CKAN/our vocabs

For archive

This is an e-mail exchange between me and Open Knowledge Foundation's 
CKAN Product Manager, Irina Bolychevsky, that I hope will be of interest 
to the WG.


-------- Original Message --------
Subject: Re: What can I say?
Date: Wed, 02 Apr 2014 10:28:26 +0100
From: Phil Archer <phila@w3.org>
To: Irina Bolychevsky <irina.bolychevsky@okfn.org>

Thanks very much Ira, this is very helpful.

At our meeting this week I was emphasising that we need to find capacity 
to contribute code to CKAN, exactly as you say. There are realistic 
possibilities of doing this that I'm hoping to firm up.

Background: we're working on what might be an extension of DCAT to cover 
things like provenance info, commitments, quality assessment, feedback 
and data usage. These are currently foreseen as being an extension to 
DCAT and a separate one for data usage but after yesterday's discussions 
I think it might all just be an extension to DCAT. There's also 
semi-separate work on expressing the output of computer algorithms to 
assess data quality. That won't be part of DCAT but it's related and 
amounts to feedback or assertions of quality depending who does the 
calculation.

That latter work is tied to a promise of CKAN extension development work 
and the DCAT extension stuff is tied to (at least) development work in 
Brazil and Spain - all subject to confirmation but everyone understands 
the need to write code. Actually, getting people to agree to write CKAN 
extensions is relatively easy. It's getting them to do the same for 
other portals that's harder (and we need two independent implementations 
of any standard to satisfy W3C process).

Add to that the likely near future launch of a new WG on application 
profiles and, again, I'm hopeful that WG will have capacity to create a 
CKAN extension that implements the ability to validate a dataset's 
metadata against, for example, a machine-readable version of the DCAT-AP.

So I hope we're working the same direction...

Thanks again

Phil.

On 01/04/2014 14:41, Irina Bolychevsky wrote:
> Hey Phil,
>
> In terms of DCAT implementation, there are several things:
> 1. There is a ckanext-rdf extension that was used on datahub.io and I think
> is still functional (maybe could be improved)
> 2. When we first implemented DCAT support, we just went with the lowest
> common denominator and came up with a very simple DCAT rendering in core
> since there was no clear consensus
> 3. Recently (and since DCAT has finally been accepted & stabilised) we have
> started work on github.com/ckan/ckanext-dcat for proper support and
> interoperability work. This is a little slow since we don't have specific
> client requirements for this.
>
> I am surprised by some of the comments since as far as I know there has
> been zero contribution from the linked data community around DCAT & linked
> data support for CKAN. CKAN is an open source project - not only would it
> welcome contribution, but we also have twice weekly dev meetings that
> people are welcome to join to review pull requests and discuss technical
> decisions and architecture plans with the core committers of CKAN. Is there
> anyone in particular who would like to work on this? Or does anyone have a
> spec for what kind of DCAT support they want to see?
>
> Hope this helps,
> Ira
>
>
> On 30 March 2014 18:40, Phil Archer <phila@w3.org> wrote:
>
>> Hi Ira,
>>
>> I've just checked in for 2 nights on the South Bank as we have a face to
>> face meeting of the Data on the Web Best Practices WG. One of our e-mail
>> threads [1] throws up the question I get asked a lot and one we've
>> discussed before. It comes in different forms but basically it's: why is
>> the CKAN implementation of DCAT so poor.
>>
>> Rather than me and/or others making unfounded statements, can I ask you
>> for something quotable about CKAN's use/non-use of DCAT now and in future?
>> For example, do you see the new extension that implements the EC's DCAT-AP
>> becoming part of the canonical package?
>>
>> The WG's charter includes provision to extend DCAT to cover areas like
>> publisher commitment, quality, persistence etc. so we're trying to get a
>> handle on possible implementations, or rather, what we have to do.
>>
>>
>>
>> [1] http://lists.w3.org/Archives/Public/public-dwbp-wg/2014Mar/0239.html
>>
>> --
>>
>>
>> Phil Archer
>> W3C Data Activity Lead
>> http://www.w3.org/2013/data/
>>
>> http://philarcher.org
>> +44 (0)7887 767755
>> @philarcher1
>>

Received on Wednesday, 2 April 2014 09:33:46 UTC