I’m writing this column in my local college library, which -- thanks to its comfy chairs and excellent Wi-Fi coverage -- has become a pleasant alternative to my home office. As I watch the students typing at the Dell PCs in the hallway, I realize that none of these kids has ever seen or used a card catalog. That’s mostly a good thing. But when I joined a group of librarians on a panel last month, I was reminded that something useful has been lost: a tradition of local annotation. This isn’t just old-fogy nostalgia. Librarians could talk to patrons through the medium of the library card, and although they weren’t supposed to, patrons could talk back to librarians — and to one another. It was a useful back channel that online catalogs could have supported. But because it wasn’t part of the official protocol, they didn’t.
The fuzzy intersection of official and unofficial data has never been a comfort zone for information technologists. In chapter 4of Klaus Kaasgaard's Software Design and Usability, Xerox’s Palo Alto Research Center (PARC) alumnus Austin Henderson says that “one of the most brilliant inventions of the paper bureaucracy was the idea of the margin.” There was always space for unofficial data, which traveled with the official data, and everybody knew about the relationship between the two.
There are lots of ways to incorporate marginalia into electronic documents — and that’s the problem. Microsoft Word supports comments. A number of OS X apps support the system-wide Sticky Notes service. But there’s no universal concept of a margin, and no standard way to embed marginalia that will always travel with a document. Henderson sees that as an essential requirement for systems that can flexibly evolve. “The system asks the user whether this marginal stuff, which is now carried electronically, makes a difference .... This is a step away, philosophically, from the business of figuring it out once and for all, and then letting the machines do it. Instead, we say: ‘The designers of the machines will figure out something, people using them will figure out more, and then the users and the machines together will actually do it.’”
The universal canvas won’t arrive anytime soon. But electronic marginalia do play an important role today. Consider e-mail. Messages received by my Outlook client are often annotated with headers like X-SpamPal and X-Spam-Flag. These experimental headers, tacked on to messages by my spam filters, help me keep my inbox clean. The use of such unofficial headers makes engineers nervous. RFC2822, the 2001 successor to RFC822 (the 1982 standard for Internet text messages), avoids all mention of X-headers and user-defined fields. But they’re handy critters that I doubt we can, or should, stamp out.
In the war on spam, marginal annotations are set to play an even more crucial role. The various sender authorization schemes now on the table — Sender Policy Framework, Caller ID, Domain Keys — stash new information into the Domain Name System. The technically correct approach would be to define new DNS resource record types. But there’s no way to get that done quickly and deployed widely. So instead, the extra stuff will be written into TXT records that can store arbitrary text strings. Purists will cringe, but I’m grateful that Paul Mockapetris, the designer of the DNS, left us the option to scribble in the margins of the system.