W3C home > Mailing lists > Public > public-swd-wg@w3.org > April 2008

[VM] Review of 16 March Editor's Draft

From: Mark van Assem <mark@cs.vu.nl>
Date: Mon, 14 Apr 2008 16:35:24 +0200
Message-ID: <48036BAC.208@cs.vu.nl>
To: SWD WG <public-swd-wg@w3.org>
CC: Guus Schreiber <guus@few.vu.nl>


Guus Schreiber asked me to review the "Principles of Good Practice for 
Managing RDF Vocabularies and OWL Ontologies", Editor's draft 16 March 
2008 [1].

General comments

In general I'm very happy to see this document coming out, agree on 
the set of practices it proposes, and learned some things I didn't 
know yet. Most of my specific comments are on structure and language.

There are still some @@TODO's that need to be filled in, especially 
the example maintenance policies in Sec 2.3. Good examples are really 
the heart of this type of document, so I recommend to complete these 
before publishing as WD.

Specific comments


- recommend to change "citation" to "reference"
- this document is not only focused on "re-use" but also on initial 
"use", so recommend "(re-)use)"

- I myself am not familiar with the term "principles of good 
practice". To me the word "best practice" (like used in the Recipes 
document) or "principles" seems more natural. (Ignore this if this is 
a typical comment of a non-native speaker ;)

- the use of the word "user" in the whole document is ambiguous to me, 
because it might mean "end user". I assume "developer that uses the 
vocabulary" is meant, so then e.g. "developer" could be reserved for 
"user of the vocabulary" and e.g. "publisher" for those who make it 

- sentences "Further ... methodologies/approaches" is unclear to me.

- what is the intended readership of this document? Specifically, does 
it include people well-versed in vocabulary development but not 
familiar with SW-technology?


- the word "strategies" is used which seems to refer to "principles of 
good practice"

- the relationship between "vocabulary" and "ontology" is unclear. 
Depending on the intended readership it might be a good idea to leave 
the whole notion of "ontology" out.

1. Introduction

- "vocabulary" seems to be presented as equivalent to *schema* 
(examples given are SKOS, DC, FOAF, OWL). However, in other 
communities the word "vocabulary" is associated with the actual 
concept schemes (e.g. AAT, LCSH, MeSH). If a distinction between the 
two is to be maintained and the intended readership includes non-SW 
people, I recommend to explain the difference.


It would improve readability to divide each subparagraph of Sec.2 into 
sub-elements (in layout), e.g. "Principle: ...", "Motivation: ...", 
"Examples: ...", "Notes: ...".


- This section starts with five bullet points. Recommend to address 
each bullit point in the succeeding text in turn. Currently the text 
directly after the list addresses an issue not in the list: 
dependencies between vocabularies.

- recommend to provide example of "good" and "bad" URIs
- what is a "subordinate URI scheme"?


- remove "*" around words and make the words bold/italic
- the WordNet example seems to cite "a plugin for Protege, an RDF 
version, an ontology and ..." as documentation of WordNet, which seems 
not intuitive to me
- the paragraphs about collaborative development are loosely connected 
to the previous paragraphs. Recommend to provide subheading to 
separate them, e.g. "Collaborative Development"

The principle seems to distinguish two kinds of documentation: usage 
documentation and change documentation; this could be made explicit.


The links to examples of vocabularies "and other artifacts" (?) that 
have a published maintenance policy do not point to information on 
maintenance policies, but about the vocabularies themselves.

Secs. 2.3.1 - 2.3.3 have open TODOs.


- recommend to emphasize that not using unique version URIs causes URI 

- par.5 distinguished two approaches: (1) parallel editing and then 
publishing one final version (2) releasing different versions, with 
the "same" elements at different URIs. It is not clear to me if the 
first approach proposes publication at the same or different URIs. I 
also don't get why the first approach is "preferable when a vocabulary 
... makes significant portions of the older version obsolete", and 
second approach "maintains compatibility across the different versions".

- the papers listed at the end could be moved to Sec.4  ("Additional 


- The word "Formal" in the section title may be confusing and doesn't 
seem needed to cover the contents of the section
- the section should explain some of the motivation of publishing 
content at the vocabulary URIs
- The section is currently relatively short as it refers to the 
Recipes for further explanation. However, it could provide some more 
"preview" to what is in the Recipes.
- provide an example (a clickable URI) that underlines what can be 
published as a resource description


I agree with Tom Baker [2] that it is preferable not to emphasize open 
research. Moreover, the current text primarily gives attention to one 
issue (automated version reasoning) that is not directly related to 
one of the principles, and has a large focus on one possible 
implementation (C-OWL).


Recommend to provide names of the documents instead of their URIs and 
explain/classify them (e.g. to what principle they are relevant)

With regards,
Mark van Assem.

  Mark F.J. van Assem - Vrije Universiteit Amsterdam
Received on Monday, 14 April 2008 14:37:23 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:07:53 UTC