Omeka as dissertation research archive

I'm using Omeka to host a small collection of largely film journals and newspaper articles on early German cinema. I had collected these for my Ph.D. dissertation over a period of several years and finally realized that even I couldn't make sense of the hundreds of sheets of xeroxes anymore.. So I decided to put them online as much for my own sanity as to help others working on similar topics. The collection/archive is called From Kinema to Caligari: Sources and is linked to a related dissertation blog.

Just a few notes here about my set-up in case someone wants to embark on something similar. I plan to write a longer review/how-to about my experience on my digital humanities blog soon.

Initially, I just added items in small batches by hand (either via the Add Items dialogs or using the Dropbox plugin) as I scanned in my originals. In retrospect, it would have been smarter to prepare the metadata in a spreadsheet and imported as much as possible in one go using the CSV plugin. This would have a) saved a lot of time b) given me a structure (the spreadsheet) to push bulk changes to the metadata via subsequent CSV uploads. An export to CSV plugin would definitely be welcome here.. Actually, importing the data caused me a lot of grief initially. First, I was importing PDF files. Omeka can't make thumbnails of PDFs. But a lot of the themes really assume that you have these.. Second, my hosting service, Dreamhost, has a very old version of ImageMagick and ghostscript installed which kept throwing exceptions when I tried to feed it PDFs or even multiple images. In the end I had to build my own versions of both from source. This (mostly) fixed the problem but not really (I don't want to get into it.. the pain is still too fresh :^) So a word of advice - don't just check the version(s) of imagemagick and ghostscript you have installed on your webhost. Also check to see if it can do the specific conversions *you* need.

In terms of organization, I quickly realized that Exhibits were overkill for my purposes. Collections, featured collections and featured items were quite sufficient. Keep in mind though that an item can only be in one collection. So some advance planning is needed. This also applies to how to fill the two arguably most important Dublin Core fields - title and creator. This is where I ran into my second (post import..) gotcha. The standard keyword search used by Omeka is based on MySQL's full-text search. This is ok.. but has some limitations. Namely, it ignores anything with less than four characters and it ignores anything in its stoplist file. So if you have a DC title like "Der Film Nr. 4" then a search with this term will return all items that include the word "Film" and ignore everything else. That is, unless you use the Advanced search which is regex (regular expressions) based. If I had known this before I started I might have constructed my title fields differently.. In any case, a PIA but fortunately there are workarounds. See the "Help" page on my site for more info. I think it'd be helpful if this and similar "known limitations" were more readily accessible / front and center in the Omeka documentation. Maybe the release notes would be a good place for this?

As far as themes and plugins I opted for something very simple since my material is hardly thrilling to look at (mostly b&w page scans of old newspapers..). "Thank You, Roy" turned out to be a good fit. Simple Pages, SimpelVocab and Docs Viewer are important plugins for me. I didn't like the appearance of the Social Bookmarking plugin and so used the older, unsupported AddThis plugin instead which seems to work (I hope..) fine. I also found the XMLSiteMap plugn to be very useful. It'd be great of this were part of the Omeka core and attached to a cron job.. Maybe once I have better metadata I'll create a new page and make use of the Timeline plugin as well. Many/most plugins are officially "out of date" yet still work ok. This can put off the cautious newcomer.. Maybe the Omeka team can confirm the ones that work and update the version & compatability info of the main/core plugins in parallel with future releases. Also the presence of plugins in the codex (but really at UVA) is a bit confusing. If they're at UVA they shouldn't be in the codex, I think. Or yes, they should be but then they should be in the official plugin list but they're not.. Then again, maybe I'm the only one who is confused by this :^)

There were a few simple things I wished I could do (or wish I could have done more easily) which you may run into as well. For example, how to add a link underneath the list of SimplePage plugin pages on the home page. An Omeka developer wrote a few lines of code for me for this purpose which I was then able to install as a custom plugin. How about a generic SimpleLink plugin? Given that I'm using the Collection description as a kind of Exhibit lite it would be nice to determine the length of the collection description that appears on the site. An Omeka developer offered a solution to this which didn't initially work - I need to get back on that. Again, I think this would be a nice option. Also, had I used the Exhibit plugin, I think a canned layout with Image on the left and Text on the right (or vice-versa) would be very, very helpful. We're not all showcasing images that can/need to stand by themselves. At one point I tried to do it myself but wasn't able to get the CSS right. Speaking of CSS I think it would also be great if the themes could be made a bit more generic. There are a lot of assumptions about their 'ideal' use & design built into them. That's fine.. but this needs to be explicit. For example, I had to do a lot of (note: for me..) fiddling with the CSS styles and dimensions of my header image to get it to behave more or less as wanted. Alternatively, something like a generic theme header editor (ala WordPress) and a generic media handler (again, ala WordPress) would be very useful. WordPress has made huge strides in this direction recently.

Ok, well.. that turned out longer than I thought. Hope it includes something useful to others who want to build similar sites and as feedback to the (always helpful, always patient!) Omeka crew.