Centralization vs. innovation?

I've worked for:

• a large library automation consortium
• a county library IT department
• a state library

Recently, I've been working with committees for:

• a library automation consortium
• a statewide portal for state agencies/egovernment

Something that's ALWAYS bothered me, that I mull a lot, is the power of centralized vs. decentralized when building systems. When I worked for a library automation consortium, I was convinced that increasing the amount of centralization would enhance services. When I began working for a county system that ran its own automation platform, I began to notice how much more quickly we could react to the changes in technology than could a large consortium.

As Web2.0 became the trend, I noticed how some of the smallest, least assuming libraries managed to extend services or create more sophisticated web experiences than the big libraries. I wondered if the innovation happened closest to the ground, where the people building the systems were directly in touch with the experiences of the end users, good or bad.

I know that the architecture of the web is an enshrinement of the principle that decentralization ensures progress, innovation, and sustainability. Should a search engine create a better algorithm, the other search engines will try to improve on it. Should a node go out, that's fine, the net routes around it. There is no single point of failure. With Web2.0 the number of communication channels have increased. News often travels most quickly online via social networking sites - Twitter & Facebook, for example. So this argues for the power of a seemingly organic method of growing systems.

Governments build things in a more official, hierarchical, and centralized way. As a result, innovation is often stifled.

On the other hand, without centralization, we are constantly reinventing the wheel and duplicating efforts, are we not? Aren't there economies of scale in having everyone pool their money and bargain with the big vendors for the best possible setup that would cover all of the members of a project (whether we're talking about groups of governmental agencies or library consortia)? Wouldn't it be great if there were standards that users of multiple libraries or governmental agencies could count on when dealing with those entities?

An additional foil to a more tightly regulated, centralized system being the answer for library or governmental systems is the concept of monopolies of power. If the system is fully centralized, the power lies all in one place. There are no competitors. There is no ability to seek alternatives to the Dept. of Revenue Services, for example. In time, the relationships between the agency or library system and the customer grow strained. The customer has no options, no way out if the agency or library system isn't giving them satisfaction. It creates a very standardized experience, but not a user-centered experience.

But maybe that's part of the issue. If we could just center on the users' needs, maybe we would know how to build systems and services more effectively. The greatest successes happen where the maker's vision intersects with the users' needs & desires. That's where both libraries and governmental agencies want to go.

I'm open to your responses to this question of centralization vs. decentralization of systems and services. I've been wanting to post it for a long time and every time I envisioned writing about the issue, I was much more eloquent than this. But it's time to stop worrying about that and just get the discussion going.

Collaborative Content Teams

So we've done quite a few things to prepare for our content migration (from a static html site to a Drupal-based site).

I've had decent luck with a test run of the Import HTML Drupal module (though I could only test it on one of my laptop-based stacks because it has some PHP prerequisites that precluded running it elsewhere (on hosted sites), such as the HTML Tidy PHP library & XML/XSL support). I also had to first detach our dynamic web templates & get rid of the header & footer include file statements to ensure that I wasn't getting extra junk...

We've defined a high-level taxonomy that were turned into content teams. The taxonomy was gleaned from the hard work of our taxonomy team (with the much-appreciated help of volunteer card sorters who took the top 100 pages of our current site & sorted them into categories they felt were appropriate, then gave those categories labels. BTW, though I didn't run stats as to how much difference there was, the card sort exercises did clearly demonstrate how much deeper (than non-librarian card sorters) the librarians wanted to go into categorization. Not a big surprise, I suppose, but I think it helps us to see how our navigation schemes have gotten so convoluted.)

The content teams are now assigned with telling me which pages in their areas can just be imported in (vs. those that have to be rewritten or rethought).

I tried to exclude from the teams' lists the content pages which would lend themselves to the creation of new content types (using the CCK module) & the use of the Views module to assemble that segmented content in what will appear to the end-user as a page, but I've probably missed a few.

Currently, I'm trying out the Links Package module in response to a discussion that came up in the first of our content team meetings. I feel like the all of the discussion and feedback from our content teams is already helping us to hone our migration strategy. The product will be better tomorrow because of our debates and decisions today. And that's a key benefit of collaborative work.

