Featured Post


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

Saturday, May 24, 2003

ssh into the matrix

I did the whole Matrix Reloaded thing last night. I could have done with less of the silly fighting scenes and more of the amazing car/bike chasing, but hey, when I was a kid we watched Bruce Lee.
The DVD would merit some freeze frame analysis. I spotted SSH at one stage being used to log into a power grid.Somewhat disappointed to see that "the source" is a freaking mainframe. Surely that should be a billion node Linux cluster?

Friday, May 23, 2003

Jon Udell

    "The inexorable logic of Web services sets aside APIs in favor of protocols. XML messages flowing through the pipes enact those protocols."

Amen to that. You know the hardest thing for a lot of programming folk to grasp about Web Services is? Its not about APIs. Forget APIs. Forget RPC. Forget SOAP. If you are thinking APIs you are going in the wrong direction. Thing *documents*. Think documents flowing. Think reliable messaging. Think XML transformations. Think URIs. Think HTTP POST and GET. Think design by contract (where XML schemata are the contracts). Think message choreography. Think two-pin in/out interfaces. Think plug-replaceable components that communicate solely via XML messages. Think loose coupling. See?

SOAs to zap the integration problem in a couple of years - too optimistic but on the right track

[Via Loosely Coupled]. A Techweb story about how SOAs will dramatically reduce integration costs over the course of a couple of years.
I'm a true believer in SOA's but that timescale is way too optimistic. Even in a perfect world where SOAs are implemented as they should be architecturally, this is optimisic.
Evidence is already mounting of a blind alley some implementors are going down in implementing SOAs. Hint - if your nascent SOA architecture diagram consists of blocks, layers, RPCs and APIs - its not an SOA (to my definition!).
If your architecture diagram has flow (often notated by circles, ellipses and pipes), synch-over-asynch MOM, XML transformations and input/output XML documents with associated schemas (service contracts) then you are on the right track.

Thursday, May 22, 2003

If an XML document has a doctype - is it an XML document?

An intriguing XML validation proxy but see this:

    "XML containing DOCTYPE declarations are blocked."

Now don't get me wrong. I'm a fan of eliding DOCTYPE declarations from XML instances in favour of specifying a schema (DTD, RNG or (blech)XSD) out of band. I've been advocated this for years.
But also years ago I was part of an effort to create a profile of XML that stripped out things like DOCTYPE (Common XML. We got eviscirated for using the word "XML" and for diluting the brand name.
As if "XML" has a clear meaning. As if "XML tools" is defined in any real sense. It isn't.
XML is is lune on the noosphere not a point on the noosphere as we would perhaps like to pretend.

The trouble with RDF

Tim Bray on RDF

    "At this point, the RDF evangelists pipe up and say “Well, Ordinary People (tm) don't have to look at the source, there will be tools to sort all that out.” Sorry, I just don't believe that. If, in 1994, you'd needed DreamWeaver or equivalent to write for the Web, there wouldn't be a Web today."

Amen. I *hate* the tools-will-do-it argument - not only for RDF but for all software. The argument is a breeding ground for complexity. It lets people off the hook (you don't need to take the time to be brief, you don't need to simplify.). It lets vendors off the hook ("X is complicated but we will flog you funky tools to manage the complexity").
A lot of XML technologies these days are big bags of complexity. A lot of it comes from this idea that XML is not meant to be read by humans. As soon as you say that, you are on the slippery slope in my opinion. Tim (Bray's) RPV notation shows that the RDF model does not have to be ugly when serialized. Same goes for Tim (Berners Lee's) N3 notation.
Syntax matters - that is all ye know on Earth and all ye need to know.

Wednesday, May 21, 2003

RDF and semantic shadows again

Semantic shadows is a mechanism I suggested in an ITWorld piece RDF and other monkey wrenches to allow RDF representation of semantics to exist as "shadow files" of the less abstract XML representation. I've just come accross another interesting semantic shadow mechanism on Sjoerd Visscher's site: xr

py2exe - I've use it a lot but it still seems like magic

Combined with a good install program, py2exe compiled wxPython apps are indistinguishable from polished C++, MFC apps. I have a good idea how py2exe works under the hood but it still feels like magic. Whats that saying about 'sufficiently advanced technology'? :-)

Tuesday, May 20, 2003

Monday, May 19, 2003

demotelco again

I've had a couple of e-mails this morning about demotelco. Its a test/demonstration site as opposed to a production site. If you are interested in using it long term (or seeing your telco supply it) please contact Newbay directly.
BTW, Newbay tell me that RSS feeds will be available on the site very soon now.

To every phone, its blog

If you have a mobile phone, and can send MMS or SMS messages then you can have a blog on demotelco. Its run by mobile blogging specialists Newbay, based in Dublin. I expect it will not be long before every mobile phone contract comes with an optional blog account. "Here is your new phone Sir. Would you like a blog with that?".