Re: [BP - MET] - Metadata quality

Hi Laufer,

This 'meta-meta' level is not out-of-scope. In fact it's precisely what is identified in the requirements. And the vocabularies (especially the quality and granularity one) are precisely about this. By creating a vocabulary, we aim to promote its usage, assuming that this will result in higher-quality metadata (a voc like Q&G will be used to create metadata of course).

So yes a lot of what you describe makes sense: it would be useful to focus on the quality of metadata, next to the current 'star schemes' that rather focus now on the technical way the data is published. This echoes with what others in the LOD community are also proposing either at the metadata quality level [1, 2] or the legal one [3]. (I guess the ODI also has several things like this).

What worries me is the potential overlap with other streams of work in the WG. Your 'metadata' includes many things, including licenses and provenance (and I agree with it, see my comments on the UC document [4])
I expect "quality/metrics" to be handled by the Q&G voc. 'License' may also happen as a separate line of work, if the requirements become stronger than expected (as discussed at [5])
It's perhaps good to have some general discussion on this before we all dive further in our own areas...


Meanwhile, I'm going to add a couple of pointers to the wiki: but that's just starting points to investigate further. In fact on provenance I have already sent something on the list [6]. Is it relevant for you?

Best,

Antoine

[1] http://www.seco.tkk.fi/publications/submitted/hyvonen-et-al-ldf-2014.pdf
[2] http://bvatant.blogspot.fr/2012/02/is-your-linked-data-vocabulary-5-star_9588.html
[3] http://lod-lam.net/summit/2011/06/06/proposed-a-4-star-classification-scheme-for-linked-open-cultural-metadata/
[4] comment 7, http://lists.w3.org/Archives/Public/public-dwbp-wg/2014May/0115.html
[5] http://lists.w3.org/Archives/Public/public-dwbp-wg/2014May/0031.html
[6] http://lists.w3.org/Archives/Public/public-dwbp-wg/2014May/0027.html

On 5/29/14 3:55 PM, Laufer wrote:
> Dear All,
>
> In our last teleconference, I commented about the idea of metadata quality. I know it may sound like something meta-meta but I will try to better explain what I am thinking.
>
> As we can see from the use cases we have collected, there are a lot of different ways in which people are publishing data on the web. It seems that we have a North in the linked data idea but many of the use cases in our list are still collections of files published in data portals.
>
> In these portals we have data and metadata. The 5 stars open data scale makes, in some sense, people think that data that have 5 stars are better described than data with, for example, 3 stars. I don’t think that this is necessarily true. Some of these data portals have a kind of best practices that ask users to publish metadata together with data, and suggest, or require, metadata content. Some portals have a fixed set of fields that must be provided by the user. They are mandatory. So, one dataset with 3 stars could have a metadata set that helps a developer in a better way than another one with 5 stars. The 5 stars don’t guarantee the metadata quality. The metadata quality scale is not the same as the 5 stars open data scale. I am not proposing to have a metadata quality scale. But I think it exists.
>
> In the Guidance on The Provision of Metadata <http://www.w3.org/2013/dwbp/wiki/Guidance_on_the_Provision_of_Metadata> page I listed a set of metadata types that is already being investigated by diverse groups and could help to enhance the metadata quality of a dataset. For each one of the metadata types we could give pointers to the groups that are working in that specific topic and some examples using our use cases. Some of these metadata types, related to data quality and data usage, for example, are in the scope of our group. In my opinion, it is out of the scope of our group to formally define each one of these metadata types. Each one of them deserves an individual WG taking care of it. And many already have these groups.
>
> As I asked in the last teleconference, I would appreciate that DWBP WG participants, which have information about initiatives working in each metadata type, could put content and links in the wiki page, in a way that other participants could study and put more detailed information about these works, with comments and examples.
>
> I think this could be a good guide for users, to help them to publish data, to choose data publishing tools, data portals, and also helping the developers of the data publishing tools to include ways of enhancing the metadata quality of datasets.
>
> As Bernadette pointed in her article, we have a set of different roles in the Data on the Web ecosystem. People with different roles may benefit from our guide, each one doing his part to enhance the overall ease and efficiency of the ecosystem usage and the effects over society.
>
> Thanks.
>
> Best regards,
>
> Laufer
>
>
> --
> .  .  .  .. .  .
> .        .   . ..
> .     ..       .

Received on Friday, 30 May 2014 09:18:48 UTC