[OAI-implementers] Harvesting DC records for DDI system

[eric casteleijn]

eric casteleijn wrote:
>> We would like to make the system search the UK Data Archive (UKDA) - 
>> as the UKDA does not use VDC we are looking at being able to harvest 
>> their records. The UKDA have a test OAI service but this only contains 
>> a reduced Dublin Core record for each of their catalogue entries and 
>> it seems that the VDC system will only harvest DDI metadata. Has any 
>> body got any experience of what would be needed to harvest DC records 
>> and put them into DDI. At first glance it seems that DC and DDI map 
>> pretty well.
>>
>> Any advice on how this would be done and how much work it would take 
>> would be much appreciated.
> 
> 
> If the needed DDI-fields are there in the DC metadata, an xslt 
> transformation could be used to translate the harvested xml to the 
> format you need. (I have no experience with DDI, so I can't really say 
> how hard or easy this will be, but I don't think it'll be very hard). 
> You could then have an intermediate server that harvests UKDA and 
> provides an OAI service with the metadata in DDI. This can probably 
> achieved in a lot of ways, but at least one of them is using the open 
> source pyoai package (full disclosure: I work for the company who makes 
> that, but it is available free of charge through cvs here: 
> cvs.infrae.com/pyoai/), which provides both OAIPMH client and server 
> software. Putting just a thin xslt-layer between harvester and server 
> should not be that hard in Python.

Sorry, I was confused, probably due to caffeine deficiency: what I 
(should have) meant to say is: you could do this with a thin xslt layer 
that takes the xml output of the UKDA server and transforms it to DDI, 
*or* you could set up a small intermediary server that harvests UKDA and 
outputs DDI, but this would have nothing to do with xslt.

I'll be having a coffee now.

eric