Re: [office-metadata] Must preserve?

[Bruce D'Arcus <bruce.darcus@OpenDocument.us>]

On Sep 13, 2006, at 8:12 AM, Patrick Durusau wrote:

> I may have just missed another way of saying it but I did not notice a 
> requirement that metadata must be preserved.

I don't have it as a separate requirement, but folded into the levels 
of conformance one:

"The enhanced metadata proposal should define levels of conformance so 
that users and developers have expectations about the level of support 
to expect in a given application. At minimum, all OpenDocument 
applications must preserve foreign content, ..."

It might make more sense to have it separate though?

> I think we need to add conversion to the "summarized" use cases as it 
> is the one that makes the clearest case for preservation of metadata.

Sure.

> If I convert from some foreign format to ODF and include metadata that 
> will facilitate conversion back to the foreign format, if that 
> metadata is dropped by any ODF application, then I have lost the 
> ability to go back to that format.

Right.

> Note that I am *not* suggesting that any ODF application must 
> recognize or process such metadata, but it must preserve what it does 
> not understand. Some ODF applications will be able to use that 
> metadata to insure clean export back to the foreign format or other 
> purposes. Nor is any ODF application required to be able to export 
> back to a foreign format. But, all ODF applications should preserve 
> the metadata that makes that possible.

It's worth noting that right now, one can embed arbitrary XML in files. 
Just add your "foo.xml" to the file wrapper register it it in the 
manifest, and you're valid.  Florian mentioned that OOo, for example, 
will preserve this.

So I think he was thinking of writing out that metadata into a standard 
form (RDF). E.g. if he has a bunch of property-value stuff he's reading 
and needs to store it somewhere, using RDF to do that is perfectly 
reasonable.

As an aside, I mentioned to Gary that I was looking at some XMP files 
that Adobe is exporting from their new Lightroom application. It 
*seems* if I understand it right that they are also outputting similar 
sort of conversion metadata.

Lightroom is a resolution-independent image editor. So it does not 
modify your RAW image file when you "edit" it; it simply describes the 
edits mathematically. So if I am reading it right, Adobe is actually 
including that information in the metadata files.

I could be wrong on that, but in principle the notion of using the 
metadata to carry this structured data sounds reasonable.

Bruce