The Centralization Conundrum

I've been thinking about this issue since I started working with library systems at a library automation consortium in 1998. The question is - which type of architecture works better and under which circumstances - one hierarchically dictated, centralized, large-scale system that provides for many different organizations - or a decentralized architecture, with each organization having its own individual system that meets local business needs.

The centralized system sounds like it would offer economies of scale. However, in many cases the centralized system is at such a scale that it costs exponentially more. The key advantage a centralized system really offers is the advantage that comes with centralized training and troubleshooting. One knowledgebase required, one set of experts who can solve all of its issues. So that's the real advantage.

The downside is that the centralized system wasn't necessarily built in a way that meets the needs of the individual organizations who've been sucked into it. The alternative - many little systems, one for each organization - allow that organization to customize things to meet their "on the ground" needs, to do real user-centered design. That's a level of flexibility and adaptability that centralized systems just don't offer. But it requires greater technological fluency on the part of each organization. The good news is that we're all more technologically fluent than we used to be and as new players enter the workforce, we grow still more fluent.

Then there is the issue of point of failure - a centralized system offers one. That's why the internet was built the way it was - decentralized. So it could continue functioning at some level no matter what.

I think about this issue of centralization v. decentralization a lot in terms of e-government. In an era when the technology for web-based services is so easy and cheap (or free), why can't agencies offer web-based services and information that are more usable and accessible? Take a look at this great article on e-government & you'll find a blueprint for how we could make government more technologically adept and user-friendly.

The old command and control model is hierarchical, centralized, and NOT web-speed. There needs to be a new model. It needs to focus on creating frameworks that allow the individual entities to do what they need to do to succeed with their customers. It needs to focus on ensuring collaboration, education, communication, interorganizational-compatibility, and usability for all stakeholders.

Putting Citizens First

I’ve been watching the #gov20 Twitter stream since the Government2.0 summit last month (the few times I actually feel like I have a moment to look at Twitter, that is!) Through the stream, I learned about a great white paper called “Putting Citizens First: Transforming Online Government” (PDF) that was put together by the Federal Web Managers council.

What struck me about it most was that you could pretty much exchange the terms "government" and "agencies" with "library" and "libraries" and the paper would often be right on the mark.

A case in point:

… promise the American people that when they need [library] information and services online, they will be able to:

• Easily find relevant, accurate, and up-to-date information;
• Understand information the first time they read it;
• Complete common tasks efficiently;
• Get the same answer whether they use the web, phone, email, live chat, read a brochure, or visit in-person;
• Provide feedback and ideas and hear what the government will do with them;
• Access critical information if they have a disability or aren’t proficient in English.

Further:

Establish Web Communications as a core [library] function

One of the biggest problems we face in improving [library] websites is that many [libraries] still view their website as an IT project rather than as a core business function. Many [library] websites lack a dedicated budget. Only a minority of [libraries] have developed strong web policies and management controls. Some have … “legacy” websites with outdated or irrelevant content. With limited resources, many find it difficult to solicit regular customer input and take quick action to improve their sites. While there are many effective [library] websites, most web teams are struggling to manage ...

...[Libraries] should be required to fund their “virtual” [community] space as part of their critical infrastructure, in the same way they fund their “bricks and mortar” [community] space.

[Libraries] should be required to appoint an editor-in-chief for every website they maintain, as do the top commercial websites. This person should be given appropriate funding and authority to develop and enforce web policies and publishing standards, including ensuring that prime real estate on [library] websites is dedicated to helping people find the information they need.

[Library management] should develop standard job descriptions and core training requirements so [libraries] can hire and retain highly qualified experts in web content and new media—not just IT specialists.

And there are other points as salient for libraries as they are for other governmental agencies, such as these report sections/headings:

• Help the public complete common [library] tasks efficiently
• Clean up the clutter so people can find what they need online
• Engage the public in a dialogue to improve our customer service
• Ensure underserved populations can access critical information online

The whole report is worth reading (and it's pretty quick to do so, at just 4 pages), so I won't elaborate on those sections here. Overall, it left me feeling that we (governmental agencies and libraries) ALL have a lot of work to do.