incubator-sanselan-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeremias Maerki <>
Subject Re: Metadata use by Apache Java projects
Date Tue, 20 Nov 2007 08:13:45 GMT
On 20.11.2007 04:39:08 Charles Matthew Chen wrote:
> Hi Jeremias & Antoine,
>    Antoine, it looks like you found it pretty easy to convert
> Sanselan's metadata into XMP format.
>    Jeremias, it sounds like you considering a new project which can
> translate data from many formats (read by a variety of projects) into
> XMP.  That sounds great!

Not really. I'm proposing a unified storage/access model metadata in Apache
projects. The necessary mapping needs to be done in the individual
projects who have knowledge about all the different document formats.

>    Sanselan could not use XMP internally to represent metadata,
> though.  Sanselan's goal is to read & write metadata (such as EXIF
> metadata) preserving not just tag values but directory structure,
> field order, field location, etc.  I'm in the process of refactoring
> the metadata data structures at the moment, actually, in order to
> approach binary compatibility as closely as possible.

I'm not suggesting you abandon that. That would be a bad idea. But XMP
could be the generic format for reading/writing image metadata for all
image formats Sanselan supports. Look at ImageIO: it has a generic
metadata format and native metadata formats for all the individual image
formats. At the moment, Sanselan only has the native metadata although
with a somewhat common storage model (Metadata.Item).


Jeremias Maerki

View raw message