March 03, 2004
Note: CogDogBlog has a new WordPress powered home at http://cogdogblog.com/. All entries from this version have been moved there, so as a guide dog service try finding this article in its new home by title search.
Meta-Data Yeti-Data
My position have been made too many times regarding the apparent over-obsession with learning object meta-data, 4 words guaranteed used together will put most ordinary humans into a mild coma. I've given thought to meta-data and our Maricopa Learning eXchange, where the "M-D" words will never appear, but certainly lurk under the cover of our "packing slip" metaphor.
Like Bigfoot, Nessie, Yeti, and other fuzzy photographed un-worldy beings, meta-data is something I have yet to see used by those ordinary humans. Not to say it does not exist, but it sure is hard to find. Some recent RSS cruising over the last few months have on occasion made me wonder, 'what would happen and what would we gain if we created 'standard' meta-data attached to MLX items?' Would there be loud operatic music? Parting of the skies? Raining gold coins?
I even quickly sketched on paper one day a mapping of our MLX database and Dublin Core meta-data. It would likely take but a few hours tinkering to write a verion of our "slip" generating script to generate the same info in proper DC format. But again so what? I asked some more expert-ed folks who could not really say enough to make me bother. Scott at EdTechPost recently wrote of a new report about "OAI" (Open Archives Initiative) compliance and repositories that had some potential but was deep in gobbledy gook.
From what I could glean, it is not that complicated to make this happen. Even without doing anything, the Dublin Core metadata editor extracted some info directly from a ML]X URL into DC data and then some tinkering resulted in a better formatted example.
However, I am completely at a loss as to what is gained by doing Dublin Core, by achieving OAI compliance. I am not opposed to doing so, and sort of interested in seeing how it works, but I am pleading for a simple explanation, something that has a minimum of acronyms and cryptic references.
Yes, please end me a copy of OAI Compliance for Dummies or Using Dublin Core for Totally Doofus Idiots. Even better if they are illustrated.
Seriously, if someone can help me understand, gratitude will be yours.
blogged March 3, 2004 09:40 PM :: category [ mlx , objects ]
Feed for comments on this entry Feed for all CDB comments
I recently started working with the OAI harvesting protocol - http://www.educate.za.net/archives/00000107.htm - and actually once you get past the reams of pointless technical data, it is an elegant and simple solution. In my experience looking at something like - http://eprints.ecs.soton.ac.uk/perl/oai2?verb=ListRecords&metadataPrefix=oai_dc - the xml returned via a - http://oai.dlib.vt.edu/cgi-bin/Explorer/oai2.0/testoai - repository explorer a lot more useful than any of the texts that I've read.
Commented by: root on March 4, 2004 02:44 AMThatnks for the quick comment, "root" is that what you are called??
Actually I had visited your blog and that was a recent prompt to think agian about this, anmd Scott Leslie had provided a URL to a nice PHP library:
http://physnet.uni-oldenburg.de/oai/
that handles the grunt work of the OAI requests.
Yet... Yeti... I ams till at a least. What exactly does one do with the harvestings? The OAI stuff is focussed on papers and looks like a great library / research tool, but what is the value for "learning object thingies"? Harvest, harvest, harvest is all about the discovery and finding of resources, but it is less than half of the goal- where are the things that help you do something meaningful with the things harvested?
That is the value I am looking for.
Commented by: Alan Levine on March 4, 2004 11:07 AMAlan, a few thoughts:
- there's lots of IEEE LOM/IMS metadata that MLX packaging slips don't contain because the specific context of the MLX already communicates them; e.g. the notion of 'educational level' - when you built MLX, you built it for a specific community of users (the Maricopa community college system) with the intent, I'm assuming, that the resources placed in there were appropriate for that community. Now if suddenly either new types of users start to look for things in MLX (e.g. high school teachers) or you start to ship these records off to other repositories (or federate or harvest or whatever) and so they become part of a larger set of results, maybe that information becomes important. When the people we trying to create the IEEE LOM, they tried to anticipate as many scenarios for use as possible and not make assumptions about context that were better reflected as data.
- on the other hand, there is metadata from IEEE LOM missing from MLX packing slips that is simply missing, e.g. 'duration' - an estimation on the approximate duration of a resource. In theory someone looking for a resource (especially amongst an absolute mass of them, an assumption again I think the creators of metadata schemes are working under) might want to constrain their search to only those resources under 10 minutes in duration. In practice, though, lots of folks question whether people actually do searchs like that at all. Still, in trying to develop what is in essence an 'idealized' description of a resource, I believe the thinking was better to make space for it than leave it out. It need not be a 'required' element, some might be populated automatically by software saving workload issues, and even if people don't search on it, as David Davies recently pointed out there might be software that can 'automatically' take advantage of the data being present (cf. http://david.davies.name/weblog/2004/03/15.html#a593).
- in regards to the OAI question, on the advantage of MLX being harvestable by OAI, it seems to me the question could be re-phrased as "What advantage does the structure that OAI (or more properly Dublin Core) imposes on its data bring to people specifically searching for *learning* objects, and is their an advantage to both searchers and creators of those objects for people being able to find them through search interfaces where they will either typically be searching for 'scholarly publishing' type material or will find MLX resources intermixed with those."
Personally, I am still waiting for some convincing demonstrations that it is in fact useful, but that said I am glad that people are preserving the ability to do so and not pursuing paths that automatically preclude our ability to do this in the future if that is the path we chose.
Sorry for the rambling, cheers, Scott.
Commented by: Scott Leslie on March 16, 2004 11:27 AM
If you really need to make contact with the chief dog around here, please submit a request via our feedback center