Wednesday, February 28, 2007

Intense discussion about performance of Java SOAP stacks

Stefan Tilkov posted an article about the discussion about Java SOAP stack performance on InfoQ. It seems that some developers of the three major Java SOAP stack projects, Axis2, Xfire, and JAX-WS, are really concerned about the topic. Although part of it was not so comfortable, the discussion can still help to improve the development of those projects. I think Steve did tell the essentials of SOAP performance that developers need to care about.

I have tried to learn all the three frameworks. To me, the learning curve and tools are also important for an open-source project to be accepted by the developers besides code reliability.

Tuesday, February 27, 2007

Switching to Google Reader

Not surprisingly, I switched to Google Reader as the tool to read blog feeds. I used to use the Sage plugin for firefox. The features of Sage are enough for my requirements. However, it is difficult to synchronize my readings (both subscriptions and reading progress) on various machines. From time to time, I need to remotely connect to the machine at office to synchronize. Google Reader is the solution if you need to read feeds from different machines. To switch from Sage to Google Reader is very easy, just export the OPML file from sage and import it in Google Reader. And then you can start to enjoy the new reader.

Wednesday, February 7, 2007

Interesting points from Grady Booch

I guess everyone using UML know Booch. Several years ago, he joined IBM when the Rational was sold. Gervas Douglas just posted some tidbits from his discussion with Booch.
Functional programming languages (like LISP, Scheme and SML) failed largely because they made it very easy to do very difficult things, but it was too hard to do the easy things.

Recently, I am reviewing and writing about the continuation and its applications to service orchestration or work flow implementations. The continuation is natively supported by LISP, Scheme and SML, but not by Java. I have no experience on those old languages, but I suspect that they must deal with the highly abstracted aspect of continuations.
The next big challenge in software architecture is concurrency. Raw clock speed has just about reached its physical limit. Chip companies are now putting multiple copies of the same CPU onto a single chip. The result is that applications can no longer just be run faster. They have to be run in parallel in some way.

I think concurrency and asynchronism are the major power and challenge of the SOA world.