- From: Kerstin Forsberg <kerstin.l.forsberg@gmail.com>
- Date: Sun, 21 Sep 2014 12:49:09 +0200
- To: Andrea Splendiani <andrea.splendiani@iscb.org>
- Cc: Michel Dumontier <michel.dumontier@gmail.com>, Joachim Baran <joachim.baran@gmail.com>, HCLS <public-semweb-lifesci@w3.org>
- Message-ID: <CAKi6PjVmQ-KcRaMp0Kp6ZN1OJs41spbXdcUpp+9um+DdyHxioQ@mail.gmail.com>
Hi Andrea, in an earlier attempt to design and launch a Metadata Registry for clinical trial data, called Clinical Reference Library1 ). To capture and manage descriptions of versions of clinical trial data standards, and of variants of actual clinical trial datasets, we applied the software pattern called Facade 2). We used it to manage variants of metadata items on different gramualrity (e.g. data element, value domain, datasets) within a shared facade. It required a strong configuration management approach and hence an standard enginering approach similar as to software engineering. Cheers Kerstin 1) http://www.slideshare.net/kerfors/designing-and-launching-the-clinical-reference-library (slide 2 highlights the Realities of clinical trials data: the variances, changes, diversities and gaps) 2) http://en.wikipedia.org/wiki/Facade_pattern 2014-09-21 12:21 GMT+02:00 Andrea Splendiani <andrea.splendiani@iscb.org>: > Hi, > > I may re-use some bits of it, but overall I am dealing with quite a > different thing. > I don't have "publications", I have evolving information sets. > Provenance/evidence and the like are there, but not so fine-grained (e.g.: > I may have the whole ontology with the same provenance/evidence, not a few > statement). In same case (small subset) I have some more fine-grained > information. In this case I may pickup something from nanopubs, though I > have a string focus on capturing evolution of knowledge rather than "facts" > (e.g.: some facts gets validated). > There is also an are I don't know how to fit in, from the nanopubs point > of view, because facts come with a history of discussion behind. > Another aspect that I think it's different is, whatever I have, it's id > centric, and entity centric in the specific (like a dictionary). > So identifiers (and the relations between identifiers and identifiers of > versions) comes first. > > best, > Andrea > > > Il giorno 19/set/2014, alle ore 20:07, Michel Dumontier < > michel.dumontier@gmail.com> ha scritto: > > > Hi, > > I suggest nanopublications to track versioning for assertions > > http://www.nanopub.org/guidelines/ > > > > m. > > > > On Fri, Sep 19, 2014 at 9:58 AM, Joachim Baran <joachim.baran@gmail.com> > wrote: > >> Hi! > >> > >> On 19 September 2014 09:45, Andrea Splendiani < > andrea.splendiani@iscb.org> > >> wrote: > >>> > >>> When a concept change meaning, it changes id ;) > >> > >> Aha! I think it might not always change ID! ;) > >> > >> Best wishes, > >> > >> Kim > > >
Received on Sunday, 21 September 2014 10:49:35 UTC