Equal Time

In a recent post I dissed academic.live.com. It's now working for me, more or less, though there are some strange things that might either be intentional and bad, or just problems with Firefox on Linux. In particular, scrolling in the internal panes is erratic, and tends to take very large jumps. I do like the multi-pane interface that pops up abstracts and other info, but I think the screen real estate isn't used very efficiently, and it means you wind up having to scroll in multiple windows, which is clumsy. Moreover, the database itself is a tad erratic; sometimes, journal names or volumes or even years are missing, which would be a fatal flaw without good links to the documents. Still annoying, as I'm often looking for either recent or old papers, and it's hard to know which a given paper is without downloading it and opening it.

And I still want to know how they actually rank things.

Since I'm a big fan of Linux and free software in general, I have to rant: OpenOffice 2.0 on Fedora Core 4 (running in Japanese mode) still crashes on me on a regular basis. Often without a message or warning at all, it just blips and disappears. Recently I was working on a presentation, and it would drop out on me every 20-30 minutes. Given that file saves can take a long time, and restarts are slow, this means I lost a lot of time. Finally, things deteriorated to the point where it crashed every time I started it, without a message at all. I spent an hour trying everything, digging through the OO registry XML files and everything...then discovered that I was out of disk space. Geez. Stupid situation, to be sure, but OO can't figure that out and give me a warning?

Setting the Standard

Digging for something else in my bibliography file, I ran across my entry for C.A.R. Hoare's "Communicating Sequential Processes", CACM Aug. 1978, and I wondered, "How many times has that been cited?" Well, according to scholar.google.com, 6027 times, so far.

That's the standard, then: a world-changing paper in CS gets cited about six thousand times in a little under thirty years.

I'm idly wondering what might be cited more times. Van Jacobson's "Congestion Avoidance and Control" might be the top networking paper, I'm not sure; but it exists in several forms, complicating things a bit. Scholar cites it as 1995, when it appeared in a collection, and claims a bit under 3,000 times for that.

Anybody either know what the top-cited paper is, or want to take a shot at one that beats CSP?

Thought-Controlled Robot

Now you can lie in your nice, comfortable MRI machine and control your robot. Or at least its hand. According to an article in the Japan Times, Honda and ATR have developed a system that can read MRI signals well enough to distinguish among a fist, an open hand, and a V sign in a few seconds, and order a robotic hand to perform a similar movement. (Yes, that's "gu, choki, pa", or "rock, scissors, paper".)

This is an improvement over other techniques for thought-controlled devices, which often require electrodes actually implanted in the brain, and/or require significant training of the user. The drawback is that it requires an entire MRI machine :-). I have no idea what this would do for Stephen Hawking, depends on which of his neural functions have deteriorated.

ITRS Emerging Research Devices

I was reading George Bourianoff's The Future of Nanocomputing, from IEEE Computer, and it led me to a section of one of my favorite references, the 2005 edition of International Technology Roadmap for Semiconductors, titled "Emerging Research Devices". 200+ references on carbon nanotubes, quantum cellular automata, etc. -- numerous alternatives to standard CMOS for achieving classical computation. On quantum computing, which it calls "coherence quantum computing" to differentiate it from classical computing performed with quantum effects, it defers to the ARDA Quantum Information Science and Technology Roadmap. All worth a look.

Molecular Tapas Bar

I sometimes learn about things in Tokyo via circuitous routes. The Daily Yomiuri has a regular section on Sundays that's produced by the Times of London. Today's issue had a column by Richard Lloyd Parry. His rhetorical goal is to show that Tokyo is getting its mojo back, like the days of the Bubble. Parry mentions people eating sushi off the bodies of naked beauty queens during the height of the Bubble. Personally, I suspect that's some sort of urban legend morphed from a scene involving live shrimp in Itami Juzo's Tampopo, though I could be wrong.

Parry's column is primarily about dinner at the Molecular Tapas Bar at the Mandarin Oriental. Food from test tubes, by chef Jeff Ramsey. If Tokyo can support food this challenging and decadent, it's back, right?

Food from test tubes? I am so there. While it doesn't sound quite as avant-garde Homaro Cantu's Moto in Chicago, which I haven't had the pleasure of trying yet, this will definitely be an interesting experience.

Our reservations are for July 2nd (they weren't as hard to get as Parry would have you believe). Stay tuned for a report.

P.S. Parry mentions longing for a MOSburger, and compares it to McDonald's, but in fact MOSburger was started after the owner-to-be visited Tommy's.

