Looking backwards at looking forward

I got a note today from a colleague at IBM Ecuador; he’s going to be making a presentation to a university next week about the Internet and Open Source, and while he was researching it, he stumbled across one of my old presentations and was wondering if I had a current version. I didn’t, but rereading it amused me enough that I thought I’d post it here.

So here is The Internet — Past, Present, and Future, as delivered to the American Intellectual Property Law Association in 1996 (with two changes: I’ve removed my e-mail address, and I’ve converted the presentation from Lotus Freelance Graphics to PDF; somehow, I think more people can read PDF these days than can deal with Freelance).

Avoiding deepest ignorance

Today was Day 2 (and last) of the 2006 Almaden Institute. I avoided the talks which looked as though they were grounded in deep neuroscience and biology — that left me two to attend.

The first was Beyond Dualism by John Searle of Berkeley. He talked about consciousness, free will, free won’t, intentionality, and subjectiveness — and that was in the first ten minutes. But, even though the talk spent most of its time in deep philosophical waters, I didn’t feel adrift.

And then I returned for the last talk of the day, on Consciousness by Kristof Koch of Caltech. This talk had demos, humor, and raised some interesting questions.

I stayed for part of the closing panel, but my phone rang and, after leaving the auditorium to answer it, I didn’t feel like returning for the last few minutes.

Instead, I went home for dinner and then a quick cache; I had hoped to be first-to-find it, but viperfin beat us by 17 minutes. There were two other brand new caches in the vicinity, but I hadn’t written down the details of either one, so they’ll have to wait.

High Order Ignorance

Today was the first day of the 2006 Almaden Institute. The past three years’ Institutes were very interesting, and two of them were even related to my work, so I was looking forward to this year’s edition. The topic is “Cognitive Computing”, and knowing that pretty much exhausted my level of knowledge going in to the first session.

There’s nothing like starting a conference out with a speaker who tells good jokes — and when he’s also a Nobel Prize winner, that’s a bonus. Gerald Edelman‘s talk, From Brain Dynamics to Consciousness: A Prelude to the Future of Brain-Based Devices, was heavy going at times (I had to go diving in Wikipedia to discover what terms like “qualia” meant), but I enjoyed those parts which I understood (which definitely included the jokes).

I didn’t expect to understand much of Henry Markram‘s talk, The Emergence of Intelligence in the Neocortical Microcircuit, but it turned out to be fairly accessible, and loaded with good graphics. His group is trying to simulate networks of about 10,000 neurons, with some early success — they need a BlueGene to do it, though.

I missed about half of the first post-lunch talk, Robert Hecht-Nielsen‘s The Mechanism of Thought, but came in during the discussion of confabulation as a mechanism to generate grammatically correct and plausible sentences. I think it’s a long way from that to Chancellor, though, and so did much of the audience — Hecht-Nielsen claims that it’s possible in the near future, with no major breakthroughs required (at least that’s how I interpreted him).

The second post-lunch talk was from Jeff Hawkins on Hierarchical Temporal Memory: Theory and Implementation, based on the work he describes in On Intelligence. It seemed plausible to me, but I do have a high order of ignorance in this area.

I skipped most of the panel discussion in favor of getting a little work done, but returned to the conference in time for the “moderate” walk to the water tower (I needed some exercise today), and then a tasty dinner and a very interesting talk by Rama on The Uniqueness of the Human Brain, which covered phantom limbs, synesthesia, and the emergence of language. His talk was far better than the typical after-dinner talk at a conference; I’m glad I stayed for it (I did, however, leave before dessert was served).

The Institute continues tomorrow, but I’ll probably only sample it instead of staying for the entire day. My inbox needs attention.

The Refill Detective at Work

Some time ago, Diane got a nice pen as part of a Thanks! award at work:

Free ibm.com pen

There wasn’t anything particularly special about the pen, but it felt nice in the hand and wrote well. For a while. But it ran dry awfully quickly, so I disassembled it this morning and found that it had a teeny tiny refill, about two inches long (the pen is five inches long).

Teeny Refill

The obvious plan was to visit the local office supply store and get a refill — but they didn’t have anything which fit. So then it was time for Plan B: Google.

But it was not obvious what I should search for. Since the refill said “Ministar E”, I Googled for that, but I didn’t find anything useful (many spam pages, though). Ditto on “Ministar pen”, “Ministar refill”, and “Ministar pen refill”.

Then I turned the refill over, and found intriguing markings.

Refill, showing ISO 12757-2 DOC

It looked like an ISO standard number, ISO 12757-2, so I Googled that, and found Dark Matter, which told me that ISO 12757-2 is the standard governing archival ballpoint pen ink. And that there was a related standard, 12757-1, which described the dimensions of refills.

A little more searching took me to the ANSI Store, where I could buy the standards: $53 for part 1 and a mere $40 for part 2.

I passed, although I did ask the librarian at work if we happened to have those standards — we didn’t.

So then I followed a sponsored link to the Colorado Pen Company, which offers many, many refills — and pictures, too. There are two which look similar to mine, one for Alfred Dunhill pens, and one for Lamy pens. The Lamy refill even has the magic phrase “ISO 12757-2” on it. But it costs five bucks, plus shipping, which seems like a lot to pay for one refill (though it is a full 4-1/4 inches long).

So I may do something entirely different. While I was chasing down the standard, I stumbled across this Digg, pointing to these instructions on using Mont Blanc refills in inexpensive Pilot G2 pens. Our local office store is having a sale this week — eight Pilot G2 pens (plus two bonus pens or pencils) for $6; they’ve got the Mont Blanc refills at two for $12. So for just a little more than I’d pay for one refill plus shipping, I could have what purports to be the world’s best writing pen. Hmmm….

Firefox cookie lesson

I ran into a problem caused by a internal site setting invalid cookies (ones with “@” as part of the cookie name) and tried to fix it by using the Firefox cookie manager to delete the offending cookie. This might have been OK, except that I’d also set Firefox to not allow a site for which I’d removed cookies to set cookies in the future.

And so removing the one errant cookie blocked all ibm.com sites from setting cookies — this had some unfortunate side effects (such as making it impossible to log into several internal sites). I couldn’t find any way to reverse this, either — not even running Firefox in safe mode helped.

So I resorted to rummaging through my profile directory and looking at any file which was human-readable. The last file I checked (of course!) was hostperm.1; that file had a line in it setting the “cookie” property of ibm.com to “2”. I deleted the line and my problems went away.

I will probably keep the Firefox setting to prohibit a site who’s cookies I’ve deleted from setting cookies in the future, but this posting will help me remember what to do if it causes me problems, too.