First stop: OOPSLA '97. Erich gave a tutorial on "advanced design
with patterns and Java":
http://www.acm.org/sigplan/oopsla/oopsla97/fp/fp23.html
But that page has no link back to him. So we continue. Next stop:
Object Technology International.
http://www.oti.com/CORP/stcconf.htm
in 1997 where he gave a talk on "Framework Development with Pattens and
Interfaces - the Feel of Java vs. Smalltalk." Mighty buzzwordy, but
okay, at least we know he's still in the pattern area. But this
document has no link to Erich, either. Or does it? A search of OTI's
complete Web presence brings us to:
http://www.oti.com/press/zurich.htm
which is a press release June 17, 1997:
> Dr. Erich Gamma, an internationally respected expert in the software
> industry, has joined OTI as the STC Technical Director. Dr. Gamma
> previously held key technical positions at IFA, Taligent and
> Ubilab. His work on design patterns and frameworks is internationally
> recognized. Dave Thomas, President and CEO of OTI noted, "We consider
> ourselves fortunate to have found the ideal candidate to lead our STC,
> and we have already assembled an excellent team around him. Attracting
> and keeping the best Swiss talent in Switzerland is a critical success
> factor for OTI and our partners. We have a tradition of building
> bridges between industry and research and look forward to building
> advanced products in Switzerland. Our Swiss partners have the
> foresight to invest for the future so that they are ready for the next
> generation of software applications."
But it still doesn't tell us what he's currently working on. Back to
Altavista. I find the "object oriented FAQ" off Ralph Johnson's page
http://www.cyberdyne-object-sys.com/oofaq2/
but find no mention of Erich. Another link back to OTI...
http://www.oti.com/CORP/employ.htm
tells me
> "I came to OTI because of the amazingly good image it has in the
> industry for delivering reliable, high-quality software that's shipped
> on time," says Erich Gamma, Technical Director of OTI's Software
> Technology Center in Zurich, Switzerland.
>
> "Our strongest research ties have a lot to do with our good
> reputation," says Gamma. "We have a lot of affiliations with
> universities and professors, and they're constantly feeding us new
> ideas from their areas of research. A lot of other companies don't
> really listen to that at all, but that's what's made us the success
> that we are today."
Ugh. Around Altavista slot 80 out of 1000 we find PhD student Bernard
Wagner:
http://www.ifi.unizh.ch/groups/mml/people/bwagner/bwagner.html
and he wrote a paper on ET++ with Erich in 1996. But no links to Gamma.
When we try to find more info about ET++
http://www2.graceland.edu/~hasman/23130/cppfa207.htm
we discover that they think Erich is at Taligent. Clearly this is out
of date information. Back to Altavista baby.
Even ECOOP '96 shows that Erich has no Web page:
http://www.ifs.uni-linz.ac.at/ecoop96/organization.html
Sigh. Even Grady Booch won't tell me where to find Erich:
http://www.rational.com/support/techpapers/patterns.html
> I'm generally not an excitable person. My idea of a great summer
> afternoon is reading a book alongside some crisp Colorado mountain
> stream, miles away from anything containing a transistor. Not quite
> the excitement of bungee jumping or, perhaps even more intense,
> subclassing from a deep class lattice while mixing in a couple of
> other abstract base classes to achieve some desired polymorphic
> behavior. Neither activity is for the faint of heart. However, there
> is one emerging development in the world of objects that does
> genuinely excite me, and that is the methodical identification of
> patterns.
>
> First, let me introduce some of the key players. There are a number of
> centers of gravity in this great search, one in the United States, and
> another in the United Kingdom. In the U.S., one such group includes
> the diverse talents of Ralph Johnson (University of Illinois), Erich
> Gamma, and John Vlissides (both from Taligent), and Richard Helm
> (IBM). Although not a part of this association, Mary Shaw, Rebecca
> Wirfs-Brock, and Peter Coad have independently contributed to the
> study of patterns. In the U.K., Bruce Anderson (University of Essex)
> has catalyzed significant study into the codification of patterns, and
> at OOPSLA '91 and '92 conferences, established workshops focused on
> the creation of an architecture handbook, whose purpose is ultimately
> to serve as a catalog of patterns.
