Discovery2Delivery

Lorcan Dempsey's blog entry on Libraries Australia introduced the buzzword "discovery2delivery" to me (although he first talked about it last November). The "Discovery to Delivery" model is a great way to describe the perpetual mission of the library and is helping me crystallize my thinking about what I want the website at MPOW to achieve.

Let's compare the typical library web site to Chapters Canada's online presence (I'm comparing to Chapters rather than to Amazon because Chapters provides a brick interface as well as just the click interface). One might argue that the these two websites have similar goals: self-promotion and connecting users with "stuff". Interestingly enough, it's the public institution that seems to focus on "self promotion" and the commercial website that focuses on "making connections". This is most obvious when you look at these websites through the eyes of somebody that wants to find a book.

OCLC's Perceptions of Libraries report emphasizes that libraries are still primarily about books for our clients, but books are almost invisible on the Toronto Public Library's homepage. In fact, the only appearance of the word "book" on the home page is in the link "Find books & materials...." But it doesn't lead you to books: it's for a page of information about the catalogue, on which there is a link that finally leads to the catalogue. Unfortunately, the single search box on that page makes it very difficult to find a known item in the catalogue.

Not only does Chapters' home page have a search box, but the rest of the page is jammed full of information about new and hot stuff. And if I type the title of a book into the search box and press enter, without worrying too much about what I'm looking for, I'm likely to find it easily. (Try searching for Bernard Shaw's "Arms and the man", and see what you end up with.)

If we look beyond "moving product" (a motive that libraries and bookstores share), libraries' claims of being the center of community are also being undermined by the store's promotion of itself as a Meetup Venue Partner. Libraries' meeting spaces, like their books, are hidden away behind library specific processes, which are usually not accessible via the web. For example, although we have introduced a new web-based study room self-booking system at MPOW recently, and the students love it, it's hidden on a page that most of them wouldn't otherwise use: the branch's home page. Why does booking a room at the library require an instructional session with a staff member?

Commercial organizations well know that if they don't make it easy for the client, then the client will find someplace where it is easy. Unfortunately, the examples of finding books and making community connections both seem to demonstrate how libraries are displaying, via the home page, the relative importance of the library's internal processes and structures over the convenience of the client base.

The library's home page is the gateway to its services. It must be structured so that common tasks are simple to do, and the uncommon tasks are simple to find. The challenge is to ensure that commonality and simplicity are decided by the clients, and not by the librarians.

Why do Users want Lists?

Why are users so adamant about being able to browse journal lists that library IT group, and vendors like Serial Solutions and Ex Libris, spend time creating systems that create the infamous "A-Z List"? MPOW catalogues all of our online resources, just like our print collection (except online resources don't get call numbers). Everything we have is in the catalogue and it's usually not too hard to find the journals in the catalogue (except for the usual suspects: "Cell", "Nature", "Science", etc.) but users keep asking for a list of our electronic resources. We (I) finally broke down a couple of years ago, and we now produce our own A-Z list from a catalogue review file. So the users can browse through a list of over twenty-eight thousand titles. If somebody decided to print it out, the "A" section would run to seventy pages alone.

I hypothesize that at least part of the reason for this behaviour is that the users think "searching" is for when they are looking for information about a topic, and that a list provides faster access when they are looking for a known item (which Cutter identified as one of the prime purposes of the catalogue). There might also be a sense that, at some level, "browsing" is an effective discovery method. But again, how effective a technique is it when the list is 28K long, and alphabetical?

Protocol Design by Committee

3M's Standard Interchange Protocol (SIP, not to be confused with the "Session Initiation Protocol" used in internet telephony) is the protocol that library self-check units use to talk to the circulation module of a library's Integrated Library System. It's a small, focused protocol that was designed by people who: knew what they were trying to accomplish, had to implement it, and had to make it work on the low-end computers that get stuffed into the self-checkout units, and over the (potentially slow and crappy) communication lines that a small library system might have.

NCIP is the NISO Circulation Interchange Protocol. This protocol was designed as a "vendor-neutral" replacement for the 3M SIP. The standards committee allowed themselves to be seduced by both new technologies and mission creep that encouraged them to attempt to design a system that could handle all circulation activities, including not just patron self-check, but also inter-library borrowing and the accounting that goes with it. What they produced was... larger than one might expect, given the size of the documents.

