[doap-interest] Auditing Releases

[Robert Burrell Donkin]

On Sun, 2008-08-24 at 09:25 +1200, Stuart A. Yeates wrote:
> There are a couple of relevant issues here:
> 
> (a) Reusing existing features is preferable to inventing new ones,
> since existing tools can likely do something sensible with existing
> features, but getting them to add new features is an uphill battle.

i agree

my only concern is that since:

"rdfs:label is an instance of rdf:Property that may be used to provide
a human-readable version of a resource's name."

the risk is that existing tools may process the information in a manner
befitting a human readable text and not in a way that befits machine
readable data. still, i'm not sure that  

> (b) Many geeks recognise a checksum / signature

i agree

> (c) I believe that mime-type is the correct way to describe the 'type'
> of this information

i agree

> Do you have an alternative idea for the representation?

i suppose that an alternative strategy would be something like that
taken by http://simile.mit.edu/wiki/Maven_POM_RDFizer etc and use a
concept definition as predicate. for example:

<wikipedia:Computer_file
rdf:about="maven:///wicket-jdk14/1.3.0-beta2/wicket-jdk14-1.3.0-beta2.pom">
  <wikipedia:MD5>f263dd7270c31dbad88f4510753cdf23</wikipedia:MD5>

<wikipedia:SHA1>a49d12523a9e66b8e68e6cb142d27c6c961bdb96</wikipedia:SHA1>
</wikipedia:Computer_file>

(these days, http://dbpedia.org/ would be preferrable as the namespace) 

AIUI the idea here is to link to a concept so that reverse searches can
turn up data. for example, searching for MD5 sum would
find f263dd7270c31dbad88f4510753cdf23 and then deduce something about
it's relationship with the subject.

in some ways, i suppose this difference in philosophy may reflect a
difference in the tool audience targetted.

- robert