- From: Peter Furniss <peter.furniss@choreology.com>
- Date: Mon, 6 Jan 2003 21:46:41 -0000
- To: "Mark Baker" <distobj@acm.org>, "Miles Sabin" <miles@milessabin.com>
- Cc: <www-ws-arch@w3.org>
> On Mon, Jan 06, 2003 at 04:54:21PM +0000, Miles Sabin wrote: > > And the RESTless version could work just as well if we substituted > > "9ajp23q9rj89aweruwer" for "getLastSharePriceOfIBM". What allows this, > > in *both* cases is the _prior_ coordination between the client which > > has, > > Wrong. > > I think it's funny (in an unfortunate way) that this benefit is so > easily taken for granted. It's called a *coordination* language for a > reason, ya know. 8-/ > > http://www.markbaker.ca/9ajp23q9rj89aweruwer > > Quick, before you type that into a browser window, tell me everything > you and your browser know about it and what I'm trying to communicate to > you by putting it in this email message. Can't resist responding to this :-) true, all my browser knows is that it's an http url, so it will make an HTTP GET and send it to www.markbaker.ca (resolving to an IP address using DNS, sending over TCP to port 80), putting the other parts of the url in the appropriate fields of the GET message. Then wait for whatever comes back, and see if it can present the information to me in some way. but I know, or suspect, that this will be a document written by Mark Baker which explains or expands on how all you need is HTTP and URIs and that it is intended to reinforce his arguments on the virtues of RESTfulness. Probably explaining how the HTTP verbs and appropriate returned URIs allow it to be used for coordination (in a sense understood by said Mark Baker, but probably not quite as I understand it). The document will be in English, with perhaps a few diagrams. It is very likely to be an html document, but will definitely be something that my (and at least any recent) browser will be able to handle without additional downloads or plugins. But *how* the browser and I know this is significant. The browser knows because the URI and the fields up to the .ca follow the patterns defined in the relevant RFCs, and the browser has been written in the light of those RFCs and to implement them. But I know my bit because I'm an instance of a self-programming, pattern-recognising inference engine with (probably) a billion years of software development behind it. Which is also how I know which link to click on a webpage that offers links to "proceed to checkout", "privacy policy" and "delete item" etc. Of course I don't have to interpret or parse the URL (nor does the browser, beyond the hostname), but there is a communication (using English and some other patterns, like an icon of a supermarket trolley) that means I know which is which. For me, it is the equivalent of that last aspect that I expect will still be unclear to me if I follow the URI in Mark's message. Peter ps1: of course, it's possible I've guessed quite wrong, Mark has got fed up with beating his head against the wall, and the URI leads a jpeg symbolising his irritation :-) ps2: just tried it, and I was wrong, though it isn't a jpeg. Obviously i'm not a very good pattern-recognising inference engine.
Received on Monday, 6 January 2003 16:50:12 UTC