Thursday, May 27, 2010

NewsWeek.com is using RDFa - let's reward them with our business

I like to "let my wallet do the talking" when it comes to rewarding companies with my business when I feel that they deserve it. Take a look at the HTML+RDFa source for the Newsweek web site. Excellent use of RDFa for semantic web markup.

Tuesday, May 25, 2010

Excellent: "Points of View" is a collection of previously unpublished essays written to celebrate Alan Kay's 70th birthday

The print books are sold out but there is a free PDF download available.

To bad for me: I have never met Alan Kay. I did get early access to Smalltalk when a Xerox representative gave me a trial version for my Xerox 1108 Lisp Machine in 1982. I also give a tip of the hat to Alan Kay and the many other Squeak developers every time I use Squeak (or more recently, Pharo).

I am not done reading through "Points of View" but so far it is inspirational (like Alan Kay!). Good job by Kimberly Rose and Ian Piumarta putting this together.

I have been working for CompassLabs

I can't talk about my work, but here is their web site.

Here is a video of our CEO's TechCrunch Disruptive Technology talk yesterday. (The judges liked the presentation!)

Sunday, May 23, 2010

Really liking Ubuntu 10.4 with my i5 laptop

I wrote earlier this year about buying an i5 based Toshiba laptop. I have been struggling a bit getting things just right using Linux, but I am there now. Installing the final non-beta version of Ubuntu 10.4 and working a little to automate wireless connections made the difference.

I still love my MacBook with OS X and use it for casual use but an i5 laptop with 4 gigs of RAM and a half a terrabyte of disk for 40% of a comparable MacBook Pro is a story that is difficult to ignore. I admit to being frugal, and lightening fast inexpensive hardware running Ubuntu makes me think that Apple should drop their prices (easy for me to say: I sold my Apple stock when I made about 4x my investment).

I have Windows 7 installed also on my i5 laptop and I am happy to have Windows on some rare occasions but dealing with very long boot up times, the overhead of anti-virus software, a poor command shell even with bash installed, and the difficulty of using a lot of development software dooms Windows 7 to a nice to have toy, but I might someday trash it to get back my 250 gigs Windows partition.

Tuesday, May 11, 2010

Very good movie by Kate Ray: "Web 3.0"

This 15 minute film is a well done set of interviews with people who have different views on the Semantic Web: Ontology's necessary, or not; possible to implement or not.

I have made my own views clear enough: I don't expect that many people will tag web pages with RDFa, etc. but that automatically generated content will be likely to contain semantic markup. Getting the Semantic Web (Web 3.0) infrastructure done right is a tough problem, and I have problems disagreeing with people who have different points of view on implementation strategy because I think that both automatic semantic markup and finding relationships between data, and more formal Ontology driven strategies will be useful.

I spend much effort in the little technical details of the Semantic Web and automated data/text mining. I am not what you might call "big idea" person, rather I have always enjoyed the dirty little details of making things work. For me, this movie provides a welcome high level perspective.

Saturday, May 08, 2010

I created a separate blog just for Clojure stuff

Recently I created a blog specific to Ruby. Today I started a new blog that is specific to Clojure.

ClojurePla.net

I created a Clojure NLP project on github

You can get a copy here. This is some of the code from my book "Practical Semantic Web Programming" with a Clojure wrapper. Natural Language Processing code for classification, entity extraction, and automatic summarization.

Saturday, May 01, 2010

Clojure: using a local Maven repository in Leiningen based projects

I had what must be a common problem so I am documenting the solution here using this reference. I have an IntelliJ mixed project with both Java and Clojure code. I wanted to reuse this project in a small Clojure/Compojure web application that uses Leiningen. I added a new target jar_to_local_mvn to the Makefile in my IntelliJ project directory that takes a JAR file in the IntelliJ project directory and adds it to my local Maven repository (the Makefile target for the dependency target jar is not shown):
jar_to_local_mvn: jar
mvn install:install-file -Dfile=knowledgebooks-0.2.0.jar -DgroupId=self \
-DartifactId=knowledgebooks -Dversion=0.2.0 \
-Dpackaging=jar -DgeneratePom=true
Then in my Clojure/Compojure web application project directory I added a local dependency to my project.clj file:
     [self/knowledgebooks "0.2.0"]
Now, when I get all project dependencies it also copies my JAR file in addition to dependencies copied from Clojars (JARs retrieved from Clojars are not shown):
$ lein deps
[copy] Copying 1 file to /Users/mark/Coding/Clojure/KBSportal/lib
The code in the IntelliJ project is very stable and is unlikely to change frequently so having to manually re-run the Makefile is not a nuisance.