LC Tissandier Collection and OAI-PMH Collection Issues
The items I decided to include in my “LC-Tissandier Collection” on D-Space I essentially chose at random based on how the images caught my eye. The two items I chose had some of the metadata fields readily apparent based on the information on the LC catalog. But some of the fields that D-Space requested were different than what LC listed, or were at least not listed as separate fields in LC. I noticed that the OAI-PMH Collection described fields using different language than I used, and that some of the information was redundant. As part of the item description field it points users to the title, which has already been listed above. And a minor difference was that the language abbreviation for French was listed in my manual metadata as “fr” as opposed to “fre” in the OAI-PMH collection. The OAI-PMH record also does not list the author at all, just the publisher of the item. I suppose I made an assumption that an automated data harvest would not make, I assigned authorship to Albert Tissandier even though it is only implied instead of being explicitly stated. For both items there was also no abstract listed by the automated harvest, this is once again understandable since the LC website does not offer an abstract for these items like D-Space asked for. I constructed my own abstract, and the harvested data set would not be able to take the “initiative” do something like that.
OAI-ORE Struggle
The issues faced with the OAI-ORE collection were different than the other two collections. The first online collection we attempted to harvest from would not allow access, I attempted the harvesting operation several times but I would only receive an error message. The second collection also caused a few difficulties but eventually worked. I did wonder when I was attempting these harvests, which I believe I attempted at least ten times between the two collections, if it would have been easier and more efficient to construct the records manually. Being dependent on an external server can cause quite a few headaches, and it was instructive to have those problems manifested in a “low-stakes” class assignment rather than for an important project later on. If I’m ever in the situation again where I need an external server to provide a service like this to me, I know that I will need to allow ample to time for the possible technical shortcomings.
As far as the actual items in the OAi-ORE collection, I did notice that like the OAI-PMH and the manual LC collections the fields were fairly consistent within the collection. However the OAI-ORE items have a redundancy at the top with the date of the item listed three times with identical information. I’m unclear of what the purpose of this would be, and I imagine a human entering this information would have just not listed the others. An automated program like this obviously wouldn’t recognize a redundant set of information the same way a human would. Even with this redundant information, the fields overall are quite clear as far as giving the user detailed information. This OAI harvesting in both collections it seems can be a very useful tool, as long as the technology cooperates.
No comments:
Post a Comment