There are a variety of ways to characterize library materials. The type of reader believed to be interested in a particular item is one. Such an indicator, generally known as the audience level, is potentially useful for a variety of activities, including the development of new ways to improve information relevance for retrieval, reference services (including readers advisory) and collection development. Audience-level filters could be implemented in existing retrieval systems to assist users in finding content based on their information needs.

This is not the first OCLC service PatREST has taken advantage of. PatREST has been incorporating data from OCLC’s xISBN service for quite some time. By pulling in the data they make available, the data PatREST is able to return becomes significantly more valuable.

Because AADL’s PatREST implementation relies heavily upon III’s XML server, I’ve added OCLC’s Audience Level functionality to the PHP XMLOPAC class code which is freely available from my files page or you can directly grab it right here.

The new versions allow users to display cover images along with the records. A new option gives you the choice of text only, images only, or images and text. Not a major change, but noteworthy. Also, in case you missed the update in my previous post, the new items gadget can now match subject headings–useful if you want to be notified about new items on a particular topic.

For the purposes of the Talis mashup competition (for the judges), the original xml files are still available under a different name (tops-v1.xml and new-v1.xml). Everyone else, here are the new ones:

• tops.xml
• new.xml

(FYI, it’s the same URL as the previous version. If you’ve already added it to your Google page, the update will be automatic)

While working on this little project, I became painfully aware of PatREST’s limitations when dealing with asynchronous execution — like that of Google gadgets. I previously thought it would be better to limit the amount of data returned in an XML hit-list and use a second record query for and detailed info. I think I may have been a little short-sighted. At any rate, the lesson learned is that the more practical experience I have with PatREST, the more I’ll know what works and what doesn’t.

The upshot of all this is that I’ve expanded the result objects in any PatREST function that returns multiple records to include more information, such as ISBN, cover image, author, and record link. For those asynchronous folks, this will make life a lot easier. The new additions have been added to an updated specification. Existing PatREST applications (I don’t think there are many at this point) will continue to work, of course.

First, two new functions have been added to the service. One provides access to tops (or popular items) lists. The other provides access to the new materials lists. I believe these are significant-enough additions to the service that they merit the 1.1 version number.

The top items query is scalable by result size and can be paginated, just like the search results. In addition, it can be scoped by material type: books, cd, dvd, or bocd (books on CD). When applied to the AADL XSLT, it looks something like this:

The new items query is similar to the top query in that it can be scaled by material type, size and be paginated. You can also search new items using their subject headings–useful for querying those new knitting books, Ed… It looks something like this:

Second, and probably more importantly, I’ve finally drafted a specification for PatREST which includes an explanation of it’s XML schema and some documentation for it’s various functions. It’s about time, I know. It can be found in the files section, or downloaded here. (PDF)

I’m not going to go in to too much detail about the two new functions in this post because the new documentation contains everything you need to know to get started with them.

In addition, Ryan has written up a couple great how-to's on getting item data and featured lists from the XML. I, for one, had no idea it was possible to grab featured lists this way. Ryan Eby has been documenting III's XMLOPAC for quite some time now and he's certainly one of only a handful of authoritative voices on the feature.

David Walker is another, and a very industrious voice at that. Today I had a good chat with him on the Code4Lib IRC channel after he showed me his totally amazing catalog based off his equally cool Shrew project. What he's been doing is exactly what I've been looking for. Even though the Shrew project is currently written in C#, he has plans to port it to PHP5, taking advantage of DomDocument. It's a project I'm completely willing to commit some time to myself, if he asks. The shrew project is "a system that converts the Innovative XML Server into an SRU/SRW and OpenSearch server." He's put a great deal of time into writing XSLT that will translate the III server's output into MARC-XML, Dublin Core, or MODS 3.0. My XMLOPAC class for PHP5 utilizes an older version of his MARC-XML XSLT, but I think when he pulls off this port, the need for my code will go away altogether--his way is preferable.

We also discussed some inherent problems with III's XMLOPAC--of which there are a number, and some potential enhancements. Chief among them would be the ability to conduct business--placing items on hold and such. A lot of work needs to be done, but I'm feeling much more optimistic about my OPAC aspirations now and, thanks to David and Ryan, a renewed sense of enthusiasm. Thanks guys!