>
> The notion of patterns is significant, especially to all things object
> oriented, because it represents a higher leverage form of reuse. The
> search for patterns encompasses far more than finding the One Perfect
> Class, but rather, focuses upon identifying the common behavior and
> interactions that transcend individual objects.
>
> The study of patterns is not unique to software. Indeed, the
> importance of patterns has long been recognized in other technical
> disciplines such as biology, chemistry, physics, and architecture. In
> a manner of speaking, the field of software engineering is just
> beginning to awaken to the importance of patterns. For example,
> Herbert Simon, in his study of complexity, observed that "hierarchic
> systems are usually composed of only a few different kinds of
> subsystems in various combinations and arrangements." In other words,
> complex systems have common patterns. These patterns may involve the
> reuse of small components, such as the cells found in both plants and
> animals, or of larger structures, such as vascular systems, also found
> in both plants and animals. In his classic work, The Sciences of the
> Artificial, Simon went on to illustrate how these patterns manifest
> themselves in social and biological systems, and how the existence of
> such patterns help to simplify their inherent complexity.
Okay, let's look at Erich's "Extension Objects Pattern"
http://siesta.cs.wustl.edu/~schmidt/PLoP-96/gamma.txt
> The Extension Objects Pattern
> Erich Gamma
> IFA Consulting
> Ceresstrasse 27
> CH-8034 Zurich
> gamma@ifa.ch
>
> The Extension Objects pattern describes a way to extend a class with
> additional interfaces and behaviour. It encapsulates an extension into
> a separate object and provides a mechanism to negotiate for a
> particular extension. The Extension Objects pattern can be applied
> when you need to support the addition of new or unforeseen interfaces
> to existing classes.
Nope. And he's nowhere else in PLoP '96:
http://siesta.cs.wustl.edu/~schmidt/PLoP-96/original-papers.html
Nope, nope, nope. I even used metasearcher
and found little more than a Javaworld article referring to the Design
Patterns book:
http://www.javaworld.com/javaworld/jw-12-1997/jw-12-perspectives.html
> I recently read Design Patterns by Erich Gamma, et al. It's a great
> book, so if you haven't had time to look through it, I strongly urge
> you do so. (See Resources for more information on the book.) The book
> is a catalog of object-oriented patterns that, through experience,
> many object-oriented (OO) designers have developed and used
> effectively. What I found most interesting about the book wasn't the
> content but how much of the Java API uses these design patterns. This
> exemplifies the thoroughness and depth of knowledge demonstrated by
> the designers of the APIs. What I find tragic is that even though I
> and many of my peers understand this point, many in the computing
> community simply don't get it. Countless commentators, editors, and
> analysts are busy predicting the demise of Java or limiting it to an
> over-hyped fad that won't amount to much. Few of the "industry
> experts" really understand technology, particularly Java, well enough
> to make objective observations.
>
> You've probably heard of Robert X. Cringely. He wrote the Triumph of
> the Nerds, which was turned into a mini-series on PBS. He has a
> commentary page, called "I, Cringely," on www.pbs.org. (See Resources
> for a link to his site.) I respect Robert's opinions; he is one of the
> few who actually researches a subject before writing about it. I sent
> e-mail to him to tell him that I generally agree with his commentaries
> but disagreed with his views on Java. He sent me this response:
>
> > You can't write anything in Java that can't be written (and run
> > faster) in C++. Java may well be the next important phase in software
> > development, but it is only a phase. Java is the Pascal of the 90s.
Wonderful. I start out looking for Erich Gamma and end up with a
Cringely quote on Java being the Pascal of the 90s.
I'm guessing that Erich and his new project doing thin X clients are
just not on the Web. Too bad, I would have been interested in reading
about them.
----
adam@cs.caltech.edu
The pendulum has gone full circle.