Date: Fri, 28 Apr 2000 15:22:40 -0700
From: | Eric Armstrong |
eric.armstrong@eng.sun.com Reply-To: unrev-II@egroups.com |
To: | unrev-II@egroups.com |
Subject: | WBI Wrong Direction for DKR |
SlashDot Style Invaluable for Intelligence |
In yesterday's meeting, I commented that gathering intelligence is the easy part -- it is the evaluating and summarizing that is most important -- so the intelligence community has a small army of people devoted to that task. This lends further weight to the concept that a SlashDot-style system would be invaluable for our purposes.
Much as I want to read a "Guide to Running to a NIC" it is manifestly clear to me that I have no idea on earth what I could possibly contribute to it. It's pretty clear to me that I don't know the first thing about it.
I'm glad there is an effort to pursue a direction, but I remain unshaken in my belief that WBI is fundamentally the wrong approach to the problem. Its a great tool to investigate, and undoubtedly has uses -- even within this project -- but as the fundamental component for the design, I'm afraid I simply do not see it.
The major points against it are:
As nice as it would be that HTML pages would be transcoded in a way that would let people link to individual paragraphs, the links would degrade over time as pages are moved or modified. That's the way the Web is today, true. But the problem would be exacerbated by an order of magnitude -- the average number of paragraphs per document. Even worse, since modifying a page would change the tags, any links I created might still be "valid", but pointing to the wrong location. The result would be gibberish as text I linked to or included inline from another document would be replaced by some other text in that document.
While an architecture built around WBI might be something we can learn from, I can only hope that it does not get promoted too widely, lest it have the same unfortunate impact on public perception that ThinkTank did.
Sincerely,
Eric Armstrong
eric.armstrong@eng.sun.com