The Best Advice

I periodically like to quote Calvin Coolidge on the subject of persistence. I gave the following advice to my son tonight and thought I should share it with all of you:

"Nothing in the world can take the place of persistence. Talent will not; nothing is more common than unsuccessful men with talent. Genius will not; unrewarded genius is almost a proverb. Education will not; the world is full of educated derelicts. Persistence and determination alone are omnipotent."

Visited Countries

OK, I'll fall in line with the rest of the geek crowd and post my World66 results:




The US:




and, sadly, just a bit of Europe:

The Mythology of Star Wars

Bernadette and I saw Star Wars Episode III over the weekend. I finally get it! For years I wondered why the great mythologist Joseph Campbell was so enamored with Star Wars. Now I understand.

Campbell, now dead, was friends with George Lucas and often spent time at Lucas' Skywalker Ranch. Campbell spent his lifetime studying the similarities between human cultures and was responsible for identifying many of them for the first time. His book The Hero with a Thousand Faces, for example, notes that hero myths follow a similar pattern around the world. A hero (any hero) starts off easy, goes into the underworld where things are tough and comes out the other side transcendent. That is, the basis for a good story anywhere is pretty much the same! Christopher Vogler even used this phenonenon in his guidebook for writers, The Writer's Journey: Mythic Structure for Writers. Star Wars clearly has this aspect in its first three stories (Episodes IV-VI).

Campbell's wonderful four-volume study of worldwide mythology, The Masks of God, culminates in his theory that post-modern humans (that's us, folks) are making our own mythologies. This is described in Creative Mythology : The Masks of God, Volume IV. He has cited Star Wars as an example of this phenonenon.

The part of Star Wars that confused me was how the hero motif fit in relation to the latter episodes (I - III). Darth Vader is really the bad guy and very much anti-heroic. Even though he had his moment in the dying gasp of Episode VI, it just didn't seem to work for me. Then I re-read Campbell's take on Judaism's great memetic leap, the concept of the Chosen People. Campbell notes that the myth of Moses, which has no historical or archeological basis (and not for want of looking!) is really a hero myth with a twist; the Jewish Partriarches go "down into Egypt" and the transendent hero which emerges on the other side is not them, but the people as a whole. The authors, as Campbell said, understood exactly what they were doing. I think Lucas did the same thing.

Star Wars, for all the rational criticisms levied against it (compare to rational criticisms levied against the Bible or the Koran), makes perfect sense as a hero myth with a twist. Episodes IV - VI are a classic hero myth, but incomplete. The completion (Episodes I - III) fill in the gaps and provide the twist: Star Wars is now a hero myth which takes two generations to complete. Anikin Skywalker goes down into the underworld and his son emerges transcendent. Campbell would have loved that!

US Venture Capital is Dead (at least for now)

Respected venture capitalist Howard Anderson (founder of Yankee Group and two VC funds, now teaching at MIT's Sloan School) has written a succinct description of the state of US VC. His bottom line: he's out and staying out.

Anderson's analysis makes sense to me. If he is right, then there was no point in starting up in the last few years and no point in starting up in the next several. Only time will tell whether this state of affairs is structural and thus long-lived (as Anderson thinks). If so, technology entrepreneurs had better start thinking of new ways to fund the movement of their ideas into the mainstream.

For now, my solution to that problem is to use a combination of Open Source Software licenses for my work, consulting around its use and deployment and subcontracting when I need a larger team. The use of Open Source licenses allows an individual to prove oneself and differentiate oneself from the rest of the pack; that fact that it extends the Open Source phenonenon is an economically-driven by-product, albeit one which is good for the community. Subcontracting leaves me without the overhead of a startup, soomething I am glad to avoid after the last decade of making payroll.

I thought Anderson's last thought was right on target: "Ever wonder what we did for a living in early-stage venture funding? I bet you think we spent the day searching for the next insanely great company. But we spent most of our lives in endless meetings with people who were lying to us: scientists who swore that their patents were solid and entrepreneurs who insisted that they had no competition. We lied right back at them: said our money was different." Hope springs eternal on both sides, and will ensure that many VCs continue to operate while they lose money. I hope I can keep that lesson in mind as I face new startup opportunities. There has to be a better way.

Thanks to Brian Sletten for the link to Anderson's article.

Impact of Northrop Grumman Purchase on Kowari Users

Several Kowari users have asked about Kowari's future since the purchase of Tucana's intellectual property. Please allow me to clarify in a public forum:

Northrop Grumman purchased, among other things, the copyright to Kowari (which makes them the copyright holder under Kowari's Mozilla Public License.

