I think Dave Weinberger has unintentionally gone off on a tangent with an interesting post looking at what he calls a Semantic Behaviour Index, following Jon Udells' thoughts about a Google desktop.
Dave quite rightly gets nervous about privacy issues of software trackling our every keystroke, application launch and mouse click, but what he describes in his table of Semantic Behaviour strikes me as too broad and not in the right mindset.
His table looks like this
| TIME | WINDOW | ACTION | CONTENT |
| 10:23:13 | "Chapter1.doc" | Typing | And so we see I was right all along |
| 10:23:14 | Desktop | Opened browser | C:\Program Files\Mozilla\FireBird.exe |
| 10:23:18 | FireBird.exe | Typed in url | "www.wikipedia.org" |
| 10:23:21 | http://en.wikipedia.org/ wiki/Voynich_Manuscript | Typed in search form | "Voynich" |
| 10:23:30 | http://en.wikipedia.org/ wiki/Voynich_Manuscript | Copied text | "Over its recorded existence, the VMs has been the object of intense study by many professional and amateur cryptographers " |
But to me, only a couple of actions are "semantic". Opening a browser doesn't count, opening a document doesn't count, even typing a URL into the browser doesn't count.
However,
Adding to, or changing some text in the word document is a semantic behaviour.
Creating or clicking a link in the word document, an email, a bookmark or a web page is also
Selecting and copying some text in the browser window qualifies
Pasting that text into another document also.
All of these actions relate to the meaning of the words or other signs like images, in the documents. If I change some, or choose to follow a link or copy and paste text, I have, in some semantic way, related to the text. Now an index of these actions not only makes sense, it could both save time and increase the density of the connections among bits of information that would help to clarify their meaning and significance.
Whatever else it might do, I'd love to have a tool that, when I start work on a document, remember an email that dealt with the topic, open that email and click a link, copy some text from the website and paste it into the document I started with, then creates a set of footnotes and endnotes that tell me where each piece of information came from and, even if I eventually discarded the quote, adds the document to the bibliography.
I hold that no document exists in isolation, its language, its rhetoric, its topic and author's professional and personal history all converge on a node where, hopefully, some new meaning is created in the discourse. The right kinds of tools will help us keep track of the semantic pedigree of documents and their contents.
Its what I hope I'm going to find what I start beta testing JOn Husband's new information management tool
Comments