W3C home > Mailing lists > Public > public-prov-comments@w3.org > March 2013

comments on PROV Primer and Overview

From: Bob DuCharme <bob@snee.com>
Date: Fri, 22 Mar 2013 10:26:18 -0400
Message-ID: <514C6A0A.5020502@snee.com>
To: public-prov-comments@w3.org
Great job. I knew nothing about PROV other than its general goals, so I 
was probably a good guinea pig to read the Primer. Because it said to 
start with the Overview, I did.

------ Notes on http://www.w3.org/TR/2013/WD-prov-overview-20130312/------

typos: defintions, Dublic (Spinal Tap reference: did you mean "Dubly"?), 
"these these", "that each document on" (that each document is on?),

Table in section 2: In the Document column, several sentences are 
missing periods at the end.

The PROV-DICTIONARY summary should have a few more words about why this 
document exists for the benefit of those reading this document as their 
very first PROV document, because the notion of collection hasn't been 
introduced yet.

PROV-LINKS entry on the table: same comment, but about bundles. (Section 
2 further on has a better short explanation of this document's purpose.)

------ Notes on http://www.w3.org/TR/prov-primer/ ------

The document calls section 2 "intuitive" four times--I would say show, 
don't tell, or at least don't tell four times. "High-level" would be 
more accurate (and more modest). Section 2 is actually not that 
intuitive, because it covers a lot of material at a pretty abstract 
level. The Primer is much easier to follow once you get to section 3.

To make it clearer about how helpful section 3 will be, the bulleted 
list at the end of section 1 could be more explicit that the first two 
bullets refer to the remaining sections of the document ("section 2 
gives a high-level overview of PROV concepts...") so that the reader 
knows when they're getting to the more concrete example. You could even 
add to the bullet about section 3 something like "in which a blogger 
investigates the provenance of a newspaper article to track down a 
potential error".

"There are other kinds of metadata that is not provenance" that are not 
provenance

"the author of an article may attribute that article to themselves" the 
authors (because of the plural "themselves")

"the agency also wish to know" wishes

If some of the example qnames were renamed to be less generic, it would 
make section 3 easier to follow. For example, "ex:article" looks more 
like a class name; ex:article1001 looks more clearly like the identifier 
for a specific article.

An added bonus for section 3.9 could be some RDFa syntax for the first 
example, given that it's about Betty embedding provenance information in 
her blog entry. Something like this, which rdflib confirmed to me gets 
translated to the appropriate triples:

   <p>According to a recent government report,</p>
     <blockquote about="ex:quoteInBlogEntry" property="prov:value"
                 typeof="prov:Entity">Smaller cities have more crime 
than larger ones</blockquote>
     <span about="ex:quoteInBlogEntry" rel="prov:wasQuotedFrom" 
href="ex:article"/>

In fact, a little PROV-RDFa cookbook, perhaps as a separate document or 
even blog entry, could help to jumpstart the use of PROV among the 
Bettys of the world.

Again, great work and I look forward to using PROV.

Bob DuCharme
Received on Friday, 22 March 2013 14:26:44 GMT

This archive was generated by hypermail 2.3.1 : Friday, 22 March 2013 14:26:44 GMT