|
Next month will see the publication of the 2003 Research Software Review. For the first time, at least a dozen software manufacturers claim to support data input or output to other systems using XML – a format few had heard of three years ago. Some merely state XML, others list TabsML, a new XML-based standard for exchanging aggregated survey data and tables. Even triple-s now has an XML version -– though as yet, most implementers are still sticking to the pre-XML version.
While XML is causing a buzz among software developers and those that go to ASC conferences, it is mere geek-speak to the majority in the research community. Yet the profession could benefit from some closer scrutiny of XML, as it has the potential to solve some perennial problems, and not just the intra-agency problem of shifting collected survey data around. Clientside, it could open up the delivery of research data directly into corporate intranets and databases.
|
XML is not a standard for describing data. It is a standard for describing a number of ways to outline data, like a language where you are given the grammar but invent the vocabulary. At least with triple-s or TabsML there is agreement on both grammar and vocabulary.
Now, there is a worrying trend among manufacturers to say that these standards are too limited to convey all the metadata and content in their own interviewing or analysis systems. They respond by inventing their own. This may provide a convenient solution internally, but when this gets touted as an export route, it does not bode well.
Suppliers can argue that their proprietary XML is, at least, open for others to read. It is true that reading an XML schema is an order of magnitude more straightforward than deciphering the proprietary definitions of the old, closed systems of yesterday. But going it alone overlooks what the X in XML stands for. XML’s raison d’être is that it can be expanded.
|
It is designed to cope with improvements and even to ignore what it does not understand. The existing limits of triple-s or TabsML are not cast in stone and their custodians are keen to co-operate with other developers on extending and improving them.
XML presents a unique opportunity. It will require effort, cash and some give and take to reach the richer standards needed. Wheels tend to get re-invented because, in the short term, it is easy and gets the job done. But it is also why we are still using the digital equivalent of ox-carts and not Concorde to exchange data between research agency and research consumer. Going it alone is no longer good enough.
Tim Macer's website is at www.meaning.uk.com
|
