On 3 Aug 2012, at 15:56, Joseph Montibello <[log in to unmask]> wrote:
> search, you could probably do worse than to install Blacklight.  It
> probably doesn't really meet the "simple" criteria - there's a lot more to
> it than I could talk about.  But getting it out of the box, turned on, and
> searching against a few records is something that you and students could
> probably manage. I've got a year of unix/ssh/command line experience and
> with a bit of mucking about, googling, and asking for help I was able to
> get a local (non-production) instance up and running, so it's definitely
> easy enough.

I'd agree - either Blacklight http://projectblacklight.org or VuFind http://vufind.org are straightforward to get running. I've found Blacklight setup using the Ruby Gem very easy both on Windows and OS X. Since they are both powered by Solr and use SolrMARC there are a lot of similarities on the indexing/searching side. However on the interface side they differ in terms of setup - so it might be this that would sway you one way of the other (or a preference for PHP (VuFind) or Ruby (Blacklight)).



>> 
>> Lesson: Interfaces, usability, accessibility
>> Exercise: Use the OPAC, populate it with some data, assess its usability

Once you've got VuFind/Blacklight setup populating with data is a matter of uploading some MARC21 records - Blacklight comes with some test records bundled, I suspect VuFind does to but can't remember

>> 
>> Lesson: HTML/CSS
>> Exercise: Use CSS to skin the OPAC, customize the HTML for your "site"

This is slightly more complex I guess - both systems can be highly customised, but in either case it isn't necessarily just a matter of editing CSS or HTML. Both use templating systems and both have configuration files that control certain aspects of the interface (e.g. what is searched, how facets display). CSS is probably more straightforward - VuFind you can just drop in CSS to override the default - not sure about Blacklight

>> 
>> Lesson: Data management, search, IR
>> Exercise: See if we can peak under the hood about how the OPAC's search
>> works
>> 

I think this would be the real strength of using Blacklight/VuFind - Solr/Lucene is a powerful combination, and used widely outside the library sector. You can also configure the indexing to a high degree - lots of options, the most basic of which I explore in http://www.meanboyfriend.com/overdue_ideas/2012/07/marc-and-solrmarc/

The thing I really like about this is students would see some of the complexity of MARC as well as some of it's utility - and where it doesn't work well

>> Lesson: Interfaces to data: databases, XML, SQL
>> Exercise: Use the OPAC as an living example to work with those interfaces

This is less well served by Blacklight/VuFind - no database, no SQL.

>> 
>> This idea primarily came from trying to get some simple XML/SQL
>> exercises that didn't suck (the setup for these environments is almost
>> as involved as any exercises itself), and the fact the previous classes
>> really liked dissecting the nextgen catalogs we've explored from a
>> software selection and 2.0 integration perspective.

Unfortunately it may be that Blacklight/VuFind don't work for your scenario because they don't provide an environment for SQL. You could do some XML stuff (there is configuration files, and Solr can be updated via XML messages) - but I'm not clear whether this is the kind of XML work you want. However, I do think they open up some other avenues that are well worth exploring, and use technologies that are going to become more relevant in the future.

Another option might be BibServer, which uses elastic search rather than Solr - but I've never tried installing it http://bibserver.readthedocs.org/en/latest/install.html