Imported from previous forum
First, let me say that after reviewing the FIXML Schema spec, I’m quite impressed. I think it represents some admirable work done by the contributors, and should help move FIXML forward considerably.
In reviewing the spec, I was curious about the consistency of a couple of the datatypes…
MonthYear - Why is this not formatted consistently with the other FIXML Schema dates and times? I agree that choosing XML formatting over FIX-Classic formatting is the way to go, but why stop at MonthYear? I think that a MonthYear format that includes the hyphens (eg- 2004-02-14, 2004-02 and 2004-02-w2) would aid consistency, readability, and understandablility, and makes it easier to build parsers and formatters.
Currency, Country & char - Was there any reason why ‘char’ has a formal length restriction, but Currency and Country do not? Presumably, a parser on ‘char’ is still going to have to pick out the first char of the string, so the restriction is in there for the validation stage only. If so, do we not want to provide the same kind of validation for other fixed-length fields?
…Bill Hollings
[ original email was from Jim Northey - jnorthey@jandj.com ]
Thanks for your comments on the schema.
Regarding month-year - we felt it was more exchange symbology as opposed to specific datatype handling and it was one we would not have to introduce the "-" into. We chose to follow XML schema datatypes for other date / time fields to make it easier to intgrate FIXML messages into off-the-shelf XML technology. Regarding MonthYear - I suppose if it were felt strongly that it would be easier to use with XML tools that would be listened to.
Regarding char - we wanted it typed so there was no information loss when moving to/from FIX.
Regarding country/currency - we left these incomplete as we are still working with ISO and FpML to determine best schema representations for these types. So I expect we will see some enhancement in the future - possibly after the FIXML-FpML Collaboration working group completes their work. We probably could have added a length restriction at this point - maybe we should.
> First, let me say that after reviewing the FIXML Schema spec, I’m quite impressed. I think it represents some admirable work done by the contributors, and should help move FIXML forward considerably.
>
> In reviewing the spec, I was curious about the consistency of a couple of the datatypes…
>
> MonthYear - Why is this not formatted consistently with the other FIXML Schema dates and times? I agree that choosing XML formatting over FIX-Classic formatting is the way to go, but why stop at MonthYear? I think that a MonthYear format that includes the hyphens (eg- 2004-02-14, 2004-02 and 2004-02-w2) would aid consistency, readability, and understandablility, and makes it easier to build parsers and formatters.
>
> Currency, Country & char - Was there any reason why ‘char’ has a formal length restriction, but Currency and Country do not? Presumably, a parser on ‘char’ is still going to have to pick out the first char of the string, so the restriction is in there for the validation stage only. If so, do we not want to provide the same kind of validation for other fixed-length fields?
>
> …Bill Hollings
>