For those who have to work with output from spreadsheets. And I wonder
if this doesn't interact with OpenRefine output as well.
kc
-------- Forwarded Message --------
Subject: new W3C CSV on the Web specs, now at Candidate Recommendation
stage - please implement!
Resent-Date: Tue, 28 Jul 2015 14:33:21 +0000
Resent-From: public-voc...@w3.org
Date: Tue, 28 Jul 2015 15:32:52 +0100
From: Dan Brickley <dan...@google.com>
To: public-...@w3.org, semantic-...@w3.org Web <semantic-...@w3.org>,
W3C Web Schemas Task Force <public-voc...@w3.org>
Hi! Short version: Please see
http://www.w3.org/blog/news/archives/4830 for the Candidate
Recommendation specs from W3C's CSV on the Web group -
https://www.w3.org/2013/csvw/wiki/Main_Page
Long version:
These are the 4 docs,
"Model for Tabular Data and Metadata on the Web—an abstract model for
tabular data, and how to locate metadata that enables users to better
understand what the data holds; this specification also contains
non-normative guidance on how to parse CSV files"
http://www.w3.org/TR/2015/CR-tabular-data-model-20150716/
"Metadata Vocabulary for Tabular Data—a JSON-based format for
expressing metadata about tabular data to inform validation,
conversion, display and data entry for tabular data"
http://www.w3.org/TR/2015/CR-tabular-metadata-20150716/
"Generating JSON from Tabular Data on the Web—how to convert tabular
data into JSON"
http://www.w3.org/TR/2015/CR-csv2json-20150716/
"Generating RDF from Tabular Data on the Web—how to convert tabular
data into RDF"
http://www.w3.org/TR/2015/CR-csv2rdf-20150716/
See the blog post for more links including an extensive set of test
cases, our GitHub repo and the mailing list for feedback. Also note
that the approach takes CSV as its central stereotypical use case but
should apply to many other tabular data-sharing approaches too (e.g.
most obviously tab separated). So if you prefer tab-separated files to
comma-separated, do please take a look! The Model spec defines that
common model, the metadata document defines terminology for talking
about instances of that model, and the last two specs apply this
approach to the problem of mapping tables into JSON and/or RDF.
The group expects to satisfy the implementation goals (i.e., at least
two, independent implementations for each of the test cases) by
October 30, 2015. Please take a look, and pass this along to other
groups who may be interested.
cheers,
Dan
for the CSVW WG
p.s. since I'm writing I'll indulge myself and share my personal
favourite part, which is the ability (in the csv2rdf doc) to map from
rows in a table via templates into RDF triples. This is a particularly
interesting/important facility and worth some attention. Normally I
wouldn't enthuse over (yet another) new RDF syntax but the ability to
map tabular data into triples via out-of-band mappings is very
powerful. BTW the group gave some serious consideration to applying
R2RML here (see docs and github/wiki for details), however given the
subtle differences between SQL and CSV environments we have taken a
different approach. Anyway please take a look!
