On tor, 2007-12-06 at 09:04 +0000, Paul Hermans wrote:
> Hi all,
> I'm rather new to the list, but I second Ann in her request not to forget
> the XML MIN version.
> Our experience is also that the full version is too verbose, too academic
> for human beings.
This was discussed on the call yesterday, and there are several
interesting aspects of this.
In order to define a new DC-XML syntax, we need clear requirements.
So far, we only have well-stated requirements for the DC-XML-full format
* Be a complete representation of the DCAM
* Be easy to process using software, to help validation etc.
* NOT prioritise human readability at the expense of any of the above
While I'm not at all against a DC-XML-Min format, we would need to state
explicit requirements in order to have some guidance in the design.
So - the most practical way forward would be to start talking about
requirements for DC-XML-Min. I have heard, so far:
* Easy to generate by humans
* Easy to inspect by humans
* Compact and "non-academic"
* NOT necessarily support the full DCAM.
Now, the above would need clarification to be useful.
1. What makes an XML format easy to generate by a human? In what
situations would a human be hand-editing XML?
2. Same but for "inspect"
3. What features need to be compacted? Does this include URI
4. What does "non-academic" mean?
5. What features of the DCAM can be sacrificed in this format?
6. What machine-processing supported is desired? Should we expect an XML
schema for this format?
Please consider providing some input on this to move the discussion
<[log in to unmask]>
Plus ça change, plus c'est la même chose