| 1.7 Standard XML
Dictionaries Through the use of XML
schema definitions, consistent dictionaries of agreed semantics can be published. Large
sites and particularly large com-panies with multiple sites and dissimilar implementations
can serve as a proving ground for the infrastructure needed for data tagging even before
addressing e-commerce issues. However, how do we get everyone with whom we have to
exchange data to agree on a common dictionary? There are several industry specific
initiatives in progress at this time. For example:
- FRXML - general accounting reporting, from AICPA
- FIXML - XML definitions of the FIX transac-tion protocol, from
FIX organization
- FinXML - data interchange for capital markets, from Integral
Corporation
- FpML - derivatives and FX processing, from JP Morgan and
PricewaterhouseCoopers
- NTM - risk management and derivatives, from Infinity
- XML/EDI - general electronic data interchange, from XML/EDI
Group
Ideally, we should just adopt the agreed standard and
retrofit all our legacy applications with the relevant standard. Two problems immediately
sur-face: the standards are still in flux, and our applications may fit the standards like
a square peg in a round hole. XML comes to the rescue with the eXtensible Stylesheet
Language Transformation (XSLT) process. Consider this disjointed set of data definitions
("dialects") between two data exchange partners. (fig 1.)
It is pretty straightforward to define an XSLT to transform
the seller's XML dialect into the buyer's XML dialect. This provides a 2-tier translation
scheme that will work nicely for two parties, as well as for in-house data exchanges
between different systems. However, when other parties come into the exchange, this
solution starts to break down. As participants increase in number, the number of 2-tier
translations required goes up much faster than the number of participants. As a practical
matter, it becomes necessary to move to a 3-tier transla-tion scheme. One party transforms
their unique dialect into the current draft of the proposed industry standard, all
exchange is done in the standard dialect, and then a second transformation is regardless
of the number of data exchange partners. These are the ultimate promises of XML: industry
wide data exchange standards, and a loosely coupled data exchange architecture that is
flexible enough to grow and to allow quick response to threats and opportunities. The rule
in IT services has always been that if you can get the data right, then the processing
logic is easy. However, data problems have been so intractable that solution providers
have focused on the logic and just allowed disjoint data definitions to propagate. Now,
XML comes forth with a practical solution to data issues which fit all significant issues.
|
