FPML Views Faq
FPML Views Faq
Introduction
Since release 5.0 the FpML schema has supported the generation of multiple ‘view’ schemas from a
common ‘master’ schema. The ‘views’ mechanism was created to address the problem of varying
the information content of schema to reflect the requirements of the business processing being
applied to it.
The first business process that FpML addressed was trade confirmation and as a result the models
for the trade structure and derivative products are strictly defined to ensure that all the data values
that would be mandatory on a paper confirmation document are defined as mandatory within the
XML schema.
When the working groups turned their attention to other business processes, such as pre-trade
negotiation and reporting they discovered that confirmation style designs where too strict and
needed to be adjusted to allow an alternative definition in which some normally mandatory
information could be omitted (e.g. business day adjustments, etc.) or replaced entirely.
To have a strict content model for each business process ‘view’ of the data we need multiple
schemas, each with its own specific grammar, but maintaining multiple schemas to ensure that an
important change to one is reflected in the others would be time consuming and error prone if done
manually.
The solution is to derive all the views from a single ‘master’ schema using information that describes
how to adjust schema components to match the requirements of the view.
Every time FpML is extended to cover a new business task the Business Process working group
considers whether the task can be accommodated within an existing view or whether a new view is
required.
Applications that process messages from more that one FpML view MUST examine both the name of
the FpML root element and its namespace URI to determine what the message was intended to
communicate.
In particular FpML does not contain any management information needed to correct book or process
trades within a bank (e.g. no books or accounting information, no system names, no processing
instructions, etc.).
For most front office to middle/back office STP data transfers an extended version of the FpML
confirmation view is probably the best match as it contains complete and detailed trade economics.
Some risk and finance data feeds may be better served by the reporting view.
Extensions to FpML (to add additional MIS type information) should be done using the techniques
described in the Architecture specification rather than by directly editing the schema.
The AWG is considering the possibility of developing a tool to allow custom views to be generated
from master schema to help make this task easier.
At present the FpML master schema does not match this vision. The initial FpML 5 master schema
was created from the confirmation model in the 4.x version and annotations have been added to
allow the generation of less restrictive views, like reporting, from it.
The AWG would like to correct the master schema but it has to be done without accidentally
introducing any unintended changes. Manual editing would be too error prone so we need to
develop some automated tools to algorithmically correct schema and compare the results with the
current views to ensure accuracy.