RE: Working on FPWD, more to do

I also like Steve’s approach, but it brings me back to an earlier question: What is *best* practice?

 

In a way, a maturity model describes what is good, better, best practice as you move up the ladder. But how does someone (us in this case) determine what is good, better, best?

 

As far as I can see, we try to define best practice based on our personal opinions – of course backed by our individual and collective knowledge and experience – but we don’t seem to consider any type of metrics or arguments that justify why something is better practice than something else.

 

I posed that question earlier on BP#1 http://w3c.github.io/dwbp/bp.html#metadata. I think that a statement like “in an open information space, metadata is essential” is an opinion, but one that needs to be qualified, especially because you could argue that in the current Web environment this has been demonstrated *not* to be true. Data can be discovered and re-used even without metadata as long as it is harvested by a search engine; actually, in the current environment of the open Web, a landing page with good SEO is probably a better way of creating high visibility than DCAT metadata.

 

On the other hand, if you want to build a catalogue of datasets like  http://datahub.io/, or want your datasets to be listed on such a portal, then of course metadata is the way to go to enable harvesting.

 

So, thinking further on Steve’s maturity model, we could have levels like:

 

Put your data on the Web and 

 

0.       Do not provide any information about your data. If you don’t, your data can only be found by people who know about it, so you don’t encourage wide re-use – NOT SO GOOD (but of course, someone might have good reasons to keep their data out of the spotlight)

1.       Provide a landing page. This allows the information to be picked up by search engines. If you’re doing some smart SEO in addition, it will make your data will make it visible, facilitating more re-use – BETTER

2.       Provide metadata that describes the data. This may increase visibility on search engines (e.g. using schema.org) but it is really essential if you want your data to be visible on portals like the DataHub; these portal services require metadata to be available for harvesting – BETTER

3.       Provide both a landing page and standardised metadata: this makes your data visible through search engines and allows your data to be included in data portals which maximises visibility and re-use – BEST 

 

Such a ladder gives advice on what to do and why: what happens if you do and what happens if you don’t. 

 

In that way, we don’t tell people what they MUST or SHOULD do, we provide advice that they can follow or not, depending on their objectives, resources etc.

 

Makx.

 

 

 

 

From: Steven Adler [mailto:adler1@us.ibm.com] 
Sent: Wednesday, February 04, 2015 9:18 PM
To: Eric
Cc: Annette Greiner; Bernadette Farias Lóscio; Phil Archer; Public DWBP WG
Subject: Re: Working on FPWD, more to do

 

I feel a little nervous about weighing in here but here goes.  I am OK with removing normative statements in this version of the BP document and I appreciate the desire to describe rather than prescribe practices.  But I also feel that we need to get more specific about our descriptions in future versions of the document.  An approach we can take in that regards is to develop our descriptions in a Maturity Model framework, which plots different levels of observed behaviors across increasing levels of maturity, allow the readers to discover for themselves how their own practices compare to other levels of maturity and decide where they are and what they want to achieve.