There should be an embargo on the use of the word 'consumer' when describing the recipient of a web service. Being on the receiving end of a web service is sometimes a very complex proposition, one that deserves the kind of consideration not normally accorded to 'consumers' by the IT industry. I was alerted to this verbal trap when reading Doug Kaye's riposte last week to my comments on the Google API. "I'm not looking at the consumer's side of such a very simple web service," said Doug, and his words have been bothering me ever since. I agree that publishing the Google service is a major undertaking, whereas most subscriptions to it thus far have been very minor affairs. But in the web services world, we are all publishers and subscribers, and I suspect that the scale of both activities relates more to the size of the individual participant than to the nature of the web service in question. In other words, enterprise subscription to web services is likely to be as complex as enterprise publishing of web services, and both roles will demand advanced skills (so that's two O'Reilly books to look forward to).
posted by Phil Wainewright 12:12 PM (GMT) | comments | link
Patterns of web services adoption
Research by the Stencil Group has found that enterprise adoption of web services is exhibiting "several systematic patterns of adoption," according to this month's newsletter from the group. Several points stand out:
Because business processes often cut across several different conventional applications in an enterprise, workers are often spending time performing 'manual integration' from one application to another in the course of their daily routine. Web services provide a way of automating these manual stopgaps.
Web services put some new spins on the concept of software re-use. One of these novelties is the notion of re-use by syndication, via the web, to partners or customers. Actually, come to think of it, software reuse is implicit in the very concept of on-demand web services.
According to one e-business manager, using web services broadens your horizons, so you think about automating "interactions, not just transactions."
Technologists often find it inexplicable when businesses refuse to adopt new innovations, even when they have the potential to deliver significant benefits. But technologists often underestimate how much hassle and disruption it takes to implement new technologies. "The cost of new technology is measured in time and frustration, not just dollars," explains this article by Michael Schrage in the Technology Review. "Genuine innovation almost always creates a disjunction between the price thatís paid to acquire it and the actual costs of implementing it." The article which I found thanks to a recommendation in Amy Wohl's newsletter this week brings to mind an article I wrote early last year after a meeting with Paul Grayson, the founder of web service company Alibre (and previously founder of Micrografx). Alibre uses web services to do something that could have been done many years ago using magnetic tape drives or dial-up modems. But before Grayson came along, no-one had taken the trouble to implement the technology in a way that made sense to buyers in its target market, so none of those other technologies ever made any headway.
posted by Phil Wainewright 3:20 AM (GMT) | comments | link
Wednesday, April 24, 2002
Microsoft denounces 'ungrounded' web services hype
So Microsoft believes some vendors are guilty of over-hyping web services this according to The Register, in a story syndicated from ComputerWire, which reports comments originally quoted in a report on Silicon.com (the audit trail is worth mentioning, just in case Microsoft decides to invoke plausible deniability). Charles Fitzgerald, general manager of Microsoft's .Net platform strategies, is said to be concerned that "there is a lot of bandwagon jumping going on, from people that are in no way grounded in this. Our concern is that those who are patently just along for the ride are setting some unreasonable expectations." As one might expect from Fitzgerald, he is said to have cited Sun as a prime example.
Presumably his choice of phrasing is intended to exempt those on the bandwagon who are "grounded" there rather than "just along for the ride." Otherwise his comments might be construed as critical of these words spoken by Bill Gates at the unveiling of Microsoft .Net in June 2000:
"What we're talking about today is a new platform. This will affect every piece of application code that gets written. This will redefine the user interface, what you see on the screen and how you interact, as much as the transition from DOS to Windows did."
Or perhaps when Fitzgerald talks of fears that people will get bored of web services before they are delivered, what he actually means is there's a danger that Bill himself is losing interest?
posted by Phil Wainewright 8:02 AM (GMT) | comments | link
Tuesday, April 23, 2002
Google API awakens the hive
Every so often in technology, something comes along that opens up new angles on the way we think about things. That's what has made the Google API such a landmark event, irrespective of what the API itself does or doesn't do (already, several noted XML developers aren't happy with the interface).
What counts here is not the granular implementation of the technology (one way or another, the engineers will get working in the end). What counts is the concept that it embodies and the possibilities that open up as a result. When the API came out, it acted as a trigger that brought home to all of us a series of ideas that have been bubbling around over the past couple of months:
Google ranking turns the Web from a mass of unsorted information into the beginnings of a collective knowledge base
Personal publishing via weblogs creates a feedback mechanism that can fine-tune (or of course distort) the ranking in real-time
An API creates the potential to vastly enhance the speed, accuracy and responsiveness of that emerging knowledge system by engaging the participants more closely.
All of those principles apply to many other existing or imaginable Web-based systems and groupings.
"Imagine for a moment that Google ranks not only content but also online Web services according to choices made by a global cross-section of trusted, knowledgeable authorities. Available on demand and in any configuration, those services include ... an entire portfolio of business functions, constantly updated based on feedback from a worldwide user base."
It's not just me. Gartner analyst Ben Pring is equally doubtful whether vendors are quite as keen on web services as they profess to be. The invaluable searchWebServices site recently interviewed him about his findings in the Gartner report, Lessons Web services companies must learn from ASPs. Back in 1999, Ben was one of the first analysts at Gartner to spot the potential significance of hosted, componentized software for software developers and integrators, and is well qualified to speak out on the topic. Here's a provocative taster of his views now: "Pring said in a best-case scenario for integrators, or a worst-case scenario for Web services customers, secure and public Web services would never be widely implemented."
I'd been holding off on posting this comment until I could link to my own article in a similar vein, which was published offline last month, and is now online at my freshly revamped personal web site. Ben is an honorable exception to the jack-of-all-trades analysts that I savage in the second paragraph of the article particularly since his comments appear to endorse (albeit in a more measured tone) my own very forthright views on the topic:
"Once businesses find out that they don't have to pay integrators to construct massive, inflexible software edifices to automate their businesses, the market for today's monolithic packaged software applications will crumble faster than a sandcastle when the tide comes in."
The emergence of on-demand web services is leading people to start talking about the concept of Service-Oriented Architecture (SOA) when designing application systems. "An SOA ... is an architecture made up of components and interconnections that stress interoperability and location transparency," says Michael Stevens in this lucid account of the key principles of SOA and how they relate to web services. He lists four core characteristics of services in an SOA:
A service is behavior that is provided by a component for use by any other component based only on the interface contract