[OAI-implementers] Re. Metadata Language Confusion
Wed, 18 Jul 2001 15:35:16 -0600
I want to thank everyone for their suggestions about how to handle the multi-language metadata records. I have been researching all of the material you have sent me with dogged enthusiasm :)
I tend to agree with Diane Hillman that it would be better to keep all of the metadata for one article in one metadata record. Having separate metadata records implies to me that each metadata record is about a separate article. And this is not the case. I will probably go with using the language qualifier for multi-language elements. Some elements do not need to be translated (i.e. dates, authors, etc.), so it's at least not as storage and bandwidth consuming as having a separate metadata record for each language for each article.
In the long run, this does need to be handled at the service provider level. As others have pointed out the storage, bandwidth and human translation costs with maintaining multi-language repositories is not practical. My problem is only a very small subset of the more general problem. I will be very interested in seeing how the multilingual version of ARC progresses.
Finally, the OLAC metadata set is interesting. It handles the problem I have although it is not primarily designed for that purpose. However, another of it's design goals is how to handle multimedia material and that is a problem I need to solve for some other repositories we intend to construct. I am starting to do a lot of streaming media production and will probably use something like the OLAC construction to handle that stuff. I am curious how this fits in with a thread here a few weeks ago that mentioned VRA and CIMI for the cataloging of multimedia data.
Thanks again for your help. If I can be of any assistance let me know.
Health Sciences Center
University of New Mexico