The Second Annual YAPC
Day 1 |
Day 2 |
Day 3 |
Wrap-up
Thursday at Second
Annual YAPC (Yet Another Perl Conference) started and ended strong,
with two great talks. Larry Wall opened the day with one of his humorous but
penetrating talks relating larger issues in his life to his understanding of
the importance of Perl. Damian Conway rocked the conference with a rollicking
exposition of the application of quantum mechanics to Perl.
I really can't do justice to Damian's talk. It was too dense, too polished,
and too hilarious to be summarized. It was one of those great moments that
only the people who directly experienced it will be able to appreciate
fully. I will tell you only that, in a little more than an hour, Damian
gave a brief history of Quantum physics; related it to computer science;
and suggested how it could be implemented in Perl with the addition of only
three operators. Okay, since you asked, the operators are *any*, *all*, and
*eigenstates*. You had to ask.
Damian's talk was part performance art, part intellectual synthesis, part
college lecture, and part Perl Wizardry. It left people gasping from trying
to keep up. Find someone who attended the conference; I can't reproduce its
effect by reporting.
Larry's talk was typical of the kinds of presentations he makes at these
conferences. He related his understanding of
Pilgrim's
Progress, a book he first read in his youth. Larry's talks don't dwell
on the features of Perl or even its futures. He wants the Perl community to
understand the philosophy and values behind the development of Perl.
Larry asserted that Christian, the Pilgrim in John Bunyan's book, shared
many experiences similar to those faced by Perl programmers. First, Larry
said, they need to acknowledge that they've got a problem. They need to
seek the advice of their mentors. They need to look beyond the easy answers
(the "quick and dirty"), sometimes taking a step back and considering a
larger, more general solution that requires more effort.
Then Larry addressed, however obliquely, the controversy and discord that
affects some parts of the Perl community. "I see signs of turbulent flow in
the Perl culture," he said. Pilgrim's Progress stresses the value of
ongoing friendship; Christian needs the help of his lifelong friends to
attain his goals, just as Perl programmers rely on each other to use and
improve Perl.
Larry's final advice was: "Live your life with a purpose. Life is an
open-ended project, like Perl."
There were a number of informative talks throughout the day. Dick Hardt of
ActiveState and
Mark-Jason Dominus brought us up-to-date on the features of Perl 5.6. The
major features include Unicode support and the implementation of fork () on
Windows platforms. (Fork will also work in MacPerl.) Other improvements are
weak references, 64-bit support, and greatly expanded and improved
documentation and diagnostics.
Chris Nandor reported on the status of
MacPerl.
MacPerl's goal is to reach the same level of functionality as is available on
Unix and Windows systems. For this goal to be reached, Apple has to update a
POSIX library called Gusi. In the meantime, Nandor announced that MacPerl will
soon be part of the Perl distribution.
Nat Torkington gave two thougtful talks. I'm running out of space, so I'll
only mention one of them today. He asked a series of interesting questions:
Why does Perl lack CORBA support? Where is the Perl equivalent of Beans?
Why does Perl XML development lag? Why has Unicode support been so difficult?
Nat believes that Perl hasn't needed to develop these kinds of technologies
because (1) They are more complex now than their practical value justifies,
and (2) Perl has simpler solutions to the same problems. He makes a good
point; Perl development follows the needs of its users, and Perl programmers
don't have problems that they can't solve in Perl (that these other
technologies would solve, that is).
Damien Conway's talk yesterday made a similar point. Nobody who listened to
Damian's talk today could imagine that he would shrink from complexity; but
his point about Object Oriented programing in Perl was that it could be
simple and easy. Talk of OO programming makes people furrow their brows and
get out their computer dictionaries; but in Perl, according to Damian, a
programmer needs to follow only three rules, to wit:
- To create a class, define a package
- To create a method, define a subroutine
- To create an object, bless a referent
That's it: a simple solution that handles the problem without the baggage
of more formal languages.
My one concern is that, although this simplicity suits the Perl community,
it won't add adherents from people interested in XML and CORBA who are
looking for a programming language. The number of Perl users grew
exponentially because Web programmers wanted a language that worked well
with CGI, and Perl was their choice. If XML catches on, will XML
programmers bypass Perl to use other languages with more XML support?
In the post-dinner trivia contest run by the Philadelphia Perl Mongers, the
team I was on, the Schrodinger Cats, came in second, aided by judicious use
of Chris Nandor's WaveLAN and the
Google search
service. I won a baseball cap, a nice berry and tan one with a penguin
on the crown. In Pittsburgh, that makes people assume I'm a hockey fan, but
we know better, don't we?
Once again, I thank my co-correspondent O'Reilly associates, Joe Johnston
and Tim Allwine, for their contributions to this report.
--Frank Willison
Visit
perl.oreilly.com for a complete
list of O'Reilly's Perl books.
Return to: Frankly Speaking
| 
