So why is there only one person running it?

A certain large local public library system has that one poor systems librarian doing the work of 8 people:

• managing the online public access catalog,
• handling the integrated library system,
• managing the web server,
• managing the database vendors and databases,
• developing web applications that interface with those databases,
• coordinating the content authoring workflow,
• editing and uploading site content,
• and managing one web editor and a part time graphic artist.

This valiant librarian is working 80+ hours per week, and the redesigned website that I and my fellow consultants have provided her will require even more of her already-scarce time to truly realize the library administration’s vision.

Think about it. To run even the smallest library branch, it requires the coordinated teamwork of several people. But the library’s website, no matter how good the information architecture, content management system, or design, cannot survive on one person’s effort. It receives many more visitors than even the busiest downtown branch of the library system. It answers far more questions per day than the busiest reference desk. It is seen by more passersby and reaches more potential patrons than the most beautifully designed or most visible library building. Satisfying the needs of all those users requires considerable effort and attention — far more than one person can supply.

Pursuant to my previous post “On Systems Librarianship“, I will be moderating a session at the annual meeting of the ASIS&T PNC on the role of Systems Librarians and Information Architects in getting people to work together. You see, the theme of the conference is “Building Bridges: Overcoming the Barriers To Data Interoperability,” and the call for proposals asked for research and reports on how information professionals were enabling the smooth transformation of data from one system to another.

In my opinion, that’s mostly a solved problem, given the wide variety of computer programs and routines that are all tuned specifically for processing data — text, numbers, images, binaries, and so on — but what is NOT solved is how to transmit meaning from one context to another. You could have the best data-sharing protocol or content management system in the world and still fail in the end due to miscommunication between the techies and… well, everyone else. For instance, look at the word “groupware”. To some, it means videoconferencing, shared drawing, collaborative VR, MUDs and MOOs, and other fancy technologies that are “virtual” analogues of real-life interactions. To others (particularly IT departments), “groupware” means shared calendaring, email, and discussion boards. Now, say the CEO or library director passes down an edict to the IT department that they need to support “groupware” ASAP. Imagine the derisive snorts and invective each camp will direct at the other when they finally realize, after weeks of confusion, what the other really means. Think of the irreparable damage to human relations between the administrators and developers that will cripple progress for years to come. Who would care then if MARC records could be seamlessly converted to XML-RDF?

It’s a human problem that requires a human solution. And it’s precisely where people like IAs and IT librarians come in. We need people who have a systems¹ view, who can see the big picture, and who are bilingual (or multilingual!) in the jargon of technology AND business. Who better to fill that gap than those who are skilled in discovering user needs and translating those needs into designs and guiding those designs through to development and implementation?

This is the topic of the panel I’ll be moderating at ASIS&T PNC 2005. The abstract follows:

[UPDATE: Yep, your registration fee for the conference will be covered. Here's the link to the chapter web site again.]

¹By “systems” I mean it in the sense of holistic “systems science” or “ecosystem” or “systemic”, not “IT Systems”.

Thankfully, there’s a couple good CMS comparison web sites out there:
http://www.cmsmatrix.org/
http://opensourcecms.com/

Moodle is a open source course management system that works pretty much the same as any other groupware system. It’s PHP-based, and the only major differences from other PHP-based systems like phpGroupWare or PHProjekt, besides being geared toward teachers and students, is that it actually works and has decent documentation.

I installed it and got most parts working in record time with little hacking required.

As I sit here in my office staring at the whiteboard, which is covered with diagrams and flowcharts hastily drawn to illustrate the flow of information from user to system to system and back again, I reflect on why this job exists at all. Why do we need librarians in IT departments anyway? Shouldn’t the librarians focus on reference and classification and selection and all those other bookish concerns? And shouldn’t computer support personnel just stay in their fortress of circuit boards and techno-jargon, honing their 1337 h4×0rz sk1ll5 and preventing the script kiddies and cyber terrorists from taking down the network?

Well, in answer to these questions, I offer the following illustration from this Sunday’s Dilbert: http://dilbert.com/comics/dilbert/archive/images/dilbert20050101046179.jpg

Notice that, in this comic, the IT department is presented as sadistic and needlessly legalistic, a stereotypical portrayal of the stereotypical computer support department. The reasoning behind the confiscation of “non-standard” equipment is that, in order to save money and staff time, the organization can only support a limited range of technology. Every boutique piece of hardware or software adds yet another expense, yet another line item, yet another headache on top of the already thin budget and overloaded schedule tech departments must handle. Such a perspective is technology and cost-centric. And everyone knows the bottom line — especially in libraries — is king. Right?

Well, let’s look at the stereotypical portrayal of librarians: maternal, dour, overly-educated-yet-technologically-challenged protectors of arcane knowledge. I have met several librarians who, at first glance, fit this description perfectly. However, being a librarian, I know that even the most stereotypical librarians care deeply about helping and educating the public. Yes, we are keepers of the sacred word, but we’re here to assist real people in finding real answers to real questions. We are user-centric.

So what happens when you have a library, full of librarians and library staff, that also happens to be full of computers — public terminals, staff workstations, printers, scanners, catalog servers, web servers, application servers, file servers, backup servers, database servers, print servers, routers, switches, wireless access points, etc. etc. etc.? You get an IT department to manage all that technology on a shoestring budget. You get aging reference librarians who are 5 years from retirement being asked all manner of tech support questions by patrons. You get stereotyped user-centric librarians criticising stereotyped computer geeks for being technology-centric and budget-conscious. You get civil war.

