My first thought is that step #1 is developing a formal statement of
the spec, probably in Jenni Tennison's "datatype library language" or
in DSDL Extensible Datatypes (ISO 19757-5).
I've never actually used either of these languages myself, but they
are explicitly intended for describing datatypes.
> First, this spec need to undergo some form of standardization,
> beyond this current process. Many people who are participating have
> made it clear to me that although their organization has a great
> deal of interest it cannot take it seriously or invest in it unless
> it is on a standards track.
> So with that as a premise, which standards body? First, I'll list
> those that I think are possible: W3C, IETF, and NISO. There are a
> few others that come immediately to mind that you will say "why not
> this one" or "why not that one", and I'm hoping to avoid that
> conversation but will discuss it if pressed. I will say this about
> ISO however: if this is standardized in W3C, it can now be
> fastracked (that might not be the current term but you know what I
> mean) in ISO. My preference is W3C. LC is a W3C member, I am the AC
> rep, and that might make the bureaucratic process nearly tolerable.
> And there are other technical reasons why I like W3C.
> However, I welcome discussion on this.