Where's the new Smalltalk?

The various recent discussions such as: Strongtalk's
strengths/weaknesses, should it be included in Squeak; the power of
eToys; Lisp's power; the Lisp Machine; what Smalltalk has brought to the
world of computing that is just now being recognized (syntax, typing,
bitmapped displays); on and on ---- got me wondering about today's research:

If Lisp/Smalltalk (and their significant predecessors) have contributed
so much to computing - and so long ago - wouldn't you think that
someone, somewhere, is developing the next significant contribution(s)?
What R&D (or just plain "D") is currently in the labs that you feel will
be a significant contribution to computing in the near future? (I
know... tough question if the work is in a secret lab. And, who can
predict the future?)

I've heard Alan say that he wished that they had something else, other
than Squeak, to use for Croquet. But, Squeak is the best thing right now
for Croquet. Why? Is the answer merely a funding issue? Are schools at
fault (I've also heard Alan lambaste Stanford regarding their Java
curriculum) ? Do gov grants ignore the little guy in favor of
universities?

What's your take on it?

Ruby is the first language in 30 years I actually find interesting compared
to Smalltalk.  Mostly it is enough like Smalltalk with a few interesting
twists, and has much better momentum in the mainstream.  It of course has
poor performance because it is not supported by a large corporation yet.
But, the ease of working at the meta class level and the approach to class
specification is very interesting.  All statements are executable, even
those that incrementally create a class definition.  So rather than a large
multi-part method send as in Smalltalk each aspect of the class definition
is a separate method send to the class object itself, with full run-time
class modification supported.  Now all it needs is a good VM and a GUI!

As to the real question you raise, I have not seen interesting programming
language research in 25 years.  I think things are heating up for the next
round of language experimentation however.  All the signs are there for
another period of broad experimentation reminiscent of the 70s.

Michael

> If Lisp/Smalltalk (and their significant predecessors) have
> contributed so much to computing - and so long ago - wouldn't
> you think that someone, somewhere, is developing the next
> significant contribution(s)?

No, but I'd hope.  I'm not old enough to say much, but when I started
studying where all the cool things were, I found my way back to Lisp and
Smalltalk very quickly, it seems that other languages have contributed very
little that these languages hadn't already done, I think the past 20 years
or so have maybe been a programming dark age, all the progress has been in
hardware, not software.  How often these days does something like Zerox Park
happen, just turning lose a bunch of brilliant guys to build whatever they
can, and letting them throw it out and start over a couple of times to get
it right?

> What R&D (or just plain "D") is currently in the labs that
> you feel will be a significant contribution to computing in
> the near future? (I know... tough question if the work is in
> a secret lab. And, who can predict the future?)

I think Ruby, as well as being interesting, has kicked Smalltalk in the ass,
forcing it to start changing, is Traits not a direct response to ruby's
mixins?  Ruby seems an interesting mix between Smalltalk and Lisp,
definitely something to keep an eye on.  Croquet, along the lines of E,
brings concurrency into the language, a very interesting thing imho, I'm
looking forward to see what that becomes and how it affects other
Smalltalks.  I think type inferencing has a place in whatever the cool
language is.  .Net, not any of the languages, but the runtime, provides a
common type system and object model, that many languages can share, I find
that rather interesting, especially as more dynamic languages get ported to
work with it, such as the recent IronPython.  

>
> merely a funding issue? Are schools at fault (I've also heard
> Alan lambaste Stanford regarding their Java
> curriculum) ?

As he should, teaching Java in the schools makes the schools little more
than a vocational school, and kills interesting research, which he so
bluntly pointed out (I dug that).

Squeak, as far as I can tell, is the new Smalltalk, it seems to be the base
that new stuff it attempted on.  Traits is certainly a huge language change,
and it has just been included, pragma's were also recently added, something
I think gained popularity in .Net, then Java.  I'm not sure what else is
coming, but it's sure fun watching it happen.

I'll provide some links to the talks, but basically, there's a
principle of a general migration that's always happening that one must
be aware of.

Tasks that were accomplished by a general purpose language in the past
become accomplished  a tool for those less skilled (leveraging the
benefits of scale). Over time, general purpose languages have turned
into ones with strengths for certain domains. Next turn of the crank,
tools will generate Domain Specific Languages augmented by tool use.
In the future, the logic will present itself to the knowledge worker
in the format that person is most comfortable with and not be the same
for each person. It'll be a mixture of graphical tools, text, and
other media and will change and adapt to the person's changing skills
and the changing agreements and pacts of the groups the person
identifies with. Rather than the person adapting to the language and
tool, the language and tool will adapt to the person, to the domain
they are working in, and to the problem they are solving.

cf.
- Intentional Programming by Charles Simonyi
http://www.intentsoft.com
- The Semasiology of Open Source (Part 2) by r0ml Lefkowitz
http://www.itconversations.com/shows/detail662.html
- Microsoft's Software Factories
(see/hear Jon Udell's interview)

This is more like how spoke language actually works and how text in
books have changed over the centuries to convey more meaning than just
letters & phonemes.

The pair programming and multi-media and annotation aspects of Croquet
might be just the thing to boot strap the next thing.

Cheers,
Darius

To see the migrations and cycles also listen to:

- Clayton Christensen's disruptive technology in The Innovator's
Dilemma and The Innovator's Solution
http://www.itconversations.com/shows/detail135.html

Language is about folding knowledge into one's memory and manipulating
knowledge by the labels one uses that are created by the folding of
that knowledge. Folding knowledge is for the purpose of saving time
and minimizing complexity.

What one knows and what one remembers will define what labels &
symbols a language will mold itself around and become its next shape.

Just like children, we change language and add to it as we use it. We
change ourselves by the language we use and the ideas contained in it.
And, we meaningfully interact with varied skilled adults and
multi-lingual adults at the same time.

Someday software languages will be that flexible.

18r
Darius

Children never change from language 1.9 to language 2.0. Nor do they
try to make sure everything they say fits within each standard
presented to them.

Darius Clarke wrote:
> To see the migrations and cycles also listen to:
>
> - Clayton Christensen's disruptive technology in The Innovator's
> Dilemma and The Innovator's Solution
> http://www.itconversations.com/shows/detail135.html
interesting.. I will. Thanks!

>>>>> "Ramon" == Ramon Leon <[hidden email]> writes:

Ramon> I think Ruby, as well as being interesting, has kicked Smalltalk in the ass,
Ramon> forcing it to start changing, is Traits not a direct response to ruby's
Ramon> mixins?  Ruby seems an interesting mix between Smalltalk and Lisp,
Ramon> definitely something to keep an eye on.

And don't forget Perl6, moving along rather quickly now, which is taking the
best of Ruby, and adding prototype-based inheritance into the VM, which will
make hosting and interoperating with Javascript or Smalltalk rather
straightforward.

--
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<[hidden email]> <URL:http://www.stonehenge.com/merlyn/>
Perl/Unix/security consulting, Technical writing, Comedy, etc. etc.
See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

> Is the answer merely a funding issue?

     Yes.


-C

--
Craig Latta
http://netjam.org/resume

On 9/19/06, Brad Fuller <[hidden email]> wrote:
> Why? Is the answer merely a funding issue?

Not only that. I think that for the last decade, the biggest bummer is
that language research has actually been Java research. I know of a
*lot* of theses and dissertations on "continuations in Java",
"delegation in Java", etcetera. Some guy hacks up the VM, writes a
paper, gets his degree, and everything he did is forgotten because
it's too "out there" to be picked up by the Java standardization
process. And it hardly counts as "R", more as "D", anyway.

BTW: I'll really like Ruby by the time it gets alive, iow supports an
image. Until that time, it's just another language in the
Perl/Python/... fold.

Darius,

I really enjoyed your comments and observations.

Ron