one of the problems facing release audit tools is understanding
documents which may legitimately lack self contained license
information. some examples:

* original binary documents (such as .png's)
* original text documents whose to which headers cannot be added (for
example, test outputs)
* generated documents
* documents with suitable licenses created elsewhere
* binary dependencies with suitable licenses (such as jar files)
created elsewhere

this leads to a lot of false positives which need to be followed up manually.

one solution to this problem might be for each audit tool to contain
some way of configuration which would allow specified files to be
excluded. i think it that it would be more powerful to ask that
self-documenting meta-data is added in subsidary documents referencing
the originals. RDF seems like a very good fit and has the advantage
that adopting this framework lessons the need for the creation of yet
another meta-data format.

i need something along these lines for RAT but i believe that this
approach may be useful for other applications as well.

i've taken a look at DOAP but i don't think that it's what i'm looking
for: i'm interested in the elements that collectively make up the
project rather than the project as a whole. i've done a quick search
and can't find anything which looks right (suggestions welcomed). i
think that a new, small vocabulary or ontology would be better.

i can see many advantages to starting this initiative here in the labs
but it would be a little unothodoxy: though there may well be code
(support tools), the focus would be on creation of an adequate
ontology or vocabulary, not on the code.

opinions?

- robert

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to