Categories: Top ::

Tech: Your 'net vocabulary of the day, Or, whatinole is UDDI, XML-RPC, SOAP, etc etc etc.
Sun, 14 Aug 2005

A lot of folks sling acronyms around, and programmers doubly so. A number of people recently asked me what the heck the a pile of terms mean so I thought I'd jot it down here in case anyone else is mystified.

Below, find some 'high level' definitions for the following terms:

XML-RPC is a simple specification (or 'agreement') for applications to talk to each other using relatively common and well understood technologies like XML and HTTP. (XML in turn is a way of presenting both data and data-formatting details within a single file, while HTTP is the protocol by which web browsers talk to web servers.) So XML-RPC itself just says that two applications who agree to the XML-RPC specification will use HTTP to chat with XML to each other. Easy right?

UDDI is a specification of a registry for web services, and the protocol with which an application can interact with this registry. The goal here is to list off companies and some of the services they offer (like a Yellow Pages), so that applications and users can enquire who does what and how to do so. (Yes, this has been done before.. like everything else, however maybe this'll stick? :) One uses SOAP to talk to UDDI.

SOAP itself is another inter-process communications standard where each transaction or data-payload is expressed in XML. (Try looking up 'SOAP' in Google and you'll see its yet-another-example of how a computing term out-ranks an every day thing such as a soap-brickette.) You might wonder what the difference is between SOAP and XML-RPC, and you would not be alone. My personal thought is that SOAP was partially arranged by UserLand (a Dave Winer company), and well.. thats reason enough to hold it in contempt ;) (DW is a fine lad, but he seems to spend 90% of his time flaming people. You're not really a good netizen until you've been flamed by him in fact.) Personal gripes aside, SOAP is a heavier duty XML-RPC -- its specification has grown and grown to handle more and more things, and as a result.. its fatter. Nowadays a lot of people just revert to XML-RPC and then carry the heavy data with pure HTTP GET's.

REST is not a standard really, but a high level philosophy for developing web service systems - an essay that clarifies in computing terms some of whats been going on for the last decade. ie: A website is a simple state machine, whereby requests are made as 'resource requests' (URL's, URI's), and state changes are made using HTTP POST etc. Stateless transactions and simple references are the game -- consider a blog whereby an article can be referenced by a specific URI over and over by numerous people, without effecting the state of the system. So yes, RESTy services are pretty much what we've all been doing, but nailed down into a nice set of papers you can read and digest.

WSDL is actually a standard (or is in the process of becoming one?), unlike REST which is a philosophy. WSDL is a standardized method by which web services can be described so applications know how to go about accessing them. UDDI (the registry) is built upon WSDL (the technology). Together, UDDI and WSDL and SOAP (for instance) aim towards a universal description of resources on the Internet (while trying to defuse this acronym bomb -- to use XML-RPC, or SOAP, or plain old HTTP? How to discover what services abound? Microsoft pushes SOAP...)

RSS (and RDF) are specifications ('empty specifications' as I like to put them, since they're mostly empty and devoid of sufficient meaning (requiring the app to 'know too much') IMHO) for making distribution of simple descriptions and URI's easy. An RSS-feed (in any number of formats such as RSS/RDF/Atom/etc) is common on blogs like this one, so that blogs can be syndicated to your door (browser), though RSS is increasingly used as a notification system for distributing things like PODCasts and other materials..

[ Category: / technology / vocabulary ] [link] [Comments]