Magazine article Information Today

Preparing for the Web Service Experience: New Design Issues and a Proposed New Language Have Emerged in Response. (Focus on Publishing)

Magazine article Information Today

Preparing for the Web Service Experience: New Design Issues and a Proposed New Language Have Emerged in Response. (Focus on Publishing)

Article excerpt

Practically every other day there seems to be yet another announcement about a new alliance, product, or company that has something to do with Web services. (See page 36 of the March 2002 issue.) For the most part, these new products support backroom operations. But there are other initiatives, such as Information Builders' announcement in late May that it's going to add Web service support to Webfocus, its business intelligence tool. This type of move might be a signal to publishers that they need to consider whether Web services are in their future.

Probably the trickiest part of the Web service equation will be how to design both the information and the interface. In this column, I'll give you a hypothetical but technically doable scenario of how publishers might use the information within Web services. As for the interface itself, IBM is offering a proposed solution with its new Web Services Experience Language (WSXL).

Designing the Information

The big questions are, "Where is the metadata going to come from to make Web services work and who's going to compile it?" (Because we're going to need a lot.) But let's play with our information for a moment to see which metadata we'll need.

In order to use Web services I'll have to set up rules that tell my information aggregator which type of information I want, similar to the way that "push" services worked. Instead of merely identifying a set of topics, I could also establish the criteria used for the selection of items. But for Web services to work, this process will need to be far more graceful than the brute-force approach used by the old push channels.

Back in the olden days of push, I would have selected a topic from a menu. (And the topics tended to be broad. Really broad. For example, "information technology" or "the Internet.") And then I might get to choose from a list of sources that also tended to be broad, such as magazines or wire services. What would happen, of course, was that tons of stuff--most of it irrelevant--would be pushed my way until I could get the channel turned off. And this, as we know, is what most everyone did.

Web services are intended to be more intelligent than push. To achieve this, the system will require good and consistently used metadata. I should be able to say that I'm interested in the online publishing industry, Internet2, Web publishing, and so forth. Perhaps I could even use the metadata tags that define my Information Today columns as my filtering criteria--preferably with a "more like this" feature that I could just click and have added to my selection profile.

Next I would select the type of information I want. I might prefer that my data come from a specific group of publications, such as information technology trade magazines, and I'd like to be able to exclude titles that I'm not interested in. I could also add that I prefer to receive my information from Information Today, InfoWorld, and The New York Times if something appears there. So if an article is in one of those publications it would take precedence over the rest. And because I'm willing to pay no more than $5 per day for this service, I only want one article on the topic delivered to me unless I specifically request additional articles.

But wait, there's more. Given that the future of the Web is based on the use of many devices (PDAs, cellphones, pagers, televisions, computers, and just about anything else that can get a signal), I could have different amounts of information sent to me depending on the device. These differing amounts are called information granularity. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.