W3C home > Mailing lists > Public > semantic-web@w3.org > November 2013

Re: Deprecate http://www.w3.org/1999/02/22-rdf-syntax-ns# in favour of /ns/rdf# ??

From: Dan Brickley <danbri@danbri.org>
Date: Thu, 28 Nov 2013 15:56:55 +0000
Message-ID: <CAFfrAFpBOYfcZ4N5=j2uyarsCrnE0g22=1wuRKnaSEzFRx-rhA@mail.gmail.com>
To: Phil Archer <phila@w3.org>
Cc: Semantic Web IG <semantic-web@w3.org>, "team-rdf-chairs@w3.org WG" <team-rdf-chairs@w3.org>
On 28 November 2013 14:27, Phil Archer <phila@w3.org> wrote:
> Dear all,
>
> An idea has been floated and I'd like to assess the community's reaction.
> The rdf and rdfs namespaces are hard to remember (I always copy and paste, I
> guess you do too), but how do you react to the idea of deprecating those
> namespaces in favour of the much easier to remember
> http://www.w3.org/ns/rdf|s ?
>
> For emphasis, there would be *no change* at all to the semantics of any
> term, but the existing semantics might be more clearly explained.
>
> For:
> ====
>
> 1. In addition to replicating the schemas at that namespace, more detailed
> usage notes could be added;

We can do that anyway. I don't know any rule that says
http://www.w3.org/1999/02/22-rdf-syntax-ns and
http://www.w3.org/2000/01/rdf-schema can't be updated. And if they
can't, then we'll be stuck with the old version floating around
confusingly. Why not upgrade both to use RDFa, tweak the wording and
add translations and an update policy?

> 2. Multilingual labels, comments and usage notes could easily be added (this
> is something I'm really keen to promote);

You've spread the same benefit into two points ;) improved comments
would be great. But we'd need to set expectations about if/when the
new files could be updated.

What we tried to do with the rdfs: namespace was to make a way for
updated comments/labels to be discovered, semi-linked-data style, i.e.
the ns doc has this at the end:

<rdf:Description rdf:about="http://www.w3.org/2000/01/rdf-schema#">
  <rdfs:seeAlso rdf:resource="http://www.w3.org/2000/01/rdf-schema-more"/>
</rdf:Description>

... which has a pointer to a French translation (of rdf: rdfs: text),
http://www.w3.org/2000/01/rdf-schema-more ->
http://www.w3.org/2000/01/combined-ns-translation.rdf.fr

The RDFS ns doc also has this in it, which suggests W3C feels free to
make changes without a live RDF WG:

 <rdfs:Datatype
rdf:about="http://www.w3.org/1999/02/22-rdf-syntax-ns#PlainLiteral">
   <rdfs:subClassOf
rdf:resource="http://www.w3.org/2000/01/rdf-schema#Literal"/>
   <rdfs:isDefinedBy rdf:resource="http://www.w3.org/TR/rdf-plain-literal/"/>
   <rdfs:label>PlainLiteral</rdfs:label>
   <rdfs:comment>The class of plain (i.e. untyped) literal
values.</rdfs:comment>
</rdfs:Datatype>
<!-- ... until here -->

> 3. You'd be able to remember the namespace.

I wouldn't bet on it :)

A lot of people write 'rdfs:Property' or 'rdf:Class'.

Is it rdf:datatype or rdfs:Datatype or both?

What about rdf:resource vs rdfs:Resource? (these are only slightly
trick questions... we have used these namespacing mechanisms for both
syntactic and vocabulary constructs)

Why not merge both while you're at it? And perhaps throw in a few of
the more useful bits of OWL? How am I supposed to remember whether to
use owl:Class or rdfs:Class (or rdf:Class)?

> Against
> =======
> 1. Everyone just copies and pastes and loads of tools have the namespaces built in so it's pointless.

To some extent they're built-in to various languages.

RDF/XML has: <Potato/> as shorthand for _:bnode1
<http://www.w3.org/1999/01/rdf-syntax-ns#type>
<http://example.com/Potato>

Turtle and SPARL have the "a" mechanism, e.g. "Foo a bar", which also
involves a shortcut for that url.

RDFa 1.1 has a whole package of namespace short names built-in,
including these. http://www.w3.org/2011/rdfa-context/rdfa-1.1 ...
would we ask RDFa parser writers to update?

> 2. Any copy or derivative work might cause confusion.
> 3. One person's clarity is another person's confusion, meaning that the
> promise of not changing the semantics might be hard to keep in some people's
> minds.

It is very hard to say what counts as a mere clarification, rather
than a change. Consider e.g. the rdfs:comment here:

<rdf:List rdf:about="http://www.w3.org/1999/02/22-rdf-syntax-ns#nil">
  <rdfs:isDefinedBy rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#"/>
  <rdfs:label>nil</rdfs:label>
  <rdfs:comment>The empty list, with no items in it. If the rest of a
list is nil then the list has no more items in it.</rdfs:comment>
</rdf:List>

> How it might happen
> ===================
> *IF* there is community desire for this then I would suggest that a
> Community Group be formed to take it on. Any publication of the schema in
> /ns space would have to make clear that the relevant standards remain
> untouched and normative so that if any errors are seen, then the /TR doc is
> the one to choose.
>
> Good idea?

It's a very good idea for W3C to consider the developer/publisher
usability of its standards. Unfortunately in this particular case,
that horse has bolted, met other horses in the forest and made many
baby horses. Some of which have themselves reached RECommendation
status.

> Stupid idea?

Not at all stupid. It is worth thinking these things through in
public. It is certainty tempting to consider, and has been discussed
occasionally here and nearby.

> Great, count me in for the Community group?
> You are a moron, please don't ever suggest anything like that ever again?
>
> If your answer is negative then I hereby deny all association :-) I'm just
> making a public version of something said to me in private.

http://xkcd.com/927/ :)

Here's a counter proposal: a) merge the machine-readable schemas from
both rdf: and rdfs: namespace docs b) encode them in modern RDFa c)
publish them with a clear statement of intent about when and how W3C
or its nominated successor(s) can fiddle with them again. We could
also publish it in W3C's git[hub] repo to encourage translation.

Dan


> Thanks
>
> Phil.
>
> --
>
> Phil Archer
> W3C Data Activity Lead (TBC)
>
> http://philarcher.org
> +44 (0)7887 767755
> @philarcher1
>
Received on Thursday, 28 November 2013 15:57:22 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:42:46 UTC