[osis-core] I don't think you want to use what you sent as 1.0 in any case.

[Todd Tillinghast]

I like the idea of keeping x- for the reasons you stated below.  I might
still like to encourage a "x-TRW..12345" structure.  Later when we define
standard systems for things like Strong's numbers, etc.., it would be easier
to convert to a standard form.

So, with this new understanding the x- seems like a good idea.  I would
still like to see an "x-RefSystem..identifier" structure but I also realize
that if people put the x- then it is for their own private use anyway.
(something like <xs:pattern
value="x-(\w(\.[\w]+){0,}))\.\.(\w(\.[\w]+){0,}))"/> )   I suspect that the
spec is final at this point, this sort of thing will have to be addressed in
the future.

Thanks, Chris.

Todd



> We need to define standard systems like Stong's numbers for lemmata,
> AGNT codes for morphology, etc. before we allow attributes without the
> x-.  We haven't and probably shouldn't have defined these yet, so we
> should continue to require x- for all values, mostly to show encoders
> that their data is essentially private use and that we will define a
> standard syntax in the future.  And we definitely need to require
> namespaces to identify the system used.
>
> --Chris
>
> > -----Original Message-----
> >
> > I don't think you want to use what you sent in any case.
> >
> > As it sits POS, morph, and lemma MUST start with "x-".
> >
> > Naturally I propose that we require the following regex in a
> > new simple type that we use for POS, morph, lemma, sync, and xlit.
> >
> > <xs:pattern value="(\w(\.[\w]+){0,}))\.\.(\w(\.[\w]+){0,}))"/>
> >
> > Todd
> >
> >
>
>