# Last edited on 2004-01-16 23:59:36 by stolfi

Decided to start keeping a blog. Me too. Why not?...

---

I stumbled into /Nature/'s news note on Gordon Rugg's "VMS is a hoax"
paper. Unfortunately the Unicamp library does not subscribe to
/Cryptologia/. I hope that Gordon is kind enough to send me a copy.

From the summaries, it seems a nice experiment, but it hardly proves
anything. It seems that he didn't address at all the strongest
arguments that we have against the hoax theory -- the Zipfian word
distribution, locality of word usage, non-lullian word frequencies,
etc.
 
Gordon's method can generate random text that looks just like
Voynichese to our eyes. This is no big news, of course: Guy's /Monkey/
can do it too. After Shannon, everybody knows that a simple stochastic
model can generate gibberish in any language that is indistinguishable
from the real thing -- but only for readers who can't read the
language, of course. The artificial text will have the right k-tuple
frequencies, and, if k is large enough, it will even have the same
word frequencies, and therefore will pass most statistical tests.

Gordon's contribution apparently was to show that a convincing
"monkey" could be built and used by hand, based on simple
table-and-grille technology that was in use in the 1550's.

It appears that Rugg's method is some Markov-like generator that uses
word segments -- possibly prefixes, midfixes and suffixes -- as units,
instead of individual characters. This too is not surprising: a good
"monkey" for Chinese or Vietnamese would certainly use the same
approach -- i.e. treat the initial, medial, and final segments of the
syllable as single characters. 

--- 

A posting to the CIVILS list reports that DENATRAN has prohibited
dashboard-mounted TVs, DVD players, and the like -- including GPS
devices. TVs and DVS can be installed in the back seat. GPS devices
may be dasboard-mounted, but must have a cover that is closed while
the car is in motion.

No one would complain about the first two. The sender was outraged
about the restrictions to GPS usage while driving. I think that it all
depends on how much attention the device demands.

A "safe" GPS should show only a few bits of information at a time,
that can be read in a fraction of a second and monitored with
peripheral vision -- e.g, only a very rough diagram of the next
intersection, and an arrow showing how to turn. In that case the GPS
would be no more dangerous than a speedometer.

However I suspect that the "hottest" GPS devices are set up to display
a detailed map of the neighborhood, possibly with text and other
complex information. (Otherwise, what would be the fun of having such
a gadget?) In that case, the device may require the driver's full
attention for seeral seconds at a time. 

That is of course a risk too high to pay for the convenience. What is
more important: the (very real) risk of loss of life, or the (largely
imaginary) benefits of using the a fancy GPS map in real time?

Until GPS displays are properly standardized and regulated so as to
remove this risk, it is best to err on the safe side. So, kudos to
DENATRAN for doing the right thing, however impopular that may be.
Too bad that their mandate does not cover voting machinery... 

PS. A follow-up message concurred that GPS restrictions were a sign of
backwardness and technophobia (hm, where did I hear that before?). The
author noted that there are no similar restrictions to GPS use in
developed countries.

Now, the last point is not relevant. In the US, at least, traffic is
infinitely more orderly and safe than in Brazil; drivers, on the
average, are much more responsible; and the punishment for any
mistakes is stiff and certain. I don't know which is the cause and
which is the consequence, but laws tend to be (have to be?) much more
"paternalistic" and explicit here than there -- it is just a cultural
difference.

On the other hand, you just can't buy a pint of ethanol in the US.
Os a test tube in Texas...

PPS. Over the past 30 years, car accidents claimed the lives of three
department collegues (one was my office mate) and several of our
students. For comparison, while many of my colleagues have been robbed
at gunpoint, no one died from that.  And I haven't heard of a single