4.8 in Southern Chiba

No big deal, at least here in Abiko. Not quite all the way to the bottom end of the Boso Peninsula.

Tommy Turns Sixty

While we're on the subject of food, Tommy's turned sixty, according to an article in yesterday's L.A. Times. Events were complete with Elvis impersonators and lots of people with personal stories of arduous treks and extraordinary effort to reach the Land of Milk and Honey (or, chili burgers and tamales).

My flight to Japan the first time I moved here (March 1992) was on a Saturday morning. A group of about twenty of my friends (including at least one vegetarian) met at Tommy's for breakfast before seeing me off to the airport. Somewhere, there are photos...

And somewhere, I still have my Tommy's Fortieth Anniversary t-shirt.

Space Kimchi

Oh, I love this. Kimchi in space! Man, I'm drooling already. Sign me up - I love kimchi.

But...must...resist...making bad joke...willpower...noooo...I give: ISS astronauts will now be rocket powered!

From the article: "Space kimchi is expected to be of great help in stimulating astronauts’ appetite with its zest and spices. In addition, it is effective in promoting the intestinal functions, which tend to be somewhat sluggish in space, with abundant fiber." You got that right!

Irradiated kimchi...wow...

Academic

Those Microsoft guys, what pranksters! Always out for a laugh.

There was a short blurb in Science recently about academic.live.com, a new search engine Microsoft has created for academic journals. Since I'm such a fan of scholar.google.com, I figured I'd give academic a whirl.

But I don't use Windows, only Linux (cue "Jaws" theme). I went to the site, and it popped up a nice search window, with FAQ below and other tidbits. I typed in something, hit search, and it moved to "Loading..." and never left that state.

I thought, "A website that doesn't work with my Linux box! From Microsoft! How unusual. Let's see if it works on Windows..." so I turned to my neighbor, who does use Windows (the Japanese version of XP), and got him to try it. Enter something, hit search, and...it hung his browser! Not even the kill button in the toolbar works. Can't iconify or background it. At least the task manager managed to kill it, though.

Just made my day...

P.S. From their FAQ:


How do you determine relevance? Are you using citation counts in the relevance ranking?

We are determining relevance based on the following two areas, as determined by a Microsoft algorithm:
# Quality of match of the search term with the content of the paper
# Authoritativeness of the paper.


I can kind of imagine what "quality of match" means, but what's "authoritativeness of the paper"? They explicitly rule out citation count (claiming that an iffy cite count is worse than no cite count), but never explain what they do use. Maybe a journal impact factor, like the one at CiteSeer? Inquiring minds want to know...

Carbon Nanotube Origami

The Daily Yomiuri has a short article (essentially cribbed from a Japanese AIST press release) about progress in manufacturing single-wall carbon nanotubes (SWCNTs or SWNTs). The article is actually a little hard to read, but it seems they've increased the efficiency of production of high-quality nanotubes by a factor of 100, for some measure. They also claim yield is up from 50% to 97.5% for defect-free tubes, though I don't see anything about the length of the tubes...wait, apparently 0.4 to 50nm in diameter, and 1um to several tens of microns long.

But what's cool enough to warrant a blog posting here is that they made a sheet of SWCNT material, 9 microns thick, big enough to fold into an origami crane (tsuru). The print edition of DY has a photo of it sitting on someone's hand, but the photo isn't in the online article. The press release has a different picture, but no scale; it's actually roughly "normal" size -- probably folded from a sheet at least 10cm by 10cm.

Now Available: Architectural Implications of Quantum Computing Technologies

My paper with Mark Oskin, "Architectural Implications of Quantum Computing Technologies," is now available from the archives of ACM Journal on Emerging Technologies in Computing Systems (JETC). It appears in the Jan. 2006 issue.

Quantum computing researchers are familiar with the DiVincenzo criteria which a technology must meet in order to be a candidate for functional quantum computing devices, and with the notion of an error threshold below which a technology meets the basic mathematical notion of scalability. This paper is the view of two computer systems engineers/researchers on the practical issues that will determine how easy or hard it is to build a large-scale, high-performance quantum computing system.

There is more news about my personal research, but that will have to wait for later postings, as I still have many thoughts to organize...

Alex @ Caltech?

Alex Doonesbury is considering going to Caltech. We're all waiting on pins and needles to see what her final decision is. Surely there are some pranksters at Tech who are capable of, um, influencing her decision?