#CAAPerth Day Two – afternoon sessions
I am in a few sessions this afternoon.
First:
Is there time for archaeology? Understanding time through modelling and representation
Chair(s): Xavier Rodier, Lahouari Kaddouri
Format: Long Paper Presentation
Schedule: Tuesday 26th 13:30 – 15:00
Room: Seminar Room 2
Venue: University of Western Australia Club
Second:
E pluribus unum – Connecting People, Connecting Data
Chair(s): Guus Lange, Matthias Lang
Format: Long Paper Presentation
Schedule: Tuesday 26th 10:30 – 15:00
Room: Case Study Room
Venue: University of Western Australia Club
<live blog>
14:44
Guus discussing work at his institution on PCA applied to undertake automatic ceramic assemblage classifications.
14:42
Interesting, Lange noting Lots of levels of abstraction from real objects to data, many interpretive steps in classifying etc.#CAAPerth
— Eric Kansa (@ekansa) March 26, 2013
14:40
Guus’ example of term variation and disambiguation reminded me of the @sotonDH SAMTLA talk from a few weeks ago. This discussed a tool that enabled exactly the form of multiple-term searching. Watch the SAMTLA seminar video here.
14:31
Guus Lange talking now: Beyond metadata; computing archaeological knowledge
14:29
Q&A: mentioned Sebastian Heath‘s nomisma project as an example.
14:23
Holly discussing a University of Southampton resource held on the ADS: Roman Amphora a Digital Resource. This resource has had 370,000 page views this year. Users stay for extended periods on many parts. It is by far the most used resource the ADS hold. There are twice as much as our entire Grey Literature Library (18,000 fieldwork reports).
In Holly’s work she has tried to map this dataset several times, due to the complexity of the process and the many possibilities. @leifuss suggested that it should be aligned to locational data for example. Reminded also of this tweet and the surrounding discussion #caauk:
@mr56k making linked data dreams come true by skosifying archaeological thesauri #caauk #whyhavewewaitedsolong
— leifuss (@leifuss) February 23, 2013
Holly found that even stripping out skos:related it was too complicated. She decided it was  better to create a variety of classifications – need to take the literals (skos:altlabel) and create an RDF node.
Big questions:
- If we start creating nodes for other people’s data this goes beyond our remit
- Is it better to be useful or preserve the depositors intellectual process, reflected in the archive
Future work:
Use SKOS thesauri as a base. Using open annotation data model and ontology. Get involved more in SENESCHAL and ARIADNE.
14:15
Update from other session:
I get asked this challenging question: How to manage likelihood of truth in a conceptual model? #caaperth #caa2013
— César González-Pérez (@verdewek) March 26, 2013
14:09
GeoNames, DBpedia, Library of Congress, NERC and Ordnance Survey are examples of Linked Data repositories used by the Archaeology Data Service.
14:04
Holly introducing Linked Data work at York. More info on this from the @ArchCRG seminar by Michael Charno available as a video here.
14:01
Moved room. Holly Wright about to start.
13:54
Q&A: Can you quantify uncertainty within this model? Yes but not implemented becuase of lack of agreement over how to ascribe uncertainty values (ironic!). I also asked about whether a provenance framework e.g. PROV had been mapped to this work. The answer is not yet but interest in it.
Interesting talk by csic.academia.edu/CesarGonzalezP… #caaperth object oriented data structures for repres space + time < @mikejewell #patinaproject
— Graeme Earl (@GraemeEarl) March 26, 2013
13:50
Benefits of employing an object-oriented data structure? CHARM provides the abstract reference model. Formal support: mark temporal features, allow for phases, capture ontic change, good to represent objectivised knowledge e.g. building materials. Representation of ubjective features allows for varying perspectives, captures epistemic change, represents interpretative knowledge.
13:45
We can use object slices – slices of the data object created over time and varying subjectivities, with each linked to the originator of this slice e.g. the name of a person who published an interpretation. Can then use diachronic composition to enable complex entities to be built from simpler ones e.g. Hall + Narthex = Church of Santa Olalla; subsequently elements added and these can also be represented. Because of the use of a conventional object representation we can reuse formal support for navigating object data structures.
13:40
Problem of hegemonic and static representations of the world in digital data structures. Using as an example a 19th C Galician church built on top of a crypt. Varying interpretations – Roman nymphaeum, Roman sanctuary to Cybele, Roman funerary place, Proto-Christian temple, Catholic temple, and deconsecrated monument. Alongside these interpretations lie a conventional phased timeline associated with the material remains and their transformation. How capture this complexity?
13:35
First up Expressing Temporal and Subjective Information about Archaeological Entities presented by C.Gonzalez-Perez