Re: Comment on Section 4.8

Hi,
The motivation for this restriction to be the in the spec was that we 
(IBM) felt that this would 1) make it easier for people to ramp up using 
Linked Data by narrowing down the number of options they have to choose 
from , 2) increase interoperability.

It was for the same reason that we had limited the number of datatypes one 
could use in section 4.1.9 [5]. After discussion at the face to face 
meeting we (the WG) decided to remove section 4.1.9 for essentially the 
same reasons you're giving [6]. Rather than simply dumping it though, we 
discussed the idea of putting into a deployment guide, which has since 
then been done [7].

Given that, I think it would make sense for the WG to consider doing the 
same for section 4.8. [8]

[5] http://www.w3.org/TR/2012/WD-ldp-20121025/#ldpr-4_1_9
[6] http://www.w3.org/2012/ldp/meeting/2012-11-01#resolution_5
[7] http://www.w3.org/2012/ldp/wiki/Deployment_Guide
[8] http://www.w3.org/TR/2012/WD-ldp-20121025/#common-properties

Side note: when referencing a section in the spec it is better to use the 
URL that points to the specific version of the document you're refering 
to, such as [4] above. The URL you gave in [3] will actually stop working 
if that section is removed in a future version of the spec, leaving the 
mail archive with a broken link.

Thanks.
--
Arnaud  Le Hors - Software Standards Architect - IBM Software Group


leigh.dodds@gmail.com wrote on 12/07/2012 08:43:16 AM:

> From: Leigh Dodds <leigh@ldodds.com>
> To: public-ldp@w3.org, 
> Date: 12/07/2012 08:50 AM
> Subject: Comment on Section 4.8
> Sent by: leigh.dodds@gmail.com
> 
> Hi,
> 
> This email revisits some comments which I raised previously [1, 2]
> about the early working draft but which I don't feel have yet been
> adequately been discussed. I'm referring specifically to Section 4.8,
> Common Properties [3].
> 
> It is clearly useful if people re-use existing properties where they
> are applicable. It might also be useful to have some general guidance
> on which properties are considered to be currently "best practice" or
> are most widely deployed, etc.  However I don't think the LDP
> specification is the place to do this. For two reasons:
> 
> * This kind of guidance is best published by specific communities who
> are seeking convergence around their data. Not on a blanket basis by a
> W3C group. There needs to be space for these kinds of recommendations
> to evolve and be widely discussed
> * LDP is meant to be defining a platform for managing data. It should,
> as far as possible, be agnostic to what data is being stored inside
> it.
> 
> A good example of community specific guidance around property usage
> can be found in [4]. Personally I think thats more useful than a
> simple picklist of properties.
> 
> My other issue with the guidance as it stands is that it diverges from
> current practice. For example:
> 
> * rdfs:label is widely used as a default labelling property, its not
> just used in vocabularies. Discouraging its use suggests that usage is
> wrong; at the very least an alternative ought to be recommended
> * There are occasions when using the old Dublin Core terms may be more
> appropriate than the newer dcterms, see [4]
> * The LDP specification recommends use of rdfs:range that are at odds
> with their specification and common practice, e.g. dcterms:title and
> dcterms:description both have a range of rdfs:Literal.
> 
> I'd like to propose that Section 4.8 as it stands be removed and
> perhaps replaced with some informative text recommending re-use of
> vocabulary where it makes sense to do so.
> 
> Cheers,
> 
> L.
> 
> [1]. http://lists.w3.org/Archives/Public/public-ldp-wg/2012Jun/0013.html
> [2]. http://lists.w3.org/Archives/Public/public-ldp-wg/2012Jul/0026.html
> [3]. http://www.w3.org/TR/ldp/#common-properties
> [4]. http://aims.fao.org/lode/bd
> 
> -- 
> Leigh Dodds
> Freelance Technologist
> Open Data, Linked Data Geek
> t: @ldodds
> w: ldodds.com
> e: leigh@ldodds.com
>