FRBR URI components can be ambiguous

[Greg Kempe]

Hi Akoma Ntoso fans,

We are starting to use actors in our FRBR URIs, and I believe the FRBR URI

  /akn/za/act/xxx/1992/

is ambiguous, because the xxx could be a document subtype or an actor.

Is there a suggested workaround? Perhaps "-" can be used for an empty subtype component when an actor is used, to indicate that the subtype is explicitly empty, eg:

  /akn/za/act/-/actor/1992/

Thanks,

Greg

--

Greg Kempe

CTO and Co-founder, Laws.Africa

https://laws.africa · gr...@laws.africa · +27 78 246 1116

[Fabio Vitali]

Dear Greg,

TL;DR: use "-" as the document subtype ONLY IF the document you are describing is the highest form of (unnamed) legislative document of the specified jurisdiction AND this (unnamed) document may come from two or more different actors, otherwise use FIRST the specific name of the document type AND THEN specify the actor.

---

In the original intention of the naming convention of Akoma Ntoso, the document type fragment represents a very wide range of documents, of which, say, "act" is meant for all types of enacted norms including legislative statutes and laws as well as executive regulations and decrees, etc.

In order to be very precise, one would use the document subtype to specify exactly which type of document, among the many that exist, you are referring to. If you do not specify the document subtype, there is the long lasting assumption that the "act" you are referring to is a piece of official legislation by the highest legislation authority in the jurisdiction specified, e.g., the National Parliament. Thus

> /akn/za/act/1992/123

refers to act n. 123 of 1992 by the Parliament of South Africa, which is the highest authority of the Legislative Power in South Africa.

In the case of other, lesser documents, the specific type of the document becomes fundamental to reduce the ambiguity, as you can have regulations, decrees etc. each with a specific name of its document type, and for this reason we maintain the "act" type and we offer a specific fragment for the local name of the document, if appropriate.

Sometimes, even this is not enough, as for instance we do have ministerial decrees that have the same name for the document subtype (e.g., "Ministerial Decree") but come from different ministers, and therefore need the actor (e.g., Minister of Commerce vs. Minister of Internal Affairs) to guarantee disambiguation.

For this reason I do not think that the Naming Convention as it is now is ambiguous: in all the situations we have seen so far, there is only ONE legislative authority generating nameless legislation deserving to be called "act" without further specifications.

As far as I know, we have never encountered a situation whereby the highest, nameless legislation comes from two different legislative authorities of the same, highest level, which is what I would imply by seeing something like

> /akn/za/act/-/actor/1992/

This form would be acceptable only in this case: that of two equally authoritative legislative bodies emitting two equally important enacted legislation at the highest level of the specified jurisdiction. If you are sure this is the case, then "-" seems a reasonable solution.

BUT,

if, as I am more incline to assume, these situations describe two identically named document types containing lesser norms and coming from different actors, such as different executive bodies, then the name of the document type becomes suddenly relevant and appropriate to specify, and I would use that name instead of "-" in the document subtype fragment of the URI.

I hope I was clear enough.

Ciao

Fabio

--

> --
> You received this message because you are subscribed to the Google Groups "akomantoso-xml" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to akomantoso-xm...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/akomantoso-xml/CAB%2Bz6DRkGwyR%2Bj2XZbhwwPvdZDhrbHPHxpTFWC6g8qVRwppcEA%40mail.gmail.com.




--

Fabio Vitali The sage and the fool
Dept. of Informatics go to their graves
Univ. of Bologna ITALY alike in this respect:
phone: +39 051 2094872 both believe the sage to be a fool.
e-mail: fa...@cs.unibo.it Where, then, may wisdom be found?
http://vitali.web.cs.unibo.it/ Qi, "Neither Yes nor No", The codeless code

[Greg Kempe]

Hi Fabio,

Thanks for this clarification and background, it makes sense to me when applying the naming convention to create an FRBR URI.

However, I still think there is ambiguity from a purely technical standpoint when consuming an FRBR URI: if I'm writing a parser to recognise an arbitrary FRBR URI and extract its constituent components, then 

  /akn/{country}/{doctype}/xxx/{year}/{number}

is ambiguous, no matter the doctype. So I still believe it's necessary to resolve this ambiguity in the standard in the case of an actor without a subtype, by indicating an explicitly empty subtype.

My example was poorly chosen, my apologies.  We're looking at using actors for judgments and not acts. In this case, a judgment is uniquely identified by the year, a court code and a serial number. So we have the option of the court code as the actor:

  /akn/{country}/judgment/{court}/{year}/{number} 

Or the court code as a sub-country code:

  /akn/{country}-{court}/judgment/{year}/{number}

Currently we're looking at the first option, which raised the ambiguity issue.

Thanks,

Greg

To view this discussion on the web visit https://groups.google.com/d/msgid/akomantoso-xml/BC978B2B-6829-416B-8533-9F4654CE5632%40gmail.com.

[Ashok Hariharan]

Dear Fabio -- 

There are a couple of other things I have noticed which are ambiguous about <FRBRauthor>. 

1) the release notes seem to suggest its optional in FRBRwork, but in the schema itself if i recall minOccurs is not set, so defaults to 1. 

2) the FRBRauthor value is a @href xsd:anyURI, however the release notes say that when used as a part of the document IRI , the IRI part for author/actor must be set to the VALUE which seems to imply a string literal rather than a URI - so far i have used it to point to a TLCRole (via @as in FRBRauthor) and used the end part of the of TLCRole IRI as the literal (e.g. "FinanceCommittee" from "/akn/ontology/role/blah/FinanceCommittee" )    in the FRBRauthor IRI part.  However, this is not what the release notes say  :-)  

Perhaps the release notes need to be updated to clarify this ? 

Ashok

To view this discussion on the web visit https://groups.google.com/d/msgid/akomantoso-xml/CAB%2Bz6DRvZwcY3%2BtZAsmjdYNQ4H%2BqeNypd6bnW9HduvqsCAgSXA%40mail.gmail.com.