Motivation for LPF Working Group

[Stephen Gadd]

Hello, everyone. Pardon the intrusion on your day, but it seems to me that we're each in or between one of two camps with regard to LPF:

1. We should aim for a Minimum Viable Product because: 
   • complexity deters non-specialists,
   • software does not support extensions,
   • we cannot possibly anticipate every possible requirement, and
   • it more-or-less does what it was originally intended to do.
2. Always maintaining backwards-compatibility, we should adapt and extend LPF as much and as often as necessary to accommodate every conceivable use case because:
   • a simple set of minimum requirements can always be extracted to comfort non-specialists,
   • in most cases other standards for representing relevant concepts have already been debated and developed, and could be adapted and adopted,
   • the lack of an appropriate standard for any single use case results in a proliferation of mutually-incompatible solutions, while
   • development of software support for extensions would be stimulated by a consensus of methodology.

Is that a useful summary, and what have I missed? Do we need to identify as a Group with one or other camp before we can move forward?

Best wishes,

Stephen

[Karl Grossner]

Thanks for this, Stephen - I agree a consensus on something like either #1 or #2 would be good to do, for clarity.

From my perspective, accommodating "every conceivable use case" is not tenable -- and I don't think it's necessary. LPF models temporally scoped places, which is a handful. Places have lots of relations to other things, events and other places mainly, and other things are involved in those events. Disclosure: I have always had an event-centered view of the big picture of historical data modeling.

Models for events and other things are usually content to leave Place as a relatively simple object at the end of a triple. If the URI to that object points at an LPF attestation or a nice rich "cluster" of annotations for a place in LPF, in WHG for example, great! LPF is aimed at being something a little more than MVP for Place as a subject.

All that said I don't think LPF is quite yet the MVP - for example, methods of asserting uncertainty and/or confidence should be there for both location and time/when, and be consistent. I like scales of 0 to 1, because they can be mapped to other likert-ish schemes people use (e.g. certain, less-certain, uncertain). And I'd like to see some proposed scope notes for "uncertainty" and "confidence." I need to go back and re-read my favorite papers on the topic, by Brandon Plewe [1, 2]

On the topic of how to get things done (i.e. make decisions) I thought the conversation last time came round to "well there's an editor, and after open discussion, the editor proposes a change most seem to favor, then everyone discusses a bit more then votes on the implementation." Do I have that right? If so, I'm happy to be the editor for the time being b/c it's so integral to WHG, which is in very active further development. And, if that's right, I need to make a proposal for uncertainty/confidence for you all to consider.

Please, comment away all!!

Karl

[1] Plewe, Brandon. "The nature of uncertainty in historical geographic information." Transactions in GIS 6.4 (2002): 431-456.

[2] Plewe, Brandon S. "Representing datum-level uncertainty in historical GIS." Cartography and Geographic Information Science 30.4 (2003): 319-334.