>In this context, I guess what I'm calling 'the Web' is the things that
>conform to the Web Architecture (sorry, I haven't thought about this
>very hard I must admit - but that seems like a reasonable response):
>
>http://www.w3.org/TR/webarch/
>
>Does iTunes conform to the Web Architecture?  Off the top of my head I
>can't answer that.  I'm doubtful that it does but I'd be willing to be
>told otherwise.

So, if you look for iTunes U content on Google, you come up with links similar to this (Carnegie Mellon's home page):

itms://deimos3.apple.com/WebObjects/Core.woa/Browse/cmu.edu

Now, the itms protocol is a bit foreign to those used to http, but if you do change it from itms:// to http:// and load it in your browser, then you get a html page back that says it's connecting you (and then launches iTunes - by loading the itms:// URI above!).

Interesting in itself, but more so when you look at some notes regarding the iTMS protocol:

http://hcsoftware.sourceforge.net/jason-rohrer/itms4all/

What does iTunes do when it's loading something identified by the itms protocol? Sending a normal HTTP request to port 80 of the server. A regular GET of the path, including specifying the HTTP/1.1 protocol.

Although iTunes itself doesn't receive/render that html page you saw in the browser. The server is doing content negotiation based on the User-Agent - if you install the 'Modify Headers' extension into Firefox then add a filter on User-Agent and get it to send:

iTunes/6.0 (Macintosh; U; PPC Mac OS X 10.5)

suddenly you get back an XML document, which describes the interface page you see in iTunes.

iTunes doesn't use HTML, but there is nothing that says a URI has to return HTML - and iTunes itself is clearly building on Web Architecture.

Whether it's a bad thing that it's not more immediately accessible is a matter of opinion. But it's entirely feasible to build a third-party client that can at least get as far as browsing the content within iTunes.

G