Re: LRR progress

[Frank Bennett <biercenator@gmail.com>]

One feature of the LRR identifiers that may need some attention is the
decision to join the element with a semi-colon:

    us;c4;va.wd;district.court

The URN:LEX schema would join the leading "scoping" elements with a colon:

    us:c4:va.wd;district.court

As the leading elements and the court identifier are different
semantic elements, this allows implementations to distinguish between
the two - and to determine whether the latter is present in the
identifier - without making fragile assumptions about the composition
of identifiers being processed.

While the LRR identifiers are a separate namespace from URN:LEX, I am
strongly inclined to adopt this feature of the latter's syntax, as it
just seems to be the right design choice.

Frank



On Sun, Jan 18, 2015 at 6:49 PM, Frank Bennett <biercenator@gmail.com> wrote:
> A number of new jurisdictions appeared in the LRR today. Most are just
> stubs, added in connection with work to move the MLZ reference manager
> to the new LRR IDs.
>
> MLZ carries a controlled list of jurisdictions, based on a JSON file
> originally hosted here:
>
>     https://github.com/fbennett/mlz-jurisdictions
>
> The identifiers in that file were not well-conceived. I had casually
> followed the URN:LEX draft at section 2.4, which recommends the use of
> ISO3166 codes for national jurisdictions - by pulling the whole set of
> IDs. What I failed to realize is that the ISO3166 is a _statistical_
> classification scheme, and many of the codes apply to entities that
> are under the jurisdiction of parent states. Some were even
> uninhabited islands ... [**1]
>
> To effect a cleanup of MLZ codes (which are potentially recorded in
> user databases scattered around the Internet), it was necessary to set
> up a counterpart for each in the new scheme (apart from the, er,
> uninhabited islands). This required that some jurisdictions with
> lingering colonial territories (Denmark, Finland, France, the
> Netherlands) be added, although their proper content is not yet ready.
>
> Small steps. The logic of this should become clear when the revised
> MLZ client becomes available, probably in a few weeks.
>
> [**1] URN:LEX controls for this by requiring that IDs be fixed by an
> official national authority. That solves one problem, at the cost of
> slowing things down considerably, but it potentially opens a thorny
> nest of controversy over what constitutes a "national authority" in
> disputed territories.
>
>
> Frank