Re: Definition of Shape

On Sat, Jan 24, 2015 at 5:48 PM, Peter F. Patel-Schneider <
pfpschneider@gmail.com> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> The glossary that I put together was at least partly an attempt to answer
> these sorts of questions.
>

Yes, and I really think it is very useful...that's why I dare to contribute
:)

In the glossary a shape (or unscoped constraint) is something that can only
> be evaluated against something, e.g., people with at least two friends
> would
> be a shape.


Do you propose that a Shape is the same as an "Unscoped constraint" ?

If you think it is right so, should we put that definition in the "Shape"
slot? I am asking because I thought that a glossary on a Data Shapes
working group should have an entry on "Shapes"...

What I was trying to stress with my proposal is that in the context of this
group a shape involves the triples surrounding an RDF node. I think it is
an important concept that can serve people who comes to the group to
understand what we are talking about...


> A scoped constraint includes both a shape and an indication of
> how it is to be applied, e.g., all people must be people with at least two
> friends would be a scoped constraint.  I would actually prefer to use just
> "constraint" for scoped constraints but I was trying not to push this
> particular preference for the meaning of constraint.
>

I think it makes sense to indicate the difference between "Scoped" and
"Unscoped" constraints...maybe, in my proposal I was suggesting that a
Shape was composed of a set of constraints...which could also be called
"Compound constraints", but I would prefer not to complicate the glossary...


> peter
>
>
> On 01/24/2015 08:08 AM, Karen Coyle wrote:
> >
> >
> > On 1/23/15 10:34 PM, Jose Emilio Labra Gayo wrote:
> >> On Fri, Jan 23, 2015 at 8:30 PM, Peter F. Patel-Schneider
> >> <pfpschneider@gmail.com <mailto:pfpschneider@gmail.com>> wrote:
> >>
> > The glossary already contains a stab as this:
> >>>
> >>>
> https://www.w3.org/2014/data-shapes/wiki/Glossary#Unscoped_Constraint.2FShape
> >
> >>>
> >
> >>> Yes, but I think we should have a simple definition of what a Shape
> >>> is and I think this definition could be it.
> >
> >> +1 for simple definition
> >
> >> I also think that in the
> >>> context of this group we can differentiate between a constraint and
> >>> a shape, saying that a shape is a set of constraints on some RDF
> >>> node.
> >
> >> Does the shape constrain? or does it define? In the Dublin Core work
> >> the "shape" was called a "description" -- it describes the graph and
> >> the desired rules for the graph. Whether one uses those rules to
> >> constrain, or to inform, or to reject, or to simply shake one's head in
> >> dismay is a matter for the application that applies the rules. "Shape
> >> expressions" is very close to this meaning, and is looked on favorably
> >> within the DC community.
> >
> >> kc
> >
> >
> >
> >>> Apart from that, we can maintain the definitions of "constraint",
> >>> "scoped constraint" and "unscoped constraint" which I also think
> >>> they make sense.
> >
> >>> Best regards, Jose Labra
> >
> > peter
> >
> >
> > On 01/23/2015 10:10 AM, Jose Emilio Labra Gayo wrote:
> >> I edited the Glossary page to add a possible definition of shape as:
> >
> >> "A Shape is a set of constraints over the properties and objects
> > of an
> >> RDF node"
> >
> >> I think it is in accordance with the use of the term in the WG
> > and does
> >> not enter in conflict with other terms.
> >
> >> -- Saludos, Labra
> >>
> >>
> >>
> >>
> >> -- Saludos, Labra
> >
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1
>
> iQEcBAEBAgAGBQJUw8z6AAoJECjN6+QThfjzx3oH/1ED2hwHUZX6C94Ph6MW19fM
> IYMUu3vS+s4XwwtPFPkqnMG8BrwtyzYKolWX5Eocd2Q+X7dAJrJMdeHQIX/EmjIZ
> 4kWFjh6fBsN8GzfDCy4qfUkVhGNf4Bzl/sKM3q2UbapNRX/VId96xSVPc/cC1jDL
> h6nRktcRuQIQx9W4Jg457nO8S+34ny6yV7t6J0CRziIyw5n5L4XgEOULblp5/mMT
> FmrDgoo7t++stZeP1HM8BSCgKmDr8dVDE9yS0+QggJU2GtpCpdrTxWqaA54rR38b
> P2SJlrd1VLJDg69vOs+KKlqNcoX2CIjOvbidur6inqBc0Ki/2prn6YSokqkiku0=
> =0HPS
> -----END PGP SIGNATURE-----
>
>


-- 
Saludos, Labra