Featured Post

Linkedin

 These days, I mostly post my tech musings on Linkedin.  https://www.linkedin.com/in/seanmcgrath/

Saturday, January 04, 2003

Nice snake. Don't squeeze too hard now, theres a good boy. Nice snake...

A new Python/Java book.. Haven't got it yet. Great cover.

Some Web Services Predictions for 2003

I write a Q&A column for searchwebservices.com about XML. Some time ago they asked for some predictions about web services for 2003.
Here is where they ended up.

Friday, January 03, 2003

Its all in the wetware folks

A piece I wrote for XML Journal some time ago came my way again today. Basically, interoperability is tractable in terms of hardware and software (Hard but tractable. As the wise man once said "strive to make the impossible, very difficult.".) The intractable part is all to do with people. Industry standard schemas in the XML world show up the extent of the people problems very nicely.

It seems the whole idea of an industry standard schema may be simple, elegant and just plain wrong. :-(

An experiment in collaborative blogfiction

I've started a blog story. The starting date for the story is January 3 2003 but it is being written from 2030 by the narrator.

If you would like to contribute, please contact me (contact information in my blogroll). Its hosted on blogger which allows multiple authors so I can sign people up with contributor access easily. Also, I'll take contributions by e-mail (please put "BLOGSTORY:" into the subject line.

I'd like to have at least one entry a day added to the story. Lets see where it ends up!

xml-dev timewarp

xml-dev is still in 2002 today. Thank goodness for meaningful URLs from which we can deduce
http://lists.xml.org/archives/xml-dev/200301/maillist.html which does indeed work.

Thursday, January 02, 2003

Wednesday, January 01, 2003

An important link for Irish visitors

Karlin Lillington's 'In defense of data privacy'.

Jython status report by Finn Bock

Jython status report from last years (2002) Europython conference. Roll on Jython 2.2. I firmly believe that high quality interpreted languages such as Jython represent the best hope for Java the platform.

Note to all Java-only programmers out there. Quit heaping abstraction on top of abstraction to work around the fact that your programming language of choice in caught in a

  • static typing is necessary
  • you must compile everything otherwise it will be fully of um, mistakes and um, slow
  • you must know everything at compile time for a program to "work"
  • you mist paramaterise throught declarative constructs exclusively
  • scripting is for kiddies, right?
  • one language fits all
  • xml is just objects in disguise

time warp and get with the program . Support your local "on the Java VM" programming language. There are plenty to choose from! Jython would be a good choice in my opinion.

Tuesday, December 31, 2002

3 + 3 = 6 but 4+4 = 9

A piece I wrote about Recombinant Growth has appeared in the ITWorld archives.

Like it or not application integration is still in N squared mode. What I mean is that the sad lack of anything approaching a workable concept of "component" has resulted in a huge amount of coding effort to bolt things together. The more things you need to bolt together, the more point-to-point integration glue you need (even if you have a middleware layer precisely to obviate it). The trouble with this (apart from the obvious ones of time and money) is that the relationship is not linear. The amount of glue you need grows as the square of the "components" being glued. The formula is n*(n-1)/2 interconnections, hence the funny title to this post. This equation is such that it only bites once you get past three components. Point to point component integrations seem to make sense for two components, then three...but as you add more the bottomless pit of integration goo opens up before you...

In order to get what Hal Varian calls Recombinant Growth, we need to make the integration model O(n) rather than O(square(n)). To do that, we need a workable definition of a component - otherwise Web Services will amount to one big damp squib. Unfortunately, concensus on what it is to be a component in software engineering seems as far away as ever.

Monday, December 30, 2002

Manufactured serendipity

The Gods of Serendipity did not shine on me two weeks ago when I wrote an article entitled Close encounters of the electronic kind for ITWorld. I wish I had stumbled accross this first.