[cap-talk] Butler Lampson's upcoming talk

[Sam Mason]

On Wed, Apr 09, 2008 at 09:23:41AM -0700, Jed Donnelley wrote:
> At 08:00 AM 4/9/2008, Winnie Cheng wrote:
> >Speaker: Butler Lampson
> >Date: 4-17-2008
> >Location: MIT Room 54-100
> >
> >Gold and Fool's Gold: Successes, Failures, and Futures in Computer
> >
> >...Other promising ideas have not worked out: capabilities, distributed
> >computing, RISC, and persistent objects. And the fate of some is still
> >in doubt: parallel computing, formal methods, and software reuse. The
> >Web was not invented by computer systems researchers. In the light of
> >all this experience, what will be exciting to work on in the next few years?
> 
> Ha!  I wish I could be there.  I love the way he turns opinion
> into "experience".

I was under the impression that RISC was well and truly alive.  Most of
the newer (desktop/server) processor designs (i.e. PowerPC and Cell)
seem to be basically RISC designs.  I also thought that most embedded
processors were RISC as well, except maybe the really tiny ones.

> OK, I know his opinion on capabilities and I can see why his 'experience'
> suggests they didn't work out.  He also puts "distributed computing"
> into that category?

Yes, I was surprised about that as well.  He refers to the web later on,
isn't that a big distributed computing environment?

> I'd be quite interested to hear his story on persistent objects
> (independent of capabilities).

Me too!  Persistence, as was (and probably still is) found in languages
like Java, tended to do nasty things.  I believe that was just because
there was a fuzzy boundary between the persistent/non-persistent worlds,
if persistence was done properly then things work wonderfully.  Google
Mail is a great example, it goes off and saves its state transparently
to the user.  Is this why he qualified "persistent" with "objects",
to say that the system should present a persistent interface, yet the
implementation will need to go off and handle the messy details of that
facade, always using application specific tools/rules.

> Also he questions parallel computing?  How odd.  We have so many
> scientific computing systems with 10s to 100s of thousands of
> processors seemingly quite "successfully" churning out modeling
> results with no apparent alternative at present and Butler questions
> whether parallel processing is gold or fool's gold?  I wonder if
> he's questioning the non determinism aspects of current systems?

Maybe Lampson was just going for a controversial abstract to get more
interest/publicity for the talk?


  Sam