Internet Alchemy

Sam Ruby has written an
essay on SOAP + REST in which he proposes four web service world-views: everything is a resource; everything is a get; everything is a message; everything is a procedure. I think he’s missing the point. My understanding of the key difference between the REST and SOAP positions is that resources should be addressable. In REST this means GETting a specific URI to retrieve the representation of, say, the weather forecast for Rome. In SOAP the URI is used as a command target, the command being to retrieve the weather for Rome. The same URI can be resused for retrieving the weather for Moscow and, perhaps, last week’s temperature chart for Athens.

Sam also makes the smart, analogy of SOAP services being the dark matter of the Internet since the lack of a GET interface makes the services inaccessible the majority of Internet clients. His key recommendation here is to maximise the ’surface area’ of the service by supporting other methods such as GET

Worth reading as another viewpoint on the whole REST vs SOAP debate.

I love the way you can use templates in the specifications of the archive names in Moveable Type. I’ve used this feature along with a custom plugin I developed to produce friendly names for all the archived entries. For example, this entry will have an archive file name of moveableTypeCamelCase.html which is a lot nicer than the default 0000321.html.

Continue Reading »

This looks like an interesting candidate recommendation from the W3C. It’s a way of specifying pattern matches for CSS media types so that, for example, you could specify that a particular stylesheet should be used when the screen width is less than 100 pixels. However, I have to question whether anyone is currently using Media Types on a production website anyway

There may be some temporary glitches as I complete the move to Moveable Type. As expected the archives are proving the most time consuming so please bear with me until I get it all sorted out. Thanks.

Moveable Type allows you to create multiple templates for archives and content pages. I’ve used this capability to produce RSS 1.0 and RSS 0.92 feed for each of the categories I’m using. I’ve also created a template to build an OCS file listing all the various syndication channels that I’ve now got available.

Here’s the template I’ve been using:

Continue Reading »

I’ve spent the last few days converting this site to run under Moveable Type which is the coolest piece of CGI I’ve seen for a long long time. As Aaron points out, MT is one of the new generation of retro tools that serve up static pages built from a database. The effects of this are fairly subtle and they’re not raw performance related – it’s a scheme that actually works with the underlying web server, using it for what it was designed for: serving pages as efficiently as possible. You immediately get all the benefit if the thousands of man-months spent developing the web server: caching, etags, proper Last-modif modified headers etc.

Weblog Kitchen is a new wiki for exploring weblogs, community editing and various
hypertext issued. Quite sparse at the moment but give it some time.

Thanks to everyone who voted on the future of RDF in OCS. The
results were as follows:

• No changes – 0 votes (0%)
• Fix OCS to be compliant RDF, keep general structure the same – 2 votes (11%)
• Fix OCS to be compliant RDF and rewrite in form similar to RSS – 10 votes (55%)
• Remove RDF, keep namespaces – 5 votes (27%)
• Remove RDF, no namespaces – 1 vote (5%)

So the consensus is clearly that RDF should stay and that the syntax of OCS should be expressed in a similar manner to RSS 1.0. I’ll be publishing a draft of the new spec soon and hope that everyone can give feedback on its evolution.