Hi Alan, 

--- Alan Kay <Alan.Kay at squeakland.org> wrote:

> It's worthwhile contemplating that most bits in
> computers are dynamic 
> relations between processes, so "state" is really a
> kind of process 

I would say ... it's the contents of a stack frame. 
:)  ahh, of course!  That's true whether it be a
reference to an object or to an integer.
 
> .... Quite a few of the ideas in Special Relativity
> on "how one knows 
> about things going on" apply to these ideas.

Could you tell me a bit more about this, please?  It
sounds very interesting...

> 
> >   The power is not the stuff, it's what's
> >between the stuff.
> 
> The Japanese call this "ma".

I hope to learn more as my life continues.


> >I have often wondered what the world's simpliest
> >Smalltalk evaluator would look like, with no vm,
> just
> >an eval loop.
> 
> The first one I did for Smalltalk in Sept '72 is
> pretty small (less 
> than a page). There is a sketch of it in one of the
> Appendices of the 
> "Early History of Smalltalk" paper. It was
> influenced by the Lisp one 
> on the bottom of page 13 of the Lisp 1.5 manual (but
> I did ST-72 
> assuming a nonrecursive base to make it a little
> more as though it 
> were running on a conventional machine). Dan Ingalls
> did a brilliant 
> job of making it really real. This could be quite a
> bit smaller if 
> done a little differently.

That would be awesome to get those published to SM.  
The first smalltalk machine, ever.  Thank you.  


> Think about what true and false are in Smalltalk.
> And what objects 
> are. You can easily synthesize arithmetic without
> needing any tables.

yep, I checked again.  I actually suggested table
lookup for addition of integers.  Let's see, for
squeak SmallIntegers of 31 bytes of representation,
that would be 4.6E18 slots, and at 4 bytes each, I
would need 1.8E13 MB of ram.   But then there are
those pesky multiplication tables...  

> 
> >  I would want to use it to help me see
> >all the way down to the bottom.
> 
> This is always a good idea. For Smalltalk-80 though,
> I'd recommend 
> the beautiful version (using byte codes) in the
> Appendix of Dan 
> Ingall's '78 POPL paper on Smalltalk-76. This is a
> pretty neat one 
> pager.

awesome, I found it here:
http://users.ipa.net/~dwighth/smalltalk/St76/Smalltalk76ProgrammingSystem.html
.  :)  It's dead simple.  I guess to start it, you
need to set a top object, then send a message to it.


> 
> >   Lisp books typically
> >have these kinds of examples.
> >
> >The real eye-opener, for me, is that the complete
> >algebra for this system is just the msg send, isn't
> >it?  That is the only computation within this space
> >that takes place.
> 
> That's the frame. The algebra is the choices of
> "polymorphisms" in 
> the base design.

oh, of course (looks around)...
 
> The term "fields" here is used differently. And the
> compact form of 
> Maxwell's equations as we know them today is due to
> Hertz, who used 
> "operators" like div, grad, and curl to collapse
> many separate 
> differential forms.

oh, yes I remember those, and they are easy once you
know how those operators work.  Sadly, I have
forgotten.


Cheers,
Rob