FpML Issues Tracker

Summary

FpML,

I have quicky glanced at the spec for FpML 2.0. I have skipped the sections on the actual definition (which I am sure mean lots more to others), but looking at the schemes and data type sections. Here are some observations that might be useful:

* Section 9.1.1 - there seems to be an inconsistency that it says at least four digits for the year, then says years in the range 0001-9999. Could be fixed either way, but depends on whether you want to create a 10th millenium bug.

* Section 9.2.1 - for the tradeId fields there is a tradeIdScheme, but this has no values specified. This makes sense since the Ids are self administered, but it might be worth suggesting that the tradeIdScheme should use something that makes the tradeIdScheme/TradeId combination globally unique, e.g. using their domain name in some part. This avoids the case where tradeIdScheme is left blank, and a number used in the tradeId field (which will quickly fail the uniqueness test).

That is all I notice at this glance...

Cheers,

Nic.

P.S. It is great to see the Architecture Recommendations being followed - especially for those of us who worked (and fought) to try and get it right.