W3C home > Mailing lists > Public > public-dwbp-wg@w3.org > August 2014

Re: dwbp-ISSUE-29: Clarify requirements on metadata

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Fri, 29 Aug 2014 14:58:35 +0200
Message-ID: <540078FB.10803@few.vu.nl>
To: "Lee, Deirdre" <Deirdre.Lee@deri.org>, "public-dwbp-wg@w3.org" <public-dwbp-wg@w3.org>
Hi,

Thanks for putting everything in context, Deirdre!

My take on this:

1. "available" can always be folded in other requirements. There's no way in bothering that metadata is machine-readable if it's not available in the first place. So we could have e.g.
'metadata should be machine-readable'
formulated as
'metadata should be made available in a machine-readable way'

2. Relevant metadata includes license, provenance and SLAs. So we could fold license, provenance and SLA as three requirements:
R-License: licenses should be made available as part of metadata
R-Provenance: provenance data should be made available as part of metadata
R-SLA: SLAs should be made available as part of metadata


3. Standardized implies interoperable, as per my position in the other thread.

If we follow this, so we'd end up with much fewer requirements:

R-FormatOpen
R-FormatStandardized
R-FormatMachineReadable

R-MetadataStandardized
R-MetadataDocumented (in principle 'standardized' implies 'documented' but I reckon that metadata may not always be standardized, in which case specific documentation should be made)
R-MetadataMachineReadable (btw I think it's not really productive to abbreviate names of requirements)
R-License
R-Provenance
R-SLA

R-VocabulariesOpen
R-VocabulariesDocumented

That's 11 instead of18 in the table.


This would even allows to fill the blanks in the grid, notably by adding:
R-MetadataOpen. There's not much point in closing data that allows access to the datasets. Even if the datasets are actually closed, it's absurd to to close the metadata that indicates it ;-)

I suspect we could add more requirements on vocabularies (standardized, machine-readable) but if these have not been identified in the use cases it's less obvious.

Best,

Antoine

On 8/29/14 1:41 PM, Lee, Deirdre wrote:
> A lot of the issues relate to how the requirement concepts (/meta-data, formats, vocabularies, license, provenance, SLA/) and requirement types (/open, standardised, machine-readable, interoperable, documented, available/).
>
> So I created the attached matrix to show current requirement definition (R represents the existance of requirement). I think agreeing on which of these requirements to keep/delete/add will resolve many issues.
>
> Can you have a quick look before today’s call and we can discuss this this afternoon.
>
> Cheers,
>
> Deirdre
>
> -----Original Message-----
> From: Antoine Isaac [mailto:aisaac@few.vu.nl]
> Sent: 22 August 2014 12:38
> To: public-dwbp-wg@w3.org
> Subject: Re: dwbp-ISSUE-29: Clarify requirements on metadata
>
> Hi,
>
> On 8/4/14 4:49 PM, Lee, Deirdre wrote:> If we define metadata as being 'data on the web', then all of the general requirements would also apply to metadata.
>
>  > The only two metadata-specific requirements would be R-MetadataAvailable and R-MetadataInteroperable
>
> Depending on the recent discussions, it may be R-MetadataAvailable and R-MetadataStandardized ;-)
>
> But if we have requirements formulated at a general MD level, could we remove the 'corresponding' ones at the specific level? At least for the requirement on Interoperable/Standardized?
>
> I.e. if we have R-MetadataStandardized we don't need R-LicenseStandardized, because License is a kind of MD.
>
> Best,
>
> Antoine
>
>  >
>
>  > -----Original Message-----
>
>  > From: Data on the Web Best Practices Working Group Issue Tracker [mailto:sysbot+tracker@w3.org]
>
>  > Sent: 03 June 2014 23:25
>
>  > To: public-dwbp-wg@w3.org <mailto:public-dwbp-wg@w3.org>
>
>  > Subject: dwbp-ISSUE-29: Clarify requirements on metadata
>
>  >
>
>  > dwbp-ISSUE-29: Clarify requirements on metadata
>
>  >
>
>  > http://www.w3.org/2013/dwbp/track/issues/29
>
>  >
>
>  > Raised by: Antoine Isaac
>
>  > On product:
>
>  >
>
>  > The requirement on 'metadata' are not clear. To me license, provenance are also a form of metadata; perhaps even vocabularies: once we have spelled out these more specific levels, it's harder to figure out what remains in the general 'metadata'.
>
>  >
>
>  >
>
>  >
>
Received on Friday, 29 August 2014 12:59:05 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:24:16 UTC