  
http://www.cs.man.ac.uk/CCS/res/res12.htm
_______________________________________________
fonc mailing list
[email protected]
http://vpri.org/mailman/listinfo/fonc

Reply via email to