Re: [legalcitem-technical] List of issues to discuss during today's call (longish)

[Fabio Vitali <fabio@cs.unibo.it>]

Dear Grant, 

>>>> Creating a software tool that does (a) is possible, but of course you run in these difficulties that you describe, and if you do not distinguish between the difficulties of (a.1) and (b), I think you are running headfirst in a wasp nest. I cannot help with the difficulties of (a.1), but I think we have a solid chance to create robust solutions for (b).
>>>> 
>>>>>> Disambiguation is necessary for a number of reasons:
>>>>>> 
>>>>>> - The normal versioning and amending reasons.
>>>>> 
>>>> This is (b) and we can help.
>>>> 
>>> GCV: One issue we have to address is that ambiguity can get introduced as some point after a reference is created. For instance, misnumbering cannot be anticipated. Requiring the necessary qualifiers be built into every reference would be cumbersome.
>> 
>> I think we need to be very careful with the terms here: the
>> **reference** cannot and must NOT be ambiguous, ever. The **citation**
>> of course can, and there is little that can be done. But once that you
>> have identified the features that are instrumental to selecting the
>> right document at any time, and that MUST be permanent, then the point
>> is creating a reference that uses these features only, and that
>> through the resolution process let you get to the document that in any
>> specific moment is the right answer to the ever-evolving citation.
>> 
> GCV2: Here is the problem. Assume that Section 101101 of Title 20
> exists in the U.S. code. It's stable and has been around for a while.
> It is routinely cited as 20 U.S.C 101101. The identifier we assign is
> "/us/usc/t20/101101". When we reference this provision, we reference
> it as "/us/usc/t20/101101". Now imagine that a legislator is running
> out-of-loop and introduces legislation that adds section 101101 to
> Title 20 of the U.S. Code. This happens from time to time. Nobody
> notices what has just happened. Now there are two perfectly valid and
> completely unrelated sections 101101 in Title 20 of the U.S. Code.
> Oops, but to fix it will take and act of Congress (an expression that
> we often use, but which applies literally in this case). Any
> pre-existing citations and persistent references will now be
> ambiguous. Perhaps we could assign a uniqueified identify to the new
> provision (say /us/usc/t20/101101-a), but we don't necessarily have
> the right interaction points in the process to do that - it's upstream
> of our involvement.
> 
> GCV2: What happens when duplicates like this occur is that new
> citations will read something like 20 U.S.C. 1011101 (as added by P.L.
> 123-456) where the qualifier specifies the origin of the provision -
> or the last amendment to the provision. This is what we must interpret
> to form new references. Old citations will be ambiguous as will old
> persistent references. I don't see how we can avoid this short of
> encoding the qualifiers in all references from the start. This seems
> cumbersome to me.

Two possible ways out, keeping in mind that identifiers MUST be permanently associated to documents and fragments in some way. 

1) Give non-positive and positive codes different base identifiers, so that you start using /us/usc-np/t20/101101, for the early instance of section 20 USC 101101 (np stands for non positive) and /us/usc/t20/101101, for the positive version, if and when it comes out. 

2) Adopt Akoma Ntoso Naming Convention, and accept the fact that the US Code (and therefore its titles and its is a work that changes and is updated to create new expressions. Therefore, suppose that title 101101 of title 20 exist2 since XX/YY/WWWW, and later on the legislator introduces a new title 101101 of title 20 on date ZZ/TT/UUUU. 

Then we consider them two versions of the same work, and differentiate them by version date: /us/act/usc/en@WWWW-YY-XX#tt_20__sect_101101 versus /us/act/usc/en@UUUU-TT-ZZ#tt_20__sect_101101. 

Ciao

Fabio


--

Fabio Vitali                            Tiger got to hunt, bird got to fly,
Dept. of Computer Science        Man got to sit and wonder "Why, why, why?'
Univ. of Bologna  ITALY               Tiger got to sleep, bird got to land,
phone:  +39 051 2094872              Man got to tell himself he understand.
e-mail: fabio@cs.unibo.it         Kurt Vonnegut (1922-2007), "Cat's cradle"
http://vitali.web.cs.unibo.it/