W3C home > Mailing lists > Public > public-dwbp-wg@w3.org > March 2016

Re: Updates and suggestions to BP17 Reuse vocabularies

From: Antoine Isaac <aisaac@few.vu.nl>
Date: Sat, 12 Mar 2016 11:00:25 +0100
Message-ID: <56E3E8B9.1060003@few.vu.nl>
To: <public-dwbp-wg@w3.org>
Hi Annette,

BP16 is purely about standardization of terms (just using words or codes being used elsewhere). BP17 is more about re-using artefacts already built, which can be re-used to express knowledge as such (XMLSchema, OWL ontologies, SKOS concept schemes).
There's also a difference of consensus: 'standardized' is quite formal, top-down, 'shared' is more bottom-up.

We already had the discussion at this issue:
https://www.w3.org/2013/dwbp/track/issues/166

Frankly at the beginning there was just one: "Best Practice 18: Re-use vocabularies"
https://www.w3.org/TR/2015/WD-dwbp-20150224/#dataVocabularies

Then a new one ("Best Practice 15: Use standardized terms") was introduced in front of all others, without caring too much about the others:
https://www.w3.org/TR/2015/WD-dwbp-20150625/#MetadataStandardized

I seem to remember some in the group dearly wanted to see the words 'standardized' and 'code lists' flashing at the top. And didn't like my suggestion to merge the two BPs:
https://lists.w3.org/Archives/Public/public-dwbp-wg/2015Aug/0083.html

These were the final resolution:
https://lists.w3.org/Archives/Public/public-dwbp-wg/2015Sep/0027.html
It was even worse because the resolution then forced be to introduce the word 'term' in the BP about re-using vocabularies (as engineered artefacts) creating confusing with the BP that was recommending to use anything standardized.
I've tried to express the distinction as I could, because I also felt it was not clear. But one cannot turn lead into gold, it seems.

Cheers,

Antoine

On 3/11/16 8:10 PM, Annette Greiner wrote:
> I think the real problem here is that BPs 16 and 17 are really saying the same thing in slightly different ways. It seems to me that at one point we had them as separate ideas, and I think maybe one was supposed to be about being internally consistent in your naming of things, and the other was about using standard vocabularies, so being externally consistent, but they seem to have wandered together over time. I wonder if someone more familiar with these two BPs (Antoine?) could take a look and tease them apart, or combine them into one. I find it odd that we have two BPs to handle a subtle difference in ways of reusing vocabularies, but one of them also extends as far as to cover a shared data model. If one can be that general, we don't really need both at all, IMHO.
> -Annette
>
> On 3/11/16 6:33 AM, Antoine Isaac wrote:
>> Hi Bernadette,
>>
>> Thanks for the feedback!
>> OK I will submit a proposal.
>> Maybe directly as a pull request.
>>
>> Antoine
>>
>> On 3/11/16 4:52 AM, Bernadette Farias Lóscio wrote:
>>> Hi Antoine,
>>>
>>> Thanks for your message! I reviewed  BP17: Reuse Vocabularies and I agree with you that the two outcomes that you mentioned are confused. Maybe, we can keep just the first one.
>>>
>>> Could you please help us to make a proposal for the intended outcome of BP 17?
>>>
>>> Feel free to use the constructions from your choice. It is just important to keep in mind that we should be able to test the BP.
>>>
>>> kind regards,
>>> Bernadette
>>>
>>> 2016-03-03 18:49 GMT-03:00 Antoine Isaac <aisaac@few.vu.nl <mailto:aisaac@few.vu.nl>>:
>>>
>>>     Hi everyone,
>>>
>>>     I've done my action on suggesting examples for BP17 "Reuse vocabularies" [1]
>>>     https://github.com/w3c/dwbp/pull/307
>>>
>>>     In the process I became stuck again with the intended outcomes. It had already flagged it some time ago [2]. At the time the discussion had focused on the editorial points. But now it's really about whether these intended outcomes should be in this BP or elsewhere, or actually whether they make sense at all!
>>>
>>>     1. I'm really not sure whether these two  outcomes should be specific to BP17"Reuse vocabularies":
>>>     [
>>>     It should be possible for machines to automatically process the data within a dataset.
>>>     It should be possible for machines to automatically process the metadata that describes a dataset.
>>>     ]
>>>     I.e. for me these are more intended outcomes of machine-readable data and metadata in general not specific to reusing vocabularies. In fact it we think they make sense for BP17 then I think we should add them to BP16 "Use standardized terms' and many other BPs. Standardized lists of codes and terms also help machines to automatically process data.
>>>
>>>     2. The first intended outcome look more specific to vocabularies:
>>>     [
>>>     It should be possible to automatically compare two or more datasets when they use the same vocabulary to describe metadata.
>>>     ]
>>>     But I also think it should be both in BP16 and BP17... And this intended outcome is confusingly written for me:
>>>     1. When two datasets use the same vocabulary, it just *is* possible to compare them. This is much stronger than what the sentence 'it should be possible to compare them' hints at. This reads poorly.
>>>     2. This sentence alludes to a situation where 'datasets use the same vocabulary to describe metadata'. Datasets here describe metadata? Like, datasets of meta-metadata? This exists, but I'm fairly sure this is not what was meant. Couldn't we just simplify and remove ' to describe metadata'?
>>>
>>>     By the way I noticed that now a lot of intended outcome don't start with 'it should be possible' anymore. If it's not mandatory, I'd like very much to get read of this construction in the vocabulary best practices.
>>>
>>>     Best,
>>>
>>>     Antoine
>>>
>>>     [1] http://w3c.github.io/dwbp/bp.html#dataVocabularies
>>>     [2] https://www.w3.org/2013/dwbp/track/issues/211
>>>
>>>
>>>
>>>
>>> --
>>> Bernadette Farias Lóscio
>>> Centro de Informática
>>> Universidade Federal de Pernambuco - UFPE, Brazil
>>> ----------------------------------------------------------------------------
>>
>
Received on Saturday, 12 March 2016 10:00:58 UTC

This archive was generated by hypermail 2.3.1 : Saturday, 12 March 2016 10:00:59 UTC