Tim Spalding

<a name="The_Pig_Flies.21_:_How_JavaScript_and_the_DOM_made_LibraryThing_for_Libraries_and_will_change_how_you_think_about_your_OPAC"></a>

The Pig Flies! : How JavaScript and the DOM made LibraryThing for Libraries and will change how you think about your OPAC

<a name="The_Myth"></a>

The Myth

• Your OPAC sucks.

<a name="The_truth"></a>

The truth

• Your OPAC sucks.

<a name="The_exhortation"></a>

The exhortation

• Stop whining and do something about it:

<a name="The_problem"></a>

The problem

• You can't get structured data out of your ILS/OPAC
• The can't plug things into your OPAC

<a name="Other_problems"></a>

Other problems

• Ajax is great, but you can't use Ajax across domains. This is called "cross-site scripting." Cross site scripting is baaaad, baaaad.
• Getting a scripting language onto your OPAC's domain can be hard or impossible.
• You may not have control over your OPAC templates. (eg., you only have control over the footer on the page).
• The HTML on the OPAC itself is childishly poor.

<a name="The_solution"></a>

The solution

• You can get your data, either:
  • Export your data;
  • Screen-scrape your data.
• Use JavaScript and the DOM
  • JavaScript allows cross-site scripting.
  • JavaScript can request and re-request all day long. (The trick is to dynamically insert <script> tags into the header.)
  • On the server side, your JavaScript must be generated (ie., with PHP, Perl, Ruby, etc.)

<a name="One_Result:_LibraryThing_for_Libraries"></a>

One Result: LibraryThing for Libraries

The web site: <a href="http://www.librarything.com/forlibraries/" class='external free' title="http://www.librarything.com/forlibraries/" rel="nofollow">http://www.librarything.com/forlibraries/</a>

<a name="Technical"></a>

Technical

• Export: Grabs ISBN, title, author. (Originally we grabbed just ISBN and scraped; what a pain.)
• Trigger: We use ISBNs. There are other ways, but ISBNs are easy because of the checksum.
• Programming: Javascript, PHP, MySQL.
• Testing: Greasemonkey. Oh yeah. We can show a library what it will look like without any work on their side.
• Cross-OPAC. We've had minimal problems with different OPACs.
• Accessibility. We link to an outside page, which looks at the referrer. This works quite well. Anyway, we're more accessible than the OPAC itself.

<a name="Examples_from_the_Danbury_Public_Library"></a>

Examples from the Danbury Public Library

• The book <a href="http://cat.danburylibrary.org/search/i?SEARCH=0670880728" class='external text' title="http://cat.danburylibrary.org/search/i?SEARCH=0670880728" rel="nofollow">Bridge Jones' Diary</a>. Includes:
  • Recommended books
  • Semi-FRBR "other editions"
  • Tags. Click on a tag and get a relevancy-ranked list of books that have that tag. This "tag browser" is a fine example of just how much you can do with JavaScript re-requests, CSS and the DOM. I didn't program it, so I feel okay saying that it's a frickin' work of art.

<a name="Problems_we.27ve_overcome"></a>

Problems we've overcome

• OPAC have such lousy HTML. We do a lot of "third blue table to the left of the ..." work. Tedious, but solveable.
• Sometimes the "browse history" on some OPACs includes ISBNs. We have to correct for that (ie., "discard ISBNs within a menu")

<a name="What_else_can_people_do"></a>

What else can people do

There is no limit to what you can add to your OPAC. Some ideas:

• Add social interactivity: tagging, ratings, reviews
• Add spell-check
• Add links to outside resources (eg., full text versions)
• Buy LibraryThing for Libraries and make us all rich as Croesus