[epsi-coord] Response of EC to standardization topic report. Fwd: Topic report on open data and standardization

[Ton Zijlstra]

hi Tom, Hans,

See below the feedback of the Commission on the TR on standardization.

Can you guys prepare a revised version? Thanks!

best,
Ton
---------------------------------------------------------------------
Interdependent Thoughts
Ton Zijlstra

ton at tonzijlstra.eu
+31-6-34489360

http://zylstra.org/blog

*Share your real life open data experiences,*
*observations and anecdotes:*
http://epsiplatform.eu/content/share-your-story
---------------------------------------------------------------------



---------- Forwarded message ----------
From: <Katalin.IMREI at ec.europa.eu>
Date: Wed, Nov 28, 2012 at 4:06 PM
Subject: RE: Topic report on open data and standardization
To: ton.zijlstra at gmail.com
Cc: Szymon.LEWANDOWSKI at ec.europa.eu


 Dear Ton,****

** **

Thank you for this draft topic report on open data and standardisation.****

We have carefully read through your document and also asked our colleague
who is an expert in standardisation issues to make comments.****

** **

Please find our comments below my current message.****

We are looking forward to receiving the revised version of the report once
our comments are integrated in it.****

** **

Thank you again.****

Best regards,****

Katalin****

** **

A general comment: the interesting question of whether to pursue the
strategy of opening up as much raw data as possible (regardless of
standards) or of publishing only standardised data does not seem to be
sufficiently elaborated on in the document. Maybe some further efforts
could be made to address this question in more details. ****

** **

As regards the technical aspects touched upon in the report, there seem to
be some elements that might be worth having a second look at: ****

** **

Page 4:****

*On the other hand, increasing emphasis is being put by the open data
community on the importance of linked data. An essential element of such
linked data is that they are standardised, preferably according to open
standards. However, public bodies often create datasets solely with the
view of meeting their own specific needs, without considering the possible
use of the data for others (as do private companies), and they are often
dependent on proprietary software that is in many cases even specifically
created for them by software and management system vendors. Hence, making
available government data in a standardised form takes time, effort and
money, and, as a consequence, conflicts with providing ‘raw data now’. *

Here, it might be worth referring to the RDF standard for data
representation that the concept of Linked Data is based on (and which can
indeed be considered as an open standard).****

When referring to the proprietary software, it could be made clearer how
this impacts on the data and interoperability, i.e. by the fact that
proprietary software often implies the usage of proprietary data formats
that make interoperability difficult.****

** **

Section 3.5: The Semantic Web****

Here, it might be worth mentioning that the Semantic Web standards are also
the basis for the Linked (Open) Data approach. ****

** **

Page 9:****

*Some file standards are closely tied to the type of data that is being
published. To a data owner, it will be immediately clear whether or not
these standards are suitable to be used in or for a specific dataset.
Examples include Google’s KML format for geocodes, .MP3 and other audio
files for sound data, etc. Even though the difference with the ‘regular’
file formats is somewhat academic, the main distinguishing factor with
these file types is the presence of a ‘context’. E.g. music data might be
expressed (even though not easily played) as a spreadsheet, but a
spreadsheet will not be stored as an MP3-file.*

Here, it seems to me as though two distinct issues were mixed up: while KML
is a data markup language (i.e. a metadata standard like Dublin Core
mentioned in the next paragraph), Mp3 is a data compression standard.****

** **

Page 9:****

*Other standards relate to the representation of the attributes and
characteristics of the data. For instance, the Dublin Core standards are
part of the Resource Description Framework of XML[1]<#13b478e0afc794bc__ftn1>.
*

Dublin Core is not part of RDF but it is a metadata standard that can be
represented in RDF or XML.****

** **

Page 9:****

*The file format for linked data files is usually RDF (for Resource
Description Framework), which in itself is a file-format standard.
RDF[2]<#13b478e0afc794bc__ftn2>allows the publisher to link to the
semantic web, thus supplying a context
for the data. *

Rather than being a file format, RDF is a data representation standard
which can be serialised in different file formats.****

** **

Page 9:****

*This is due to the fact that the RDF contains links (uniform resource
identifiers or URI’s) to entities or relations that are stored within
ontology-databases such as DBPedia.
<http://opendatastandards.org/?t=faqs&faq=what-is-a-standard>*