Dear John & John,
during the last LegalDocML TC, on June 29th, the group has analyzed your comments and discussed about a possible way to satisfy both your concerns and our need to preserve the consistency of our overall design of the Akoma Ntoso proposal.
In this message we would like to advance to you, informally, the gist of the solution, and once we receive your approval, proceed to actually draft and emit a new version of the documentation for the formal approval procedure.
Briefly, the most relevant and pivotal comment is on the naming convention and on your request to allow "on an equal footing" other naming conventions than Akoma Ntoso's own one. We understand the request and the underlying need, and we are inclined to accept that as long as we can express a few simple and reasonable requirements.
In brief, we would like to adopt your concept of "functionally equivalent naming convention" (FENC), specify that only FENCs are acceptable in Akoma Ntoso XML documents, and provide a few reasonable indications of what we mean by functional equivalence.
A functionally equivalent naming convention is defined as follows:
[ELI + OP] Disagree with most of these requirements. Why are they needed? Some of these requirements are highly subjective (when is something "memorizable", "global" etc.)? Who would measure these and decide on the possible compliance as functionally equivalent naming convention?
Any naming convention that declares itself to be functionally equivalent, is persistent and well publicly documented must qualify.
1) *recognizable*: a syntactical means exists to recognize the specific syntax used for the URI (e.g., a specific prefix);
[ELI + OP] This would largely be the case for ELI (and ECLI), but why this requirement? Also the /eli/ part is _not_ a mandatory part of the ELI identifier syntax (and the UK doesn't have it…).
2) *published*: a sufficiently detailed description of the syntax is publicly available and backed by a recognizable institution;
[ELI + OP] OK (assuming that "recognizeable institution" = relevant stakeholder). Otherwise, badly defined
3) *FRBR compliant*: A full distinction between "distinct intellectual creations", "specific intellectual forms", "physical embodiments" and "exemplars" of relevant documents must be explicitly supported and aligned with the FRBR conceptualizations. Support for items is not necessary nor requested.
[ELI + OP] Why is this relevant at the identifier level? As it is formulated, it is unclear if ELI complies to this or not
4) *CEN Metalex compliant*: the seven rules in CEN Metalex requirements (section 6.1 of ) must be fully implemented:
"To allow for the discovery of IRI identifiers, names must be:
1. Persistent: names at all levels must maintain the same form over time regardless of the political,
archival and technical events happened since their first generation;
[ELI + OP] OK
2. Global: all relevant documents by all relevant bodies must be represented;
[ELI + OP] What does this mean? Documents are globally accessible or the scheme is globally applicable? If the last, why?
What are "all relevant bodies"?
3. Memorizable: names should be easy to write down, easy to remember, easy to correct if they were written
[ELI + OP] In principle OK, though subjective
4. Meaningful: names should mean something; It should be possible to make assumption about the kind, freshness
and relevance of a citation by looking only at the document’s name;
[ELI + OP] OK (but what does freshness mean)
5. Guessable across levels: references to different levels of the same document must be similar; e.g.,
given a reference to an _expression_ a user should be able to deduce the name of the work;
[ELI + OP] OK
6. Guessable across document classes: references to different instances of the same document type must
be similar; and
[ELI + OP] OK
7. Guessable across document components: references to different components of the same document at the
same level must be similar."
[ELI + OP] OK
4) *active*: at least one working, accessible, available, robust resolver must exist that provides dereferencing of URIs/IRIs according to the specific syntax;
[ELI + OP] No resolver must be required, unless this includes is the general http resolution mechanism
5) *equivalent*: at least one working, accessible, available, robust converter must exist that converts URIs/IRIs according to the specific syntax into equivalent URIs/IRIs according to the Akoma Ntoso Naming convention;
[ELI + OP] In this case the Akoma Ntoso Naming convention is the privileged partner and not a choice on "an equal footing".
In practical terms, who would maintain such a converter?
6) *evident*: Akoma Ntoso XML documents identifying themselves (in <FRBRUri> and <FRBRThis> elements) using a FENC URI, must also provide equivalent <FRBRalias> elements with the URI ref corresponding to <FRBRThis> according to the Akoma Ntoso Naming Convention, one for each of the first three FRBR levels.
[ELI + OP] Unacceptable. As an effect of this everybody would have to support the Akoma Ntoso naming convention in any case, the only difference being that in this case ELI is the primary schema and Akoma Ntoso the obligatory second one. In this, the Akoma Ntoso naming convention retains a privileged position and is not at all "on an equal footing"
Any Naming Convention that complies with these requirements is termed a *functionally-equivalent Naming Convention* and its URIs can be used in any situation where Akoma Ntoso URIs/IRIs are appropriate.
Finally we resist at allowing custom syntaxes for inner-document ids in eId and wId, because
a) inner document identifiers are a reflection of the overall XML structure, which is still Akoma Ntoso, and not of the document-level URI syntax adopted,
b) allowing multiple syntaxes for the same attributes would create havocs in any decent resolver and editor trying to deal with documents coming from different sources, and
c) a good destination for custom ids already exists, attribute guid, that has exactly the stated purpose and allows custom ids without polluting the id space expressed by eIds and wIds.
[ELI + OP] Using the prescribed syntax must not be requirement for compliance to Akoma Ntoso nor for the use of the eId / wId attributes. If desired, an additional compliance level could be defined for those users of Akoma Ntoso who want to use a resolver. The use of a resolver must not be forced on users of the schema who have no interest in using this functionality. So, usage of a predefined inner identifier structure should be an option (CAN), not an obligation (MUST), except perhaps in said specific compliance level
We plan to work on a new draft including this kind of flexibility, and would appreciate your opinion within the next week.
Thank you for your comments and opinions on this.
Monica and Fabio
Il 21/06/2016 08:48, John Dann ha scritto:
of the LegalDocumentML
As Chair of
the ELI Task Force, and on behalf of
Office of the European Union
(comments also sent individually)
(comments also sent individually)
we would like
to send the following comments.
universal naming convention, given the differences
in national legal systems, is very complex and does
not cover all national legislation cases and
especially hinder existing naming conventions.
In line with
the principle of proportionality and the principle
of decentralization, each country and company should
continue to operate its own national Official
Journals, Legal Gazettes or legal databases in the
way they prefer. We should therefore carefully
consider not to impose a naming convention in order
to respect the legal and constitutional differences
between countries, and authorize on equal footing
other naming conventions, e.g. URN-Lex, ECLI, ELI
flexibility would help the implementation of the
revised version of AKN.
fully support the comments of the Office of
Publications of the EU sent earlier by email.
LE GOUVERNEMENT DU GRAND-DUCHÉ DE LUXEMBOURG
central de législation
Tél. (+352) 247-82961 . Fax (+352) 46 74
E-mail : email@example.com