- From: Eric Stephan <ericphb@gmail.com>
- Date: Thu, 29 Oct 2015 12:38:15 -0700
- To: Laufer <laufer@globo.com>
- Cc: Phil Archer <phila@w3.org>, Annette Greiner <amgreiner@lbl.gov>, "Purohit, Sumit" <Sumit.Purohit@pnnl.gov>, Public DWBP WG <public-dwbp-wg@w3.org>
- Message-ID: <CAMFz4jhZBg9WX-Sctt_GRwof_8YTEMX+1eJO945-5PWD5PtYeA@mail.gmail.com>
Hi Phil, Busy day today, so this is just a quick note to express: 1) While I didn't know there was a specific term linksets, it makes a lot of sense because communities outside of the spatial data perspective are grappling with this problem. 2) Some examples that come to mind: a) The Atmosphere Radiation Measurement program has instrumentation that produces measurements. These measurements must be restructured in such a way that simulation modeler consumers can use their data. So temperature measurement data streams A, B, C are integrated to create a "Value Added Product" D to provide a composite of the best temperature measurements for the simulation modeler consumer. If I were reviewing results M produced by the simulation modeler, I might want to look at D and how D originated from streams A, B, C for a full disclosure record. b) Heterogeneous data collected in datasets or measured in a variety of ways and the composite of represent either a profile of a physical phenomena (performance metrics). c) A series quantum chemistry simulations are run to predict the behavior of molecules under various conditions. Each simulation represents a separate dataset. A linkset ties together a portion (e.g. lines 2200-4500 in file A of dataset Z) of each dataset distribution that represents an insight or anomaly. Does this help? Eric S On Tue, Oct 27, 2015 at 10:00 AM, Laufer <laufer@globo.com> wrote: > > > Hi, Phil, > > I think it is interesting. What I can see (at first sight) is that the > group is interested in how to give metadata about a relation between > linksets. In a broader sense, when our group was discussing versions, > we talked about the idea of datasets having different types of relations. I > think that here we have a relationship class that has properties that are > associated with a relation between two datasets. I think this also applies > to the case of diff between versions of a dataset. Summarizing, we have a > relation between two datasets (or linksets) that has properties. Another > thing is what vocabularies will be suitable for describing each set of > properties of this relation. > > Laufer > --- > > . . . .. . . > . . . .. > . .. . > > > > Em 27/10/2015 3:07, Phil Archer escreveu: > > Annette, Eric, Sumit, > > I have you in mind although this is open to everyone in the WG (Laufer?)... > > Again from the Spatial Data WG, there is a lot of discussion going on here > about Linksets and how to publish them. There is interest here in working > on a Note on this, but it needs to have non-spatial people concerned as > well since it's applicable to any kind of data. > > So we're talking about publishing information about how to interpret a > dataset in relation to others (A's garden is B's yard, A's foo includes B's > bar and baz) > > Does this resonate with you at all? > > Phil. > >
Received on Thursday, 29 October 2015 19:38:43 UTC