For example, when a self-checkout unit wants to verify that a user is valid with SIP, it sends this message:

2300020060221 211723AOLochalsh Public\ Library|AAdjfiander|AC|AD|AY1AZEA14

When a self-checkout unit is using NCIP, it sends this XML:

<?xml version='1.0' encoding='UTF-8'?>
<!DOCTYPE NCIPMessage PUBLIC "-//NISO//NCIP DTD Version 1//EN"
     "http://www.niso.org/ncip/v1_0/imp1/dtd/ncip_v1_0.dtd">
<NCIPMessage version="http://www.niso.org/ncip/v1_0/imp1/dtd/ncip_v1_0.dtd">
    <AuthenticateUser>
        <IntiationHeader>
        <FromAgencyId>
            <UniqueAgencyId>
            <Scheme>local-name</Scheme><Value>Taylor-SC-1</Value>
        </UniqueAgencyId>
        </FromAgencyId>
        <FromAgencyAuthentication>xyzzy</FromAgencyAuthentication>
         <ToAgencyId>
            <UniqueAgencyId>
            <Scheme>local-name</Scheme><Value>UWOLS</Value>
        </UniqueAgencyId>
        </ToAgencyId>
       </IntiationHeader>
       <AuthenticationInput>
           <AuthenticationInputData>djfiander</AuthenticationInputData>
       <AuthenticationDataFormatType>
           <Scheme>http://www.iana.org/assignments/media-types</Scheme>
           <Value>text/plain</Value>
       </AuthenticationDataFormatType>
       <AuthenticationInputType>
           <Scheme>http://www.niso.org/ncip/v1_0/imp1/schemes/authenticationinputtype/authenticationinputtype.scm</Scheme>
           <Value>User ID</Value>
       </AuthenticationInputType>
       </AuthenticationInput>
    </AuthenticateUser>
</NCIPMessage>

Need I say more?

Appropriate Linking, Like Love, is in the Air

A couple of weeks ago Dan Chudnov commented (in #code4lib) that he didn't like putting Amazon.com links on web pages when he referred to books, because he's a librarian, and it doesn't feel right to be pointing to a commercial entity when the book is probably available at the local library. Similarly, Karen Schneider just this weekend felt conflicted about her blog book reviews, for much the same reason. Then, Lorcan Dempsey noticed that the COPAC experimental library catalogue provides COinS in the full record display.

While one might take issue with Dan and Karen's (apparent) distaste for the marketplace, a deeper point is that, just like those bibliographic databases that unconditionally provide links to online fulltext that may or may not be available to a particular user (*cough* PubMed *cough*), the link that is prefered by the creator of the content may not be appropriate, for any number of reasons, for the reader of the content. Even if Dan had no qualms about providing a link to, or maybe even deriving a benefit from affiliating himself with, Amazon.com, and even if I wanted to buy the book, that doesn't mean that Amazon.com is the best place to go. In fact, I just recently spent quite a bit of money at Amazon, but it was at Amazon.ca. But providing Open WorldCat links, as Karen suggested, isn't the right thing either, because (a) my local public libraries aren't members of OCLC and, (b) I might want to buy the book.

These incidents seem to emphasize yet again the importance of providing appropriate links for users. But how can Dan or Karen provide appropriate links for me, when they don't know what libraries, bookstores, or online retailers, I might frequent? Lorcan's post points in the right direction, however: content providers (and not just databases and library catalogues, but also bloggers, and other publishers of original content) need to begin providing metadata about the bibliographic units that they are discussing, so that the reader can decide what to do (and where to link to) with the metadata.

Openly Informatics (recently purchased by OCLC, coincidentally) provides a Mozilla Firefox extension that recognizes COinS and links to the user's own OpenURL resolver. Unfortunately, I don't think that's good enough for the majority of people. MPOW's OpenURL resolver doesn't, for example, provide links to amazon (of any nationality) or to the local public library. I need a personal link resolver; something that runs on the desktop that knows those places from whence I borrow or purchase books. My personal resolver should also provide a fallback, or "smarthost", OpenURL resolver, so that journal links that I run across can be passed along to a "full-service" link resolver, like the one that the university runs, so that I can find appropriate copies of articles.

In fact, I don't think I need a full link resolution server. I just need a shopping/library assistant: a firefox extension, much like Openly's, in which I could configure the online bookstores that I frequented and the libraries that I used, and which presented an OpenURL-style menu of searches based on the the contents of embedded COinS in web pages. Now I just need to find the time to write such a beast. And convince everybody else to start providing COinS.

The Librarians' Playlist

I've been working on a playlist for librarians for a while, mostly from my own collection. So far I've got

• "What do you do with a BA in English" - Avenue Q - Original Broadway cast recording
• "My baby loves a bunch of authors" - Bargainville - Moxy Fruvous
• "If I only had a brain" - Wizard of Oz soundtrack (for the undergrads)
• "Lobachevsky" - Remains of Tom Lehrer - Tom Lehrer (trust me on this one)
• "Too much information - Ghost in the machine - The Police
• "Private investigations" - On the night (Live album) - Dire Straits
• "Hats off to the stranger" - Sunny days again - Lighthouse
• "The Internet is for porn" - Avenue Q - Original Broadway cast recording