- From: Esko Nuutila <esko.nuutila@tkk.fi>
- Date: Fri, 4 Apr 2014 12:51:24 +0300
- To: Abraham Bernstein <bernstein@ifi.uzh.ch>
- Cc: "public-rsp@w3.org" <public-rsp@w3.org>
- Message-ID: <CACeuo1367Y92gv=E6zrE5dsqn5dvGNgsaVzOpYEyJH+NjkAZ_A@mail.gmail.com>
Hi! I think that the stream should be a sequence of RDF datasets. These datasets then could contain timestamps and intervals describing the time related properties of the data. After all, RDF is a general data representation language, originally developed for representing metadata: The *Resource Description Framework* (*RDF*) is a family of World Wide Web Consortium <http://en.wikipedia.org/wiki/World_Wide_Web_Consortium> (W3C) specifications <http://en.wikipedia.org/wiki/Specification> [1]<http://en.wikipedia.org/wiki/Resource_Description_Framework#cite_note-1>originally designed as a metadata <http://en.wikipedia.org/wiki/Metadata> data model<http://en.wikipedia.org/wiki/Data_model>. (http://en.wikipedia.org/wiki/Resource_Description_Framework) Timestamps and intervals are metadata. I believe that adding another layer of metadata on top of RDF is unnecessary. It also makes RDF streams less compatible with existing tools like RDF stores. In addition, we should have a well defined library of time and interval operations, which could be used in SPARQL queries. I think that the idea of attaching timestamp or intervals to stream data items is valid in a situation where the actual data items are of some simple or restricted type, for example, real values coming from a sensor. In the case of RDF streams, however, it is more flexible to have all that data represented in RDF. /Esko On Fri, Apr 4, 2014 at 11:30 AM, Abraham Bernstein <bernstein@ifi.uzh.ch>wrote: > Dear all > > I just to the Query semantics page ( > http://www.w3.org/community/rsp/wiki/RSP_Query_Semantics ). Reading the > text I saw that the page says: > > G = {( subj, pred, obj)} > > and > > A stream is a sequence (G, τ ) where τ ∈ T the set of all timestamps. > > This worries me! > > As we discussed a number of time (I believe I raised the issue back last > September http://lists.w3.org/Archives/Public/public-rsp/2013Sep/0027.htmland it seemed to me that we got agreement on the issue > http://lists.w3.org/Archives/Public/public-rsp/2013Dec/0010.html) *intervals > are elemental*. Many things do not happen in one moment of time but are > really things that are happening over time. Hence, I would prefer a > formulation like the following: > > A stream is a sequence (G, τ ) where τ ∈ T the set of all *intervals*. > > τ = [t_s, t_e], where t_s denotes the start time and t_e denotes the end > time. > > As a short hand notation we could say that when t_s = t_e then one can > only write one of them. > > What do you guys think? > > Avi > > > ----------------------------------------------------------------- > | Professor Abraham Bernstein, PhD > | University of Zürich, Department of Informatics > | web: http://www.ifi.uzh.ch/ddis/bernstein.html > > -- Esko Nuutila Aalto University School of Science and Technology Department of Computer Science and Engineering P.O.Box 15400 (T-building, Konemiehentie 2, room B218) FI-00076 AALTO, FINLAND tel. +358 50 5750278 mailto: esko.nuutila@aalto.fi
Received on Friday, 4 April 2014 10:39:09 UTC