McBride, Brian wrote:
> I’m looking at converting the information represented in the XML files 
> on http://ukparse.kforge.net/parlparse to RDF in linked open data 
> style.  If anyone is already doing this, please let me know so I can 
> either help or get out of the way.

There has definitely been discussion in the past by people like Keith 
Alexander and Graham Higgins - e.g. 
http://semwebdev.keithalexander.co.uk/blog/posts/twfy-rdf-mapping and 
http://project.knowledgeforge.net/semwebparlparse/

> I’d like to understand the semantics of the id attributes in files like 
> all-members.xml.  For example, the first member element describes Diane 
> Abbott representing Hackney North and Stoke Newington from  1997-05-01 
> to 2001-05-14 and has an id of “uk.org.publicwhip/member/1”. 

Yep. You also want people.xml which is meant to match all the member IDs 
of one person to one person ID.

> Basically I would like to understand how these ids might change with 
> time.

That particular ID will certainly never change.

 > I’d also like to understand the semantics of the various hansard ids.

They are/were for matching up with the historic Hansard data Parliament 
made available, they're not actually used by TheyWorkForYou itself.

> Leaving aside processing errors etc, is it possible that in a future (or 
> earlier) published version of all-members.xml a different id might be 
> used to identify the same “member” or record i.e. same person 
> representing the same constituency for the same period?

No. A member ID is even more limited than that - it's the same person 
representing the same constituency for a particular party/name for that 
period. ie. a name change/ party change mandates a new entry and a new 
member ID. This does mean that it's possible for a correction to be made 
to a previous ID (say someone got married halfway through a Parliament, 
changed their name, and we missed it - unlikely, but say - then we would 
need to create a new member ID for their new name and change the date 
range for the old member ID).

> Leaving aside processing errors etc, is it possible that in a future (or 
> earlier) published version of all-members.xml the id 
> “uk.org.publicwhip/member/1” might be used to identify a different 
> “member” or record i.e. where the person or the constituency or the 
> period is different?

Only to correct an error as above (I'm not sure if that's what you meant 
by processing error) - it (and all its details) would never change if 
they're correct, and the ID would certainly never point at a different 
person.

> Are any of the ids used in the various xml files intended to remain the 
> same across time and this be suitable for constructing URIs?

All member and person IDs are meant to remain the same. There are/ will 
be anomalies in the person IDs where e.g. we have two entries for 
someone who was both an MP and a Lord and we didn't "join them up" - in 
which case one of the two person IDs will be removed (and not reused) 
and both member IDs will be matched to that person ID.

Constituencies are a bit more haphazard - historical ones are again 
based on the historical Hansard data and the from/to of those are 
broadly just name-based, but with some other things too. I wouldn't 
infer anything about constituency IDs beyond they cover a range - ie. 
for the next general election, it's probably easiest to create new 
constituency IDs for all the English/Wales/NI constituencies given 
they've all had boundary reviews, even if actually a few of them have 
had no changes. Plus some have had substantial changes but kept the same 
name.

> I’m figuring out how to create URIs for members, constituencies etc so I 
> need to know the scope in which the various ids in the xml files 
> identify things.

Hope that's helpful.

ATB,
Matthew

_______________________________________________
Mailing list [email protected]
Archive, settings, or unsubscribe:
https://secure.mysociety.org/admin/lists/mailman/listinfo/developers-public

Reply via email to