[update] Ryan reminded me that Rob Casson is the individual who was kind enough to provide hosting for the wiki [/update]

[tags] OPAC, XMLOPAC, XML, XSLT, Library, Code4Lib, PHP, C#, SRU, MARC, DomDocument, Web Services [/tags]

In addition, I've included an XSLT file written by David Walker which I've modified slightly. It allows me to easily extract subject headings from the XML, so yes, this class now will return an array of subject headings!!

So, be sure to download the latest version. As always, you can grab it anytime from the files section.

When last I wrote about this, it was little more than a working proof-of-concept, but I've been working on it a lot this week--partly driven by DaveyP's experiments and Ed's tinkering, but mostly because I've been planning to do this for a long time now.

So, to cut to the chase:
PatREST now provides an easy, RESTful URI to do searches on quite a few fields. Essentially, you will plug the search key into the URI, followed by the search term. You can also request optional paging for keyword, author searches by appending hits-per-page and page-# to the URI. It looks like this:

http://www.aadl.org/rest/search/[searchkey]/[searchterm]/[hits-per-page]/[page-#]

The following search keys are available:

title
author
callnum
keyword
subject
gvtdocnum
stdnum (ISBN/ISSN)
titlekey
controlnum
barcode
record (Bib #)
bibnum (Bib # - Same as record)
itemnum

Go ahead and try this keyword search. It'll look something like this:

You'll notice that the XSLT stylesheet presents it nicely, all the data may not be displayed, but the XML is sound. Go ahead and view source to examine the schema. I've departed quite a bit from my original schema in order to provide a little bit of metadata for processing purposes. The XSLT allows you to click on a title which will take you to another RESTful record:

You'll notice that the URI for this record looks like /rest/record/1035670/ ... It's /rest/record/[bibnum]/. The XSLT allows you to click on either the title or cover image to go to the regular OPAC record. Again, view source for the XML schema. This schema has changed little and you'll notice that I'm taking advantage of OCLC's xISBN service.

The real treat in all this, however, is the ability to access your personal records RESTfully. To do this, I make use of the RSS token that is provided to every cardholder who has registered for an online account. This is the token that authenticates RSS readers against our system. It is a 32 character MD5 hash that looks something like "316928e0d260556eaccb6627f2ed657b". Accessing personal data is as easy as using the following URIs:

For checkouts, you would use /rest/checkouts/[token] (ie http://www.aadl.org/rest/checkouts/316928e0d260556eaccb6627f2ed657b). You'll then get output at looks something like:

For holds, the URI is /rest/holds/[token]. Output looks like:

Both of those results allow you to click on the bibnum and access the REST record for that item. Again, check the XML schema with your browser (no point in putting it here).

Please hack away at this and send me your comments/suggestions. You're absolutely correct that it doesn't adhere to any existing standard, but that's because I didn't want it to.

On a slightly less geek-oriented note...

I've gone several rounds with Talis's Richard Wallis before, so I want to call your attention to a post he made the other day in which he suggest that a) DaveyP and I collaborate and b) I/we consider using industry standards. I'd like to respond by saying that Dave and I have been communicating extensively. Ed Vielmetti has been involved as well.

Richard writes:

I encourage John, Dave, and those that follow them to take a look at these standards like we have

I can't speak for Dave, but I'm very familiar with these standards and so is Ed. I think Richard is completely missing the point of this whole project which is to put a friendly, accessible, and useful development interface into the hands of patrons. Ed offers just a small example of the type of features that existing standards don't readily make available:

Already there is a lot more useful function in it than you can get in SRU, e.g. permalinks for card records and a sensible way to get item availability.

Bear in mind, that comment was made prior to my work with RESTful holds and checkouts, which, as far as I know, don't even have a standard XML schema--I don't think it's been done before! (I'd love to be corrected on this) This project will also evolve to the point where holds can be placed RESTfully, items renewed, fines paid. It'll also extend to our other features like checkout history, wifi device management, personal card catalog management. Adding this functionality is not only going to allow the public to develop highly useful applications, but it'll be a framework by which we ourselves can build new services in-house. Vendor's obviously haven't stepped up to this, so we are--and we believe in our patrons. They deserve this kind of access to their public library.