Informally, the MPL license ensures that licensees may use the code and make modifications. The applicable part of the license requires modifiers to submit modifications back to the original developer (formerly Tucana, now Northrop Grumman) if the modifications are distributed to any third party.

So, all is well for Kowari users. All you have to do is submit new modifications to the Kowari source code that you intend to distribute to Northrop Grumman instead of Tucana. No big deal. I intend to blog it when they decide who should get those modifications. Of course, IANAL, so check with one if you like.

Kowari and Sesame Compared

Chris Wilper did a nice comparison of Kowari and Sesame. The comparison mostly highlights the design criteria differences in the two projects: Sesame is nicely lightweight and easy on resources. Kowari is much faster on the queries and performed much better on the more intensive operations.

I will note that Chris used a version of Kowari which is a year old (for some reason) and was wrong about Kowari not supporting xsd#int. Still, kudos for publishing this interesting data.

Consulting to Northrop Grumman

Northrop appears to be quite serious about becoming a Semantic Web company. They have retained my consulting company, Software Memetics, to train them on TKS internals, consult regarding future releases, help define new features and even to help them with their relationship with Kowari.

Former Tucana engineer Paul Gearon is in the US now to start the source code turnover. Andrae Muys will be coming to the US next month to continue that effort. Frequent NoFluffJustStuff contributor and JavaOne speaker Brian Sletten and author/entrepreneur Darren Govoni and I will be conducting the long term strategy and feature work.

I expected them to ask for a source code turnover, but I certainly didn't expect them to make such a serious monetary commitment. Between the purchase of the Tucana IP, their significant internal spending and all this consulting, they have laid down some serious dough. It is awesome to see Semantic Web technologies making such an impact in the federal market. The big question on my mind is what they have planned for the commercial sector. Stay tuned.

Perl Syntax Error Not Caught at Compile Time

I have a bad habit of munging the syntax for perl 'open' commands. Instead of the correct syntax, e.g.

open (TESTFH, "< test.txt") or die "Couldn't open data file: $!";

I tend to misplace the right parentheses, like this:

open (TESTFH, "< test.txt" or die "Couldn't open data file: $!");

Consider this test script:

#!/usr/bin/perl
open (TESTFH, "< test.txt" or die "Couldn't open data file: $!");
while () {
  $content .= $_;
}
print $content or die "Couldn't print data: $!";
close TESTFH or die "Couldn't close data file: $!";

This code will generate an error like this:

$ ./test.pl
Couldn't close data file: Bad file descriptor at ./test.pl line 7.
$

That is, the file handle was never opened properly, but did not fail. The WHILE block does not execute and the print statement has nothing to print, but does not fail. A simple syntax error which should have been caught at compile time has promulgated through to a nasty and poorly-reported run time error.

Unfortunately, the use of the 'strict' pragma ('use strict;') does NOT catch this error at compile time.

This example was generated on Perl v5.8.6 for darwin-thread-multi-2level, but I have seen it on Linux as well. Earlier versions of Perl 5 have this problem. I haven't tested Perl 6 yet.

SWCASE05

A workshop on Semantic Web Case Studies and Best Practices for eBusiness called SWCASE05 will be held at the International Semantic Web Conference (ISWC 2005) in Galway. I have been asked to serve on the Program Committee by Robert Tolksdorf (Freie Universität Berlin), Alain Leger (France Telecom) and my colleague Guus Schreiber (W3C).

Maybe I'll go to ISWC this year; I hadn't planned on it, but the W3C Semantic Web Best Practices and Deployment Working Group face-to-face meeting will be held 4-5 Nov in Galway, the two days just prior to ISWC.

Mea Culpa: Spotlight is Actually Useful

I have been pretty critical of Apple's Spotlight and still have some UI suggestions for its designers (such as increasing the amount of information shown about individual results), BUT I have found myself slipping into using it recently with much success.

I couldn't find a contract this morning. Spotlight reported it as its top hit. I couldn't locate an email message in the folder where I thought I filed it. Spotlight found it. I couldn't find a Web page I thought I had bookmarked. Spotlight found it without trouble. Wow.

There have been a couple of times when I failed to find what I was looking for with Spotlight, but I have to say that I am more efficient with it. I should be very relieved; Spotlight is metadata-based searching, after all. If Spotlight failed to work, then the whole SemWeb thing would be questionable.

Well done, Apple, and pay no (OK, "little") attention to my earlier criticism.