Sunday, December 2, 2007

ooPSLA 2007 - part 4b

ooPSLA 2007 in Montréal, Canada - Oct. 21 to Oct. 25, 2007 (this is the part 4b (and last) of my braindump on ooPSLA 2007. Please see part 1, part 2, part 3, and part 4a)
Another side effect of attracting so many experts in any one field is that ooPSLA typically ends up having great panels. This should be expected since, almost by definition, one becomes an expert in academic settings (and to some extent in industry as well) when you do good work but also can defend it when placed in unfriendly environments---where colleagues with criticisms or alternative approaches and ideas.

This year was no exception; tons of really cool panels with many well respected experts and famous figures. I attended four panels and want to briefly highlight three: Simula 67, Silver bullet, and SOA.

Simula 67 panel
Forty years ago the language Simula 67 was introduced to help facilitate the modeling of reality and thus allow help with simulations and other computing tasks that try to replicate real life. This led to what Brooks called the most effective advance toward a silver bullet in computer science we have ever known: object-oriented languages and systems.

(Photo credit: self with iPhone.  From left to right, Guy, James, Bertrand, Anders, and Ole)

To celebrate the 40th anniversary of this historical language, Joe Yoder, panel chair at ooPSLA 2007, managed to get in the same room:
  1. Guy Steele Jr. and James Gosling, both of Sun Microsystems, and co-creators of the Java programming language;
  2. Bertrand Meyer of ETH Zurich and creator of the Eiffel language and;
  3. Anders Hejlsberg of Microsoft and main creator of the C# language;
  4. Ole Lehrmann Madsen of Aahus University and co-creator of the Beta programming language.
Like most such 'roads' down memory lane, a lot of it was good old reminiscing time, however, a few poignant moments marked my memory. First, someone commented, to wild applauds, something to the effect that ooPSLA has become too much of a collection of papers adding or analyzing some aspect of the Java programming language and that needed to change. I have attended enough ooPLSA to remember when the program was not mostly about Java extensions and the likes. As a proponent of the virtues of dynamic languages and as a happy polyglot, I could not agree more.

The last thing that stuck to my mind is Ole's recollection of the grander than life Kristen Nygaard (the creator of Simula), who singlehandedly not only co-created the field, and in so doing impacted computer science profoundly, but Ole also reminded us that Kristen was an important figure in his country's politics and helped shape his country's position with regard to the European Union. I could not help but notice that such multifaceted characters are nowadays becoming rarer and rarer.

Silver bullet panel
An interesting, fun, and at time comical panel "celebrating" our resignation that the No Silver Bullet thesis is still very much in effect and that years of efforts have hardly put a significant dent into making good software easier to create. There was a slew of panelists, which made it a bit diluted, but noteworthy attendees were David Parnas of the University of Limerick, Ricardo Lopez of Qualcomm, Martin Fowler of ThoughtWorks, as well as the Silver bullet man himself, Fredrick P. Brooks of the University of North Carolina.

Two memorable moments are worth discussing. First, Brooks gave a short summary of his Silver Bullet arguments and how over the year it has still remained valid as well as how it is sometime misconstrued. Brooks attempted to remind us that his primary argument, though empirically deducted, has a strong mathematical foundation... I am less convinced of the latter, though, now that I have a brand new signed copy (my third) of the book, maybe it deserves some torn-age by a third read of the text.

The other interesting tidbit to mention is really the courage of 
Martin Fowler (ThoughtWorks, Inc).  After an 
interesting metamorphosis that lasted almost a full minute, Martin reemerged as the storied werewolf on Brooks's book cover. What transpired was Martin playing the mythical beast that never seems to die. Parnas mentioned that we are able to hurt the werewolf with lead bullets but nothing significant to make him disappear. It took courage for Martin to do this and keep a straight face (voice) and at time he
 came across with satire of Colberesque quality... However, the act also became boring at the end and at times distracted from the important themes and the overall conversations while also limiting the time of other panelists.
(Photo credit: self with iPhone of Silver bullet panelists.  Notice Martin's werewolf costume on far right)

SOA panel
The final panel I want to mention is the SOA panel which assembled a group of industry representatives such as Kerrie Holley (IBM Fellow), Dave Thomas (Bedarra Research Labs), Brenda Michelson (Elemental Links), Nicolai Josuttis (IT Communications), and John deVadoss (Microsoft).

After the typical short introduction by each panelists, which saw a regurgitation of your typical SAO kool-aid, though this time with a focus on the importance of looking at the business-side of SOA; that is, separating the architectural concepts from implementations (e.g., Web services or WS-*), and the importance of standards. What followed was a series of interesting "friendly jabs" (pointed questions and comments :-) between panelists and audience members. Since I was part of the conspirators, I'll limit my comments simply to say that I will always attend any panel where Dave Thomas is present. While I don't agree with everything he says, Dave's flamboyant character and speak, along with no-BS attitude, generally results in the truth being spoken...

Of the other panelists, Kerrie was maybe the most calm, cool, and collected and he stayed on message. As far as I could recall, Kerrie's message was that SOAs must reflect the business imperatives and be a projection of the business functions of an organization. While I would agree in theory with with this vision, in my opinion this aspect is less important to debate since as in most interesting ideas, the devil is in the details... Since some of the discussion also centered around the ageless REST vs. WS-* and company implementations, the details argument becomes more interesting. For me, this is analogous to saying that a BMW M3 and Ferrari F350 are sports car. At various levels they are, however, the latter is THE sports car by excellence; the differences dear Watson is in the details...

Parting thoughts
Visiting Montréal outside of the winter is always a treat. A wonderful city with kind folks working hard to hold on to their disappearing roots and culture. As a native french speaker, I always enjoyed to visit places where I am addressed in my mother tongue and also get to meet
 many of my countrymen. The Haitian immigrant population in Montréal is substantial and many of my old high-school and neighborhood friends have made the city their home. Got a chance to also visit many young cousins that I had not seen in 15 years or more.
While the conference allowed me to go down various memory lanes, the best parts and may be most memorable moments, were getting to meet, chat, argue, and even friendly fights, over wine and beer, with colleagues from both academia and industry. Just to name a few, and as a recommendation to never miss meeting the following if you are at a conference where they are presenting or are just attending: Eric Mejyer of Microsoft, Gregor Hohpe of Google, Brian Foote of Industrial Logic, Dave Ungar and David Bacon of IBM Research, as well as (who can forget) Dave Thomas of Bederra Research Labs.
(Photo credits: Self with iPhone of David presenting the best student paper award during the McCarthy keynote and of Dave at the hotel lobby)

No comments: