Paperless office? Only on paper

The "paperless office" is back in the news, nudged by statistics showing that we are consuming less paper than before.

Affluent countries such as ours use a lot of paper, about 531 pounds per person in 2000. By 2005, that had dropped to 502 pounds. Are we really, after so many false predictions, about to do away with paper altogether, in favor of the many advantages of digital media?

The answer is a resounding "no" in the short term, and a qualified "yes" in the long term.

Clearly, the ability to do away with mountains of paper carries huge business advantages, not to mention helping us lead more orderly lives in the home.

Those who love and collect books blanch at the thought, but ultimately there is a case to be made for digitizing that runs not only to convenience and utility -- digital books are easily searched and no problem to store -- but to environmental concerns in the production of paper.

But the move to paperless isn't going to happen any time soon, because we're simply not ready for it.

Paper is simple, computers complex. Our attempts at PC organization are all too personalized, lacking coherent standards, meaning the document I put in a folder today might be hard to find next month. I can run a search for it, but when I find it, I recall that I use OpenOffice, you use Word, and you'll lose some of my formatting when I send the file to you.

What gets in the way

Tags are an attempt to get around the search problem. Rather than worrying about folders, you simply add a few keywords to each item to be indexed. But your idea of the best tags may differ from mine. Experience with bookmarking sites such as del.icio.us has shown me that I often can't remember my own tag choices, much less anyone else's.

Can all these issues be overcome? Sure, but daily computing needs to be utterly intuitive before we abandon that old instinct to play it safe and print out a hard copy.

Positive changes loom on the horizon, to be sure.

Virtual computing software allows me to move easily between operating systems, surmounting the Mac/PC/Linux divide, which will one day be rendered irrelevant.

Also coming down the road is the "semantic Web." It shows hope of making information organizing more sensible by improving the computer's ability to categorize what we save.

Twine organizes data for you

A case in point is a still experimental Web site, Twine (www.twine.com).

Envisioned as a Web-based personal organizer with smarts, Twine is all about information organization and retrieval. Throw any digital documents you choose at it (e-mail, Web pages, personal documents, photos or news feeds), and Twine will scan it to extract useful items and build an index. In other words, you don't create the tags, Twine does.

Twine's tags are interesting because they rely on the software performing a semantic analysis of what it sees. This is fundamentally different from what a search engine does, because search results depend upon keywords. Twine can find a document that fits your concept even if the specific keyword isn't mentioned in the document. Natural language processing and semantic indexing is being used to show you the connections between any items you have saved on your PC.

We have to be careful with the term "semantic Web" because it's being overused, but in Twine's case, it's safe to say the idea is to process the meaning within documents.

The man behind Twine, Radar Networks' CEO Nova Spivack, likes to make the distinction between the Web as a place where pages are served up on demand and the Web as a database that offers ways of seeing the relationships between pages. We're nowhere near such a point, but tools like Twine are pushing the concept to see whether and how it translates into a business model.

The more you index, the more relationships Twine sees. It's a machine learning process that theoretically improves the more you use it.

Twine also includes a social networking component, allowing you to set up individual collections -- "twines" -- whose tagging and categorization are improved by the input of other users.

Ultimately, the right technologies, many of them being built now, will enable the Web to become more responsive by allowing machines to process not just "twines" but routine Web pages to extract meaning.

Twine is still being tested as these natural language technologies continue to be tweaked. The distance between its concept and a widely used finished product highlights why we're not ready for the transition to a paperless office.

Because despite wonderful strides in computer note-taking and organizing tools, we're still faced with the muddle of conflicting standards, operating systems, organizational choices and upgrade issues that make reliance on digital solutions alone a dicey proposition for average users.

Dominating my office are two big filing cabinets housing writing that goes back 30 years. I have everything from the last 20 years in digital form, but I haven't scanned the earlier materials, nor am I in any hurry to do so.

In writing this column, I printed out a number of background documents, scanning them by eye instead of machine and marking them with a yellow highlighter. The advantages of paper are still palpable and will be until computers get a whole lot smarter.