UNLESS… this is the point of this whole spiel… unless you have Systems Librarians there to fill the gap between users and technology, between the IT department and the librarians. They bridge that crucial disconnect, providing the user-centered perspective to the bit checkers and bean counters while keeping the limitations and challenges of technology management in mind.

Systems Librarians are bilingual, translating techno-jargon into natural language and back again; they are cyborgs, chimaeras, the moderators between the binary of circuitry and the poetry of meat. They are catalysts of change. IT Librarians bring a holistic perspective of the operation of the library to every interaction and can see possibilities for improvement where non-technical library staff or non-librarian technical staff would see only barriers and the fog of indecipherable complexity.

Thus, if I often switch abruptly in this blog between apparently disparate topics such as LDAP and library committee formation, know that it’s just part of my job as a Systems Librarian. I’m just exercising my babelfish.

It’s looking more likely that we will go with a best-of-breed approach. In this scenario, we would select separate packages from separate vendors and glue them together with some common interface. Microsoft provides products for such a scenario, allowing one to tie together Office, Outlook, SharePoint, Content Management Server, SQLServer, and Server together with Exchange. Unfortunately, licensing fees, disdain for MS products, incompatibility with everything else, and security concerns have kept us from selecting the Microsoft route — except as a last resort.

However, a couple open-source alternatives to MS Exchange exist:

• SUSE OpenExchange (aka OPEN-XCHANGE)
• exchange4linux

Both currently use WebDAV and XML as an interface for communication between components of a groupware system.

Due to the needs of the groupware project I’m working on, I’ve been exploring the world of LDAP.

Unfortunately, I’m trying to query our Windows Active Directory to discover the LDAP attribute names for use in a PHP script — and the Microsoft documentation for the LDAP interface to Active Directory is cryptic and completely unhelpful. I’m sure what I want to know is in there, but I don’t have time to read it all. Thankfully, some nice people have written some more useful guides to the LDAP attribute names in Active Directory.

Especially of help is the ldapsearch command line tool, which allows one to query an LDAP server and get all kinds of useful user, group, and permissions information. Well, at least it allows one to do it if one knows what one is doing. Through trial and error with the syntax, I figured out a Linux command-line query that would allow me to see the attributes available in the LDAP-ified Active Directory:

"objectclass=*"

Check out the ldapsearch manual pages for an explanation of what all those flags mean. Once I can get our PHProjekt installation to map the resulting LDAP attributes to meaningful username and address book entries, we’ll be in business.

• Why do we need groupware anyway?
• Who are the stakeholders?
• Who are the end users?
• What do those users need?
• How are they meeting those needs now?
• What constraints (technological, budgetary, etc.) are we working with?

I should also know the who, what, where, when, and how of this project and its possible implementation. In a large library with tens of thousands of users, all this should be figured out by a task force. The committee should ideally be made up of:

• representatives of the end users (staff)
• someone from Administration
• someone who does computer training/support in the library
• someone who’s studied the organizational problems in our library
• someone from Systems (like me)

This process should be typical in any major IT project in a library. At this point in our process, it’s not clear whether we’re really going to commit to a groupware solution. To gauge need and interest, we’ll set up a prototype system as a “straw man” and present it to the administration. If they feel it’s worth spending time and money on, we may form a groupware task force to figure out which vendor meets our needs. The task force would need to write up an RFP, schedule demos by the vendors, collect proposals, evaluate proposals, etc. etc.

But first, let’s just focus on selecting our “straw man” prototype. Like I said back in my first post on this subject, I’d like to create a feature comparison table to track which system does what. Unfortunately, most commercial groupware vendors are less than forthcoming about the specs of their systems in order to protect their intellectual property. Thus, we’ve decided to evaluate open source packages only. An added benefit is that we can really get into the guts of open source systems and figure out how they’re built.

Our development server is running Linux and PostgreSQL, and our university’s authentication service is LDAP-compatible. We also need a system that is being actively developed (i.e. updated in the last year). Any system that does not meet these minimum requirements is ignored.

So, without further ado, here’s the resulting groupware comparison matrix (Excel spreadsheet). You’ll notice that the vendors are sorted according to the number of results in a Google search. It’s a rather arbitrary metric, but it’s useful for seeing how popular (or controversial) a system is [Credit goes to Bill for this idea]. Although there was no clear winner, the system we will probably go with is phpGroupWare. It’s lightweight, supports a good number of critical features (PDA syncing being the most critical) and has the added bonus of running under Apache, so we don’t have to run another server environment.

[Update 12/21/2004]

Well, after playing with phpGroupWare, I’ve decided it’s too unstable and disorganized to implement. Plus, recently discovered security vulnerabilities make me wary of considering it as an option. I’ve decided to play with PHProjekt instead. There’s a possibility the PDA syncing could be made to work with it. Maybe…

So, as you surf the web, be it for work or play, you review the sites you like (or don’t like) with those toolbar buttons. And, on those days when you just want something new, you can use StumbleUpon to explore new sites in any of the topics you’ve chosen. There are some spyware concerns with StumbleUpon, but none more than Google Tooolbar. I’m not too worried about it — yet.

The thing that fascinates me about this tool is the great potential for a kind of “readers’ advisory” or Amazon Recommendations for the public Web. It also allows some degree of personal knowledge management, as you can add a review of any site you come across. Those reviews are shared with others, but you can also view the collection of all your reviews in one place. Even without this aspect, the signal-to-noise ratio has the potential to be very low, so information encountering and foraging behaviors can become very efficient.

One gripe, though… they don’t have a “Society > Libraries” topic, but I sent in a suggestion to have it added.