Sunday, April 30, 2006

Physicists Get Water to Flow Uphill

Physicists working at the University of Oregon's Linke Group have harnessed the Leidenfrost effect to make water flow uphill using only heat and a carefully-shaped surface. The BBC has the story and a video.

The plan is to harness this effect to automatically cool microchips.

Just when I thought all interesting physics was non-classical...

Wednesday, April 26, 2006

USB Memory Devices and Mac OS X

For many years (about 13) I had a low user id number on my various Unix boxes of 112. This was a hangover from the old days of Sun OS 4.1.3 when anything under 100 was restricted. I carried this uid from Sun OS to Solaris to Linux to Mac OS X because I routinely used NFS (which operates on the uid) and didn't want my systems administrators to have to change my uid on all file servers.

I finally made the switch to a high uid (above 500 - the new standard) when I bought a PowerMac G5 last year. That made switching the files from my Powerbook G4 much more difficult, but not impossible. I thought I was done.

Then something strange happened. In a seemingly unrelated series of events, my three USB memory sticks started to leak capacity. A particular 1 GB stick would only hold 40 MB, even though there were no files on it! I found the problem today. My old uid had raised its head one last time.

Mac OS X segments files on removable media by - you guessed it - uid. The file structure on the 1 GB USB stick had these files on it:
[macaw]$ file .*
.: directory
..: sticky directory
.DS_Store: data
.TemporaryItems: directory
.Trashes: directory
._.TemporaryItems: AppleDouble encoded Macintosh file
._.Trashes: AppleDouble encoded Macintosh file
The .TemporaryItems and .Trashes directory were full of files owned by my old uid - which my new uid had no right to delete. The solution was to use sudo on the command line to simply remove those directories and their associated files (including the top-level ._.* files). Ejecting the device and remounting it showed a comfortable 976.1 MB.

Kowari is Forking. Welcome, Mulgara!

The unofficial Kowari fork is about to become official. The new name will be Mulgara, another Australia marsupial which is remarkably similar to a Kowari :)

The Mulgara Web site is partially up. The domain is obviously registered and an Apache server is responding on a Fedora OS. Check back soon for more news!

Andrae's Subversion repository is rumored to be opening for public submissions shortly, as well. In the meantime, source and binary code for Kowari "version 1.2" are available.

It will be a relief to get this project back on track given the substantial customer interest expressed since January. I look forward to participating in the Mulgara project.

New WordNet in RDF

The latest version of Princeton University's WordNet has been made available in RDF by the WordNet Task Force of the W3C's Semantic Web Best Practices and Deployment Working Group. The new RDF version is here.

This conversion is based on the newest RDF/OWL Representation of WordNet document, now being considered for First Working Draft status.

Monday, April 24, 2006

China and Asymmetrical Warfare

Victor Corpus, a former Phillipine armed forces chief of intelligence, wrote an interesting peice for the Asia Times Online entitled, "If it comes to a shooting war ..." about a potential shooting war between the US and China. There were ideas of interest; (a) Foriegn military officers are seriously considering the consequences of the current US policy of pre-emptive strike and (b) America's military might is insufficient to achieve its stated goals. The latter point was made by way of highlighting American forces vulnerability to asymmetrical warfare.

One aspect of assymetical warfare was certainly demonstrated on 11 September 2001. A handful of terrorists conducted a very effective bombing raid on New York City, killing thousands. Nations, however, can also conduct asymmetrical warfare. The Center for Nonproliferation Studies reported on China's anti-sattelite weapons capability in this article: China's Space Capabilities and the Strategic Logic of Anti-Satellite Weapons

Thanks to Nova Spivack for blogging a pointer to the first story.

Current Kowari JARs ("version 1.2")

Andrae's company Netymon has kindly hosted an unofficial Kowari fork in lieu of a resolution with Northrop Grumman. Andrae has posted source and binary code for Kowari "version 1.2", which is what he is calling the natural completion of the release originally scheduled for January, 2006. Hopefully this will ease the pain of those waiting for an official release.

Netymon is also hosting source and binary for a version of Kowari with Andrae's latest iTQL extension.

Tuesday, April 18, 2006

Friday, April 14, 2006

Defining N-ary Relations on the Semantic Web

The Semantic Web Best Practices and Deployment (SWBPD) Working Group has published Defining N-ary Relations on the Semantic Web as a Working Group Note.

In Semantic Web languages like RDF and OWL, a property links two individuals or an individual and a value. The Note presents patterns and considerations for representing relations between more than two individuals or values.

Thanks to the Ontology Engineering and Patterns Task Force for producing this Note.

Thursday, April 13, 2006

Kowari's Legal Status

One positive outcome of Kowari's ownership and control scuffle has been the number of companies, some large, who have come out of the woodwork in support of the project. And they have lawyers.

The Kowari community has been assured that corporate lawyers, including some who were involved with the writing of the Mozilla Public License version 1.1, have reviewed the situation and that Kowari is safely and in perpetuity an Open Source project. That is, Northrop Grumman Corporation may not stop a fork of the code should that become necessary. That is reassuring news for the development team and should also help to reassure Kowari's user base.

Hopefully, Northrop will take steps shortly to make good on their promises of openness so that a fork can be avoided.

Although I am not under NDA with most of these companies, I will refrain from naming names except in one case. I was contacted (via a mutual colleague) by Eben Moglen, the patriarch if not the father of Open Source law, while in Australia a couple of weeks ago. His offer to help resolve the situation added substantial comfort. Although nobody wants a fight over Kowari's intellectual property rights, it is clear that the Open Source community is on firm legal ground.

The one murky issue is Kowari's trademark. Northrop owns it and the MPL v1.1 explicitly excludes trademarks from the license. Since Northrop has not, and most probably will not, license the use of the trademark to the Open Source community, one has to assume that we cannot use it. That means that we can build projects or products with Kowari, but we can't claim it in an advertising sense. The only references would be in required legal documentation and in the package names. That issue alone may force a fork to occur, since branding is important to Open Source developers too.

I should mention that I have still not had direct dealings with Kowari since resigning in January, although I hope that can change in the near future. If Northrop plays ball, so will I. If the project forks, I will follow the fork.

Experimenting with Advertising

I have, with some trepidation, turned on Google ads for this blog. Only public service ads are showing for the moment, until Google's lawyers review and approve my application.

Mostly, I am curious to explore the world of online advertising. I occasionally provide advice to a friend's Internet advertising company and need to have more personal experience with the space. Too, I am always surprised at how many people read my drivel, so perhaps the income will outweigh the disadvantages. If that does not turn out to be the case, I will turn it off.

Tuesday, April 11, 2006

Gee, HP Really Does Suck...

I stumbled on this little gem today while browsing software engineering faculty at the University of Maryland: HEY, HP, ARE YOU LISTENING????. It was written by Assoc. Prof. James Purtilo and mirrors my experiences nicely.

I've complained in the past about HP's software (see Why Does HP Software Suck? and Part 2).

Prof. Purtilo's staff page is also the home of this nifty tidbit:

"Somewhere, something went terribly wrong"? According to Jared Diamond, that something would be the migration of human food production to agriculture. Strangely, that correlates with this image :)

Thursday, April 06, 2006

SPARQL Specifications are now W3C Candidate Recommendations

The W3C Advisory Committee has announced that the three SPARQL specifications are now W3C Candidate Recommendations. This is big news for the Semantic Web community. Congratulations to the DAWG!

The specifications are:

SPARQL Query Language for RDF
SPARQL Protocol for RDF
SPARQL Query Results XML Format