Okay, I am going to milk my 15 minutes of fame as an E 2.0 “influential” to pitch you some pure vaporware. When I am not starting flame wars around E 2.0 culture change, I manage a research team within the Xerox Innovation Group, that is building a technology called Xerox Trails. The technology allows you to blaze and follow “trails” through Web content. Right now, the consumer incarnation of the technology, a product called “Trailmeme,” is in limited invitation-only beta. Read on for an invite code. What I’d like from you E 2.0 evangelists and champions is help brainstorming and dreaming up the ideal enterprise version of this technology, which is on our roadmap for a year or so down the line. At a higher level, I am interested in discussing a more conceptual question: how do you make sense of the huge mess of documents on a typical Intranet, hosted on multiple internal sites and technologies? This is the problem of enterprise document integration (EDI).

The Problem of Enterprise Document Integration

A typical large-scale Intranet is a cheerful mess of Websites running on different platforms, often from entirely different generations and technology paradigms. Within Xerox for instance, we have an ancient 1.0 level default Intranet (sorry any Xerox IT guys reading this, but you know it is true!), several repositories running on our homegrown Docushare CMS, some Sharepoint installs, and of course, a whole bunch of special-purpose local workgroup sites running on systems ranging from Confluence to Mediawiki. This is as it should be — a few slowly-evolving sensible enterprise defaults, complemented by some faster-moving local solutions for local problems. Besides this general content management infrastructure, like every large corporation, we naturally have all those specialized platforms for functions like elearning and HR.

This information environment though, doesn’t lend itself to easy organization and re-organization around changing and often transitory purposes. How, for example, would you pull together all the online resources needed to help employees from a newly acquired company orient themselves? Or give your employees an easy way to explore all the information related to a particular market?

The traditional 1.0 answer (which we know does not work) is “portal.” Portals are simply too hard to set up, too slow to adapt, and too centralized (I often rant about “Portalitis” — multiple silos each wanting to centralize information around themselves, leading to a proliferation of portals with a lot of redundancies). In reality, valuable information exists all over in a distributed way, and is constantly shifting.  Most of us are reduced to emailing bunches of cut-and-paste links and attachments to each other for every situation.

This is the problem of enterprise document integration, EDI: The problem of rapidly assembling and sharing an organized set of online resources to suit a given situation.

I don’t claim to have the complete, definitive answer, but I do think the idea of a “trail” is one potential solution. So here’s my quick soapbox pitch, before I get back to discussing the broader EDI problem. Keep in mind that what I am describing below is the simpler consumer version that we already have out there. The stuff isn’t yet ready to take on full-fledged EDI.

What is Xerox Trails?

Trails allow you to collect, organize and present (in short, “blaze”) collections of Web content using the user experience metaphor of a “trail” (as in hiking). You navigate a trail using a “Trail Map” view and a “Following” view. Show is easier than tell, so here is an example trail I created on our destination site, trailmeme.com, about the “Social Media vs. Knowledge Management” debate I triggered a year and a half ago, organizing the original article and some of the most interesting responses. Here is a screenshot of the “Trail Map” view of the trail (Click here to go to the live, clickable map: you’ll need to have Flash enabled in your browser).

And here is the “Following” view (think of it as a “ground-level/street-level” view). The idea is to support navigation with the sorts of signs you might see on a highway or a hiking trail at the ground level, by framing the original source with a useful navigation overlay. Click here to start “following” the trail.

This is the basic user experience. We also have a plugin that implements this experience on WordPress blogs (our project blog has this enabled), and we have a similar plugin for MediaWiki in the works. These allow you to create the “trails” user experience on your own site. Both are open source (GPL) and if you are interested in supporting trails on other platforms, talk to me.

And of course, since we are Xerox, there is a print and publishing angle to all of this. Trails (when appropriate) can be sequenced and converted to instant PDFs with one click (currently only supported on the WordPress and MediaWiki plugin version).

So that’s the technology itself.  Now let’s get back to the EDI question.

The Varieties of EDI Problems

Here is a quick half-dozen variants of the EDI problem that occur to me, that frequently appear in the enterprise. One of the big things I am interested in hearing from you guys is: are these the right EDI use cases? Are there better/more important ones?

1. A bid team for a major and complex enterprise sale collaborating to rapidly assemble existing resources into a raw material package to support the proposal writing
2. An HR/eLearning department creating an ad-hoc, just-in-time “course pack” out of stuff already scattered across the intranet, to support a course on the eLearning system
3. A workgroup involved in competitive analysis gradually putting together a “market view” picture by putting together internal material, analyst reports, press releases, competitor news and so forth, into a sort of constantly evolving online intelligence dossier
4. Creating an ad-hoc org chart out of wherever people choose to post their Intranet (or even public) profiles (even for companies that have a single social network/expert locator system, the ordinary model of organizing with tags and a couple of matrix dimensions reflecting the static organizational structure, rarely reflects the real org chart of ad-hoc teams, workgroups, dotted-line reporting and so forth)
5. Creating a weekly or monthly newsletter out of press coverage of your company
6. Creating a collection of information for a new “community of practice” type group, without creating yet another silo with redundant copies of information

Some of these problems, of course, can be solved with other technologies, including internal social bookmarking or collections of RSS feeds in a truly SOA-enabled Intranet. Where trails add value is where there is a need to add some meaningful sequencing (“read this before that”), so that the collection becomes a guided narrative of sorts. Some questions that my research team and I are interested in include:

1. What are the most critical EDI problems and pain points?
2. When the existing solutions to these problems work poorly, how do they fail?
3. Lost opportunities: what problems are being finessed or not faced at all, due to the difficulty of the problem?
4. How transient are information organization needs? How much organization lasts for days, versus months, versus quarters, versus years?
5. How dynamic are information collections, how often do they change during active and quiescent phases of the respective workflows?
6. What sort of security and permissions architecture constraints impact EDI?
7. What regulatory and “need to know” constraints impact EDI?

Invite Code and an Invitation to Dream

One of the smart things Xerox does is organize “dreaming sessions” between researchers and customers (who otherwise rarely meet).  That’s kinda what I am hoping to create here, virtually. If there is enough interest and I am able to scrounge together a budget, we might even do a face-to-face dreaming session at some point.

So here you go, here is how you can experiment with the technology in its current consumer incarnation:

• Go to the login/registration page of trailmeme
• Request an invitation to register with the coupon code: e2blog

You should see the invitation in your email in 10-15 minutes. If you don’t, check your spam folder. Once you register, you can create your own trails. The code will work for the first 100 users who sign up. We’ll release more batches of invites soon.

Keep in mind that this is NOT a commercially launched Xerox product. It is a labs project that is in research beta, so yes, some things (okay, several things) are kinda klutzy at the moment. But I hope you will be won over by the sheer coolness of it all.

There is a FAQ, and quick-and-dirty screencasts to help you follow/blaze trails. You can keep up with our updates  on the project blog at blog.trailmeme.com

So play around, create some trails, and feel free to post any comments/feature suggestions/high-level conceptual thoughts on EDI, anything you like. You can also email me directly at venkatesh.rao@xerox.com if you want to share private comments or just want to connect and talk about this.

If you are itching to use this inside the firewall, and you happen to run WordPress inside your org, you can of course try the plugin right away.

And I’d appreciate any reblogs. Feel free to share the code too.

Venkatesh Rao is a researcher in the Xerox Innovation Group. His personal blog is at ribbonfarm.com. While his posts on this blog usually come with the disclaimer that the opinions are not that of his employer, in this case, that is obviously not true.

Comments RSS