Re: Additional review and telecon (Tuesday) for Best Practices

In addition to Dave and Makx's comments (which I agree with pretty much 
throughout), I'd like to add the following.

Section 5, Metadata. More could be said there, for example, a link to 
DCAT and, as an example of how to use it, the EC's DCAT Application 
Profile 
(https://joinup.ec.europa.eu/asset/dcat_application_profile/asset_release/dcat-application-profile-data-portals-europe-final) 
written by Makx.

7. Licences

The problem here is not that it's too American - there's useful info 
there - but that it's really only American. The line about Crown 
Copyright might lead to a reference to the UK Open Government Licence 
(http://nationalarchives.gov.uk/doc/open-government-licence/version/2/) 
for example. Many governments are in the process of working out how to 
licence their data (there are whole EU projects about it such as 
http://lapsi-project.eu/ but sadly that site is not guaranteed to be 
persistent so I'm hesitant to suggest linking to it.

8. Convert Data to Linked Data

The sentence "This involves a data modeling step..." should link back to 
the modelling section.

The sentence that ends "...mapping the source data into a set of RDF 
statements via a script" suggests, or might suggest to some, that there 
is 'a script' - i.e. a single tool for doing this. I would say something 
like 'programmatically' or 'through an automated process', 'using one of 
a variety of tools' etc.

I think sections 10 & 11 should also be merged. URI construction is an 
important part of designing for persistence.

That'll do for now - we can talk more on the imminent call.

HTH

Phil.


On 17/12/2013 08:33, Dave Reynolds wrote:
> Hi Bernadette,
>
> On 17/12/13 04:53, Bernadette Hyland wrote:
>> Hi,
>> Thanks for the review you did do.  See inline below.
>>
>> On Dec 15, 2013, at 5:04 PM, Dave Reynolds <Dave.e.Reynolds@gmail.com>
>> wrote:
>
>>> # 4 Data Modelling
>>>
>>> Not sure what value this section adds, it says too little. Should
>>> either say more or say that modelling advice is out of scope.
>>
>> Will beef up with a bit more description to make it clear the modeling
>> step is really important.   In the real world we face people who heard
>> there is a 'convert to RDF' tool and they cannot distinguish between
>> auto-magic conversion vs. well-modeled content that uses standard
>> vocabs & is expressed as something others can re-use.
>
> I'm not sure beefing this up is the way to go.
>
> Fully agree that the modelling step is really important. But it's not
> clear to me that we could reach consensus on best practices for
> modelling within the short time left. We already have sections on
> ontology selection and ontology creation which are arguably the major
> part of modelling. [Though our advice on creation doesn't actually tell
> you anything about how to develop one.]
>
> My suggestion would be to say words to the effect of:
>     * Modelling is really important.
>     * Advice on detailed modelling is out of scope for this document.
>     * A key part of modelling is selecting or developing appropriate
> ontologies, some advice on this is contained in the following sections ...
>
> Cheers,
> Dave
>
>

-- 


Phil Archer
W3C Data Activity Lead
http://www.w3.org/2013/data/

http://philarcher.org
+44 (0)7887 767755
@philarcher1

Received on Tuesday, 17 December 2013 14:40:42 UTC