[osis-core] reference systems: Issue 1

[Harry Plantinga]

The proposal assumes that if there is a reference system
xxx.yyy, then there is also a reference system xxx and a 
xxx <--> xxx.yyy transformation. The transformation can be
the identity in some cases, but there is always a transformation.

There's no "fallback" in the sense that no transformation is
required, but fallback is always possible in the sense that
we can perform the necessary transformation with existing
information.

Bible, Bible.French, and Bible.French.LSV need not be related
in any way except that there is a Bible <--> Bible.French
transformation and a Bible.French <-->Bible.Frency.LSV.

-Harry


> In some cases the identifiers separated by "." only specify a 
> more specific work and in other cases a transformation is required.
> 
> In the case of Bible.NIV.1993.en:Ps.45.3 I can simply drop 
> ".NIV.1993.en" and use "Ps.45.3" as Bible.Ps.45.3 or drop 
> ".1993.en" and use "Ps.45.3" as Bible.NIV:Ps.45.3 without 
> translation, BUT with Bible.French.LSV:Ps.45.3 there is a 
> transformation required at by French.  Do we want to create 
> namespaces that require transformation of "fall back"?  If we 
> do how can we indicate that a transformation is required. 
> 
> Possible solution: use upper case letters to identify that an 
> identifier requires a transformation.
> 
> Example:
> Bible.niv.en:Ps.45.3 does not require a transformation when 
> falling back to Bible.
> 
> Bible.French.lsv:Ps.43.3 does require a transformation when 
> falling back to Bible:Ps.44.3 but does not require a 
> transformation when falling back to Bible.French:45.3!
> 
> There is a problem when using numbers because they have not 
> case.  We could force identifiers in the "namespace" to start 
> with case oriented identifier and/or we could reserve a 
> transformation identifier.  (Ex:
> Bible.niv.-1993.en)
> 
> Todd
>