So the line of argument was that Open Refine, R2RML and GATE have
something in common as low-level 'something else in, RDF out' (to put it
crudely) technologies, whereas MIMIR, KIM, Solr integration, Stanbol,
etc, are higher level application building technologies.
I'd definitely say that the decision could be questioned, which is why I
raised it, but just to say that the intention is to arrange things into
digestible chunks (for webinar and book chapter), rather than to exclude.
Let's say the focus of Chapter 3 is "what are you starting with, and how
do you make RDF?", the focus of Chapter 4 is "how will you let people
interact with that data?" and of 5 "how do you pull these individual
things together?"
Feedback on this is, of course, as welcome as on the individual chapters.
Barry
On 29/03/13 09:34, Sebastian Schaffert wrote:
Hi Barry,
your project - of course I agree :-)
However, I don't really see what e.g. GATE would have to do with "providing linked
data" in any different way than Stanbol. Like Stanbol, it is also a framework and
API, so I'd move this to Chapter 5 as well.
Now reading the abstract on your webpage, I see you are actually referring to
GATECloud, so this might be a slightly different case - our company RedLink
(www.redlink.at, currently in the process of incorporation) will provide
Stanbol-as-a-Service in a similar fashion in the coming weeks (including a
for-free service). :)
Greetings,
Sebastian
Am 29.03.2013 um 10:17 schrieb Barry Norton:
Sebastian, Bernadette, Kingsley,
Just to note (lest the conversation slip too far from the original request)
that application-building, APIs, frameworks etc. are the subject of a later
EUCLID chapter (5) on which we will also consult.
We look forward to following up on some of these points then.
Note that OpenRefine and SKOS are being covered in the current chapter (3),
whose webinar is scheduled on 22nd:
http://euclid-project.eu/events/webinar-providing-linked-data
One valid question is whether you agree with how far we've taken text analytics
in Chapter 3 (generic services like DBpedia Spotlight, Zemanta, then use of
GATE) versus what we leave for Chapter 5 (Stanbol. etc.)
Barry
On 29/03/13 08:54, Sebastian Schaffert wrote:
Hi Maria and Bernadette,
Am 28.03.2013 um 20:05 schrieb Bernadette Hyland:
Hi Maria,
Happy to see you're compiling a survey of topics & tools. May I suggest adding a
category called "Linked Data Frameworks" (a peer to Linked Data Browsers).
For example, in the Linked Data Frameworks category it may include: Callimachus
Open Source, OpenLink Software's Virtuoso Open-Source Edition and TopBraid
Composer, and others. Note, I don't think these products are direct
competitors, rather a class of enterprise products that build on Linked Data.
I'll leave it to the respective companies to describe their products.
I'd like to add Apache Marmotta (http://marmotta.incubator.apache.org) and the
Linked Media Framework (http://code.google.com/p/lmf/
). Apache Marmotta implements a complete Linked Data Server with additional
features like resource-based updates, reasoning and versioning. It aims to be a
reference implementation for the Linked Data Platform once the recommendation
stabilizes. The Linked Media Framework builds on top of Marmotta and provides
integration with SKOSjs (thesaurus management), OpenRefine (for RDF-izing
legacy data), Apache SOLR (for Semantic Search) and Apache Stanbol (content
ananlysis and interlinking).
Both are aimed at developers and users who want to build Linked Data applications,
specifically in combination with Content and Media Management Systems (hence the name
"Linked Media Framework"). Apache Marmotta is currently undergoing its first
Apache release (some minor license details still to be settled). We expect it to be
available after the Easter holidays. The Linked Media Framework will follow with its next
release probably end of April, building on the most recent versions of Marmotta, Stanbol
and SOLR.
Greetings,
Sebastian
Sebastian
