Slashdot Mirror

just look at that GA tutorial.

It's a 4MB postscript , but it's worth reading. GA and Simulated Annealing are quite different. the main difference is that GA use a population of trial solutions, and crosses them, with SA use only one trial solution at a time.
If i recall well, there is also a difference in the convergence: it is proven that for a infinite time, SA will converge toward THE global optimum, while there is no proof that GA will converge, even toward a local optimum.

just look at that GA tutorial.

It's a 4MB postscript , but it's worth reading. GA and Simulated Annealing are quite different. the main difference is that GA use a population of trial solutions, and crosses them, with SA use only one trial solution at a time.
If i recall well, there is also a difference in the convergence: it is proven that for a infinite time, SA will converge toward THE global optimum, while there is no proof that GA will converge, even toward a local optimum.

The Web100 Project is working on putting automatic TCP tuning into the stack. This will allow a TCP connection to use all of the available bandwidth, without breaking any of the internal algorithms or stomping on other connections. It is already possible to tune most TCP implementations by measuring the bandwidth*delay product and tweaking the socket buffer size; the NLANR TCP Tuning page has instructions.

I thought it was hard the first time I typed on one of the cheap modern keyboards that provides very little resistance and no click. Typing without a keyboard is taking that to an extreme. And I just started paying more attention to the way I type. My junior high school typing teacher would be appauled. From what I can tell, I use three finger and my thumb on my left hand and one finger and my thumb on my right. Not to mention frequently stretching them. I do wonder how it will interpret odd typing styles.

Then, of course, there is the whole issue of how well it can discriminate chords. I use Emacs, The One True Editor (C-0 M-x all-hail-emacs), which is well known for some of the secondary meanings of its acronym including "Esc Meta Alt Ctrl Shift". We just express it more compactly as M-A-C-S-. Humor aside, will I be able to type M-C-v or C-@ or other three key chords with ease?

Fact 1 -- Deja News is in the Echelon building:

Deja News, Inc.
9430 Research Boulevard
Echelon II, Suite. 350
Austin, TX 78759

Fact 2 -- Cycorp makes what are arguably the best tools for scanning the web for concepts.

Fact 3 -- Cycorp was a spinoff of MCC.

Fact 4 -- Deja News, Inc., Cycorp and MCC are within walking distance of each other.

Fact 5 -- Bobby Ray Inman was the first director of the MCC.

Fact 6 -- Bobby Ray Inman is a spook's spook.

I may be a bit biased here since I was invited to go to work at the MCC when it was in its early formative stages (before Austin had been selected). My office was, at that time, at Arden Hills operations at Control Data Corporation, just two stories above about an acre of supercomputers that had signs hung on them that read "Fort Meade".

As Seymour used to say to the "insurance" agents located at the "Thorp Insurance offices" out in the middle of the corn fields near his farm where his tribe was building the Cray-1:

"Just don't let my people know you're here."

In my (informed) opinion the overwelming consesus amongst climatologists and biogeographers is that climate change is real - and this is backed up by both modelled and experimental data - see Myneni et al, Nature 386 (1997) for some convincing evidence from our group.

Of course, you are perfectly welcome to download a GCM such as CCM3 and go through it line by line to see whether it is "real science" or not.

If you want to discuss what is "real science" or not email me or if you're in the Boston area, come round - my work address is on my web page.

Nick

FYI, ALL current climate/NWP (that's numerical weather prediction to you) models use radiative transfer models - yup, using numerical quadrature to solve scattering and absorbtion models (Rahman et al.) You'd be mental to do ray-tracing (i.e. Monte Carlo) on an entire atmosphere. Anyway, the radiation bit is only one part of a GCM - there's atmos circulation, ocean circulation and SVAT (soil-veg-atmos transfer). If you wanna run your own GCM on your home machine (providing it's a UNIX/Linux box) check out CCM3.

Nick

Guys like Alvin Toffler will continue to publish wave after wave of best-selling bogus prophecy while NCAR continues to spend our money running the same old models on bigger and bigger machines. Consequently, I'm sure they are highly appreciative when folks insult systems that "put your credibility where your mouth is" like weather futures or Idea Futures.

And don't even get me started on The Digerati.

I have been thinking about how to do "collaborative moderation" for a few years now. To summarize my ideas in the context of Slashdot:

1. What you call "moderator", I call "participant". Any human person can be a participant. A participant must be registered (though possibly psueodonymous), and possibly authenticated to prevent cheating.
2. Participants rate each message based on its "value" (not agreement).
3. Participants accrue "weight" based on some algorithm which is a function of the value of the messages that they have posted in the past. (weight might change for other reasons, too)
4. Messages have values based on some algorithm which is a function of rating and the weight of the rater.
5. Users can sort/filter messages based on their values.

Obviously, what this creates is a "consensus value" of a message. This would be the first thing for slashdot to do. (If I understand what you are doing now, you assign a weight of 1 to the 408 moderators, and a weight of 0 to all others).

It's also very interesting to have a second message rating, namely "agreement".
1. This helps the person doing the rating to explicitly seperate message value from agreement.
2. Its allows to compute distance between two participants based on how much they agree.
3. It allows messages to have other kinds of values besides consensus values, for example the "agreement value" of a message uses the agreement distance between participants to weight the value ratings. So you can say "sort the messages based on how much people agree with me".
4. Strong clusterings of participants might emerge that indicate different "viewpoints". A viewpoint might be thought of as a synthetic participant. Users can examine values from different viewpoints (ie weighted by viewpoint agreement), and thus escape from the "tyranny of the majority" (if they wish).

These ideas are embedded in a bigger system for "wide area collaboration", focused on "a better Usenet". I wrote a paper on it at:
http://www.unidata.ucar.edu/staff/caron/collab/chi _cscw.html

I would be very interested in comments, critisism, and collaborators. I would be interested in helping slashdot implement any such system (Im a programmer: Java and C).

...given similar L2 cache sizes and reasonably modern hardware, it's not really that much better than an UltraSPARC or high end PowerPC based AS/400 system for serious number crunching.

For air pollution forecasts, have a look at this.

fwiw

The UNAVCO (University Navstar Consortium) has prepared a programme package TEQC (Translation, Editing, Quality Control) for reading and formatting to RINEX (the open standard for geodetic GPS data) from a considerable number of both geodetic and simpler GPS receivers. It can read both stored data files and input from the serial port.

This software is (currently) binary only, but free and Linux is one of the primary platforms supported.

Here.