On 3 February 2016 at 19:25, David Croft <[log in to unmask]> wrote:
> The biggest unsolved issue that we found though was not data that
> couldn't be automatically processed, but imprecise data, date fields
> that covered 300 year time spans for example or "circa 19th century".
> In the end we used multiple imprecise fields to do our searching but
> without manually processing I don't currently see any obvious way to
> handle the imprecision of much of the data as recorded at the moment.
That said, you could link to (or create) identifiers that represented those
fuzzy dates or locations. For example, PeriodO http://perio.do/ is a
gazetteer of 'scholarly definitions of historical, art-historical, and
archaeological periods' - so you could publish your definition of the
'early Roman England' period and link your records to it, and others with
collections from the same period and place could re-use your identifier and
link their records to your 'early Roman England'. Similarly, Pelagios
http://pelagios-project.blogspot.co.uk/p/about-pelagios.html allows you to
link to identifiers for 'places in the historic past'.
Comma-separated (CSV) files are easy for people to use and the records are
kinda human-readable, but the records we share in any format would be
easier to re-use if they had just a bit more structure - if they used URLs
as well as strings for people, places, events, etc. For example,
http://dbpedia.org/page/Frances_Oldham_Kelsey is less ambiguous than
'Frances Oldham Kelsey' and makes it easier to link information about her
together across collections. As Richard says, linking to people, historical
events, etc, would be even more useful. However, if you've read this far
I'm probably preaching to the converted, and we all know that resources for
cataloguing are pretty scarce so realistically, for now, we might only get
new or a selection of re-catalogued records with identifiers that link
outside the institution.
I've speculated wildly on other reasons why cultural institutions don't
publish more structured data in
http://www.openobjects.org.uk/2015/07/the-good-the-bad-and-the-unstructured-open-data-in-cultural-heritage/
and I'd always be glad to hear more success stories to balance the gloom!
Or failing that, more reasons why organisations aren't publishing linked
data - e.g. is the gap between people who'd like their data to come with
identifiers and the people who publish it bigger than we thought?
Crowdsourcing (*cough* http://www.libcrowds.com/ *cough*) might help fill
some of the cataloguing backlog, but it's no panacea.
While I'm linking to blog posts, my notes for/from the Nesta/ACE/AHRC
Digital Culture survey discussion are at
http://www.openobjects.org.uk/2016/01/the-state-of-museum-technology/ It
was a packed event and the data is available online, so I'm sure others
here have opinions on it too.
Cheers, Mia
--------------------------------------------
http://openobjects.org.uk/
http://twitter.com/mia_out
Check out my book! http://bit.ly/CrowdsourcingCulturalHeritage
<http://bit.ly/CrowdsourcingCulturalHeritage>
I mostly use this address for list mail; my open.ac.uk address is checked
daily
****************************************************************
website: http://museumscomputergroup.org.uk/
Twitter: http://www.twitter.com/ukmcg
Facebook: http://www.facebook.com/museumscomputergroup
[un]subscribe: http://museumscomputergroup.org.uk/email-list/
****************************************************************
|