"Smalltalk has some great opportunities ahead, but we can't just say we're the best, we have to innovate."
Stéphane Ducasse represents for the world of Information Technology what his namesake does for international cuisine. Both are French, perfectionists to the extent that they personally take care of the smallest detail and have both set the bar so high to lead the way in the field. In the case of this University dean of computer engineering, research in Smalltalk is his great project, a computer language with great development potential taught at Citilab right from basics. Ducasse is the creator of Botsinc, an environment for getting started in programming wellknown at home and co-responsible for Pharo, a new way of understanding and working with this language. He is currently preparing the next International Smalltalk Conference (ESUG 2010) which is being held from 11th -17th of September at Citilab.
Smalltalk was created as a programming environment that was ahead of its time.
Are those characteristics that now make it a very attractive option for development groups?
It's clear that Smalltalk was really ahead of its time back in the
80s, now you still have an environment where you program completely
by interacting with objects. What you see for web development is that you can
directly interact with your web-server and web-application, so for
example, with Seaside you can debug on-the-fly your web application,
that is a really interesting value. What you also see is the fact that
you can have access to all of the objects in the system, like the
stack, this is what allows you to
define solutions that are really not traditional. If you look at Seaside
for example, this is not something that you would do easily in
other languages, so those kinds of characteristics, they are really
interesting for making sure that we can innovate. Now, it is clear
that it would be nice if Smalltalk would have scripting syntax, and we are
working on that, for example.
To give you an impression from France, when we started the Smalltalk
mailing-list in '98 there were 3 of us, now there are 250, so it does
not mean that you have a lot of business, but you have business, and
i see more and more companies developping in Smalltalk. That is not
like the mainstream language but there are still people that take the
opportunity and i think that the web is really helping because people
can deliver fast their application and they can build fast flexible
applications too, can prototype them fast, so that's three
professional competitive advantages. I know a company developing in JSP (JavaServer
Pages) that completely switched to Smalltalk, although this is not like the growth
that there was in 1996 for example.
In '96 Smalltalk got a 40% increase every year, so, this (increase)
would not happen anymore, but I see also a lot of small companies
where people develop software and make a living- so that's good!
What impact are having educational initiatives based on the Smalltalk environment (Scratch, Botsinc, etc) in the education community?
What I saw is that Scratch, that is developed at MIT, is really having an
impact, so that is really nice. I reckon that they will continue the
E-toy distribution with the $100 laptop. It is also having a lot of
users. I have a friend who is developping a Dr. Geo II for the $100
laptop and he had a lot of downloads, a lot, so
this is very nice!
Now Botsinc, this is what I did,
it is more anecdotal because i never push more than I
should, and, yes, I think that they are doing a good job there.
Are companies using increasingly Smalltalk in its commercial developments?
What we saw is that when Java arrived it killed a lot of opportunities
and things like that. Since 2000 i think that it has changed, since
2000 we really see companies that are re-using Smalltalk, for example,
you should know that the AMD 64 bits that you maybe have in your machine
is developed in Smalltalk. Thirty per cent of the world market of
container-shipping overseas is controlled by Smalltalk systems.
JP Morgan, one of the founders of the Dow Jones, is also building
extremely complex banking systems with Smalltalk, and those companies
are really investing in Smalltalk. I know a lot of people that have
been hired by J P Morgan. I know companies that have been doing
simulation of petrol extraction so i think that the bad years are
over. We really saw that in ESUG with the amount of people that are
participating in the conference, and the years that it was the
Java-only dogma are over. Now everybody can use any dynamic
language. Smalltalk can regain from that
perspective. What we see is that more and more companies, like I said
at the beginning, use Smalltalk. This is not a huge increase, but you see all of
these projects that got frozen are hiring new people and starting new
projects and we saw that since '98, and '98 was a really bad year,
well, not a bad year but the community and conference were small. Two
years ago at Amsterdam there were 170 people coming from all over the
world, and the year before was similar and this year it should be the
same in Barcelona, so I don't see any problem there.
What are the challenges of ESUG in this edition, as the reference meeting for the Smalltalk user community?
The goal of ESUG is to make sure that we build a community where
people really meet each other. This is not a conference like an
academic conference, this is a conference with coeur where people
meet and circulate information within the community. That is the first
role of ESUG. I don't really know what will happen - but that will be
fun so you people should come! It is always a really interesting
conference.
Something really important that Alan Kay said was that the best way to
predict the future is to invent it, I think that the Smalltalk
community should go back to this state-of-mind, so, for example we work on
Traits. Traits are a kind of better Java Interface. Now the
Traits we developed have been imported into Perl 6, Scala uses a
version of them, they were integrated into Dr. Scheme - maybe
JavaScript will have them. What it proves is that we can invent
something and you can get an impact in Smalltalk.
What are the challenges for the future?
Now for the future there are challenges like: What is a secure
language? What is a secure, dynamic language? What is a secure,
dynamic and reflective language that can modify itself and still be
secure? Can we really have minimal systems that could be built
automatically from distributed sources? And things like that. So, what
are the better modular systems that you want to have for the future,
scripting syntax, the multicore? How do we approach multicore with
a system like that? IBM were doing some experiments and they used
Squeak, a version of Squeak to run on a 64 bits
microprocessor - and on each of these microprocessors they had a
Squeak running and communicating with the other virtual machines.
So these are the challenges for the future. I think that Smalltalk
really has a chance for the future but we should really innovate, not
just continue to say "Yes, we are the best".
