Robot4583 at your service...beep beep

Cataloguing was a big concern for me in Eprints. This is because I had already entered every item in other repositories: some in Drupal, some in Dspace, a few in both. So in the interest of being consistent with those previous entries I looked into importing the metadata from those CMS's into Eprints.

In Drupal, this simply did not seem possible because I didn't use Views for my collection. It could be that if I created a view for it, I could export it, and I may try that in the future. I expect there are modules to make this all work better but when you are evaluating systems on the time span that I am, you don't necessarily ever reach this step.

In Dspace I was able to export my collection to a zipped folder, but I couldn't get the hand of batch importing into Eprints if that is even possible. (It should be.) So instead I manually uploaded all 4 or 5 files (including one classified as "other" and therefore requiring a description, which I made DspaceBundleDescription for each) from each dspace item from that bulk export, then manually added title (copied and pasted from dublin core xml doc), abstract, creators, and status. All this me from having to enter details. I also started adding keywords (listed as subject in the Dspace-exported dublin _ore xml doc). And I selected Plain Text for format for the DspaceBundleDescription "Other" doc for each item, and also selected "Additional Metadata" for the dublin core xml docs.

Nonetheless I still had to then detail the record, adding at the very least: Title, Creators, and Status (I selected Submitted). All then had to be approved by me.

Eprints gave me many areas where I had to repeat identical metadata in this collection and this definitely led to human error. I minimized this as best I could by opening the Dspace-exported dublin_core.xml doc for each record and pasting from it and also constantly referencing it as I described the item in Eprints. But this felt like it should be unnecessary, and was certainly less foolproof than a computerized entry system.

One day, they'll sort out repositories so that humans only have to do what humans are good at: enter new concepts. For now, though, this is Robot4583, signing off.

What is this repository about again?

Well my installation of a third CMS and the creation of my 8th or 9th repository is complete...and I am beginning to feel a little weary of puppet resources.

I did enjoy this installation, however. So many defaults accepted! I am normal! My configuration has been anticipated and all but arranged!

Interesting that you have to go to the CLI monitor to make design changes. It seems like it shouldn't be that way, and yet, it makes the line between the designer and the content manager very clear, especially when the content is entirely distinct from the design, as is the case with many repositories. The user interface is better organized than Dspace, whose creators seem to have never heard of tab menus, and less confusing than Drupal, whose creators got a little too high on menus. And the continued expectation of communion with the command line puts everyone in their place, which seems more attractive as I work with users who expect a mile when given an inch in design options.

So it's more attractive and belled-and-whistled than Dspace yet simpler than Drupal. Not incredibly modular, but easy to introduce without a tutorial. Overall I may like Eprints better than Drupal or Dspace for a few collections I am working on, which makes me see how important it is to evaluate before you launch. You can't very well change the delivery vehicle when the resources are already on the train!

A 2nd repository, for Breast Cancer Resources

I found Drupal to be the right repository management software for my 675 collection of puppet resources, but I am going to be using Dspace in the next two weeks to create a second repository, of breast cancer resources.

This is because I have worked with breast cancer resources in the past, in the creation of a metadata schema, controlled vocabulary, and relational database, for IRLS 515. I've been interested in continuing to develop these modular components to work together, and consequently I just got accepted in the GPSC Student Showcase for my work with breast cancer resources, which I'll exhibit under the name "An Integrated Approach To Information Discovery In Consumer Breast Cancer Research." The resources used for the work so far have been online, with our series of integrated projects designed to direct consumers to them while reporting on the quality of information that is accessed.

I decided that at the Showcase, through lively graphic and digital representation I will also show where this integrated approach to information discovery is going. The user interface I am currently creating will deliver a sample set of online results, along with research I find in Creative Commons stored in a local repository. I don't expect it to be complete by any means, but I feel members of the University of Arizona's rich scholarly and medical communities would appreciate peering into the development of this project while it is still transparent, and while I am still able to incorporate feedback and opportunities for interdisciplinary collaboration.

That's all she rote

I really enjoy the way Bruce will instruct us to do something we did weeks or even months before, and expect us to figure out how to do it again. He is very offhand as he does this, isn't he? As if he doesn't realize that most of us are going to have to go back into this video from week 2, that tutorial from week 1, the install guide from 672 (which hopefully we have, since D2L sure doesn't anymore!), and click back and forth between three different pdfs all the while. UNTIL four hours later, you successfully reach line 2 of the 9-page TechAssignment document for the week.

Obviously this is the best way to make us learn these steps on our own, and in fact I feel fortunate he walks us so carefully through the steps of each install the first time. But I always wonder, is he laughing when he writes those simple statements?

Installation of DSpace wasn't horrific for me. But the more complex an install is, the less I am able to know what the meaning of each step is, and this install was more rote than any I have done this semester. It only helped, oddly enough, to be using VirtualBox instead of VMware. This is because I learned this week that I may be the only one in our class doing this, when I discovered that my version of Linux is older than most of the other students', who installed theirs in 672. (I installed one on a VMware machine in 672 as well, but had to switch computers. Now I hope to never go back; VirtualBox is much less awkward about mouse and keyboard integration than VMware.) So I sometimes have to troubleshoot a little differently than the others, and this gets me thinking.