By Harry Fuecks

Walking the walk vs. talking the talk

By Harry Fuecks

About two years ago you may remember we were in the throes of the SOAP revolution. “Web services everywhere!” was the cry and have to confess I’m one of those guilty of having gone for it, for a while. There were going to be these giant UDDI servers that would aggregate everyone’s web services and the Internet would never look the same again…

Two years later, seems that old RSS thing, first defined by Netscape back in ’99, is “broken” (more analysis here). Infoworld also report struggling with RSS. The basic problem is the impact of automated web requests. RSS readers and aggregators are not a far cry away from that haXor Perl script that’s trying to brute-force your website’s login mechanism right now.

Personally don’t think RSS is “broken”. Its just Really Simple Syndication turns out to be not-so-simple when the number of clients reading it > lots and organisations you’d expect to have the answers turn out to be just as fallible as the rest of us.

So where does SOAP fit into this? Perhaps Bug 257304 on changes to Firefox’s extension update mechanism is an indication;

This web services business isn’t going to scale, or have a hope of being shaken out thoroughly for 1.0, there are problems with SSL etc. Thus we are falling back to RDF to provide updates from

As always reality is the dividing line between those talking the talk and those walking the walk. Put another way, wanna sell technology?

avoid all technical people and ensure they only sell to sales/upper management/marketoids, because they’re the group that’s too detached from the [insert word here] to know any better.

Anyway, excuse my Friday cynicism. Probably the result of a certain “small person” waking up at 3am this morning…

  • Captain Proton

    Web services are a form of distributed computing. But the first law of distributed computing is: don’t distribute! :)

  • the first law of distributed computing is: don’t distribute! :)

    Found the link to Martin Fowler talking about this: Errant Architectures. Interestingly Martin’s ideas seem to have warped into the notion of Service Oriented Architecture and coarse grained remote APIS (but don’t forget SOAP used to be called Simple Object Access Protocol).

    Think the main point that’s getting missed is distributed computing means dealing with alot of complexity – the sort of stuff you can’t nicely “abstract out”. Simon’s just blogmarked a couple of reads on how to deal with RSS bandwidth consumption. Meanwhile ICE seems to be flying in the face of reason and rather than regarding distributing objects as fundamentally insane, is trying to fix specific things they feel CORBA got wrong.

    Guess the bottom line is this stuff can be made to scale but it takes time and focus on fine technical detail. Or watch out for things claiming to be simple…

  • arborint

    Sounds more like a healthy dose of practicality for next Monday than Friday cynicism. I think what people like ICE and the previous discussion here about RPC point to is — they aren’t building blocks if you can’t really build on top of them.

Get the latest in Front-end, once a week, for free.