Re: [Quickfixj-users] repeating group problem

[Steve B. <st...@te...>]

Again, it's not a significant issue. I see validation and metadata
management as two separate responsibilities and representing two types
of functionality that can be used independently. For example, some
of the QF tools use the metadata but aren't concerned about
validation. It's possible someone would want to use the data dictionary
for message parsing but disable all or most validations by providing
a custom validator. For extended validators, the "something else" might
be custom validation logic such as a rule-based validator or some other
advanced validation mechanism. There might be a desire to store metadata
in a different format than XML, such as a database or a Spring Framework
configuration (for Java, of course). The tight coupling between the
validation and the data dictionary implementations make this a bit more
challenging.

Steve

> I'm not sure I necessarily agree about pulling validate out of the
> data dictionary.  I suppose if there is a use case where you want to
> validate against the data dictionary and something else.  If you are
> just validating against a data dictionary however, the validate
> method seems to be in an appropriate location to me.
> 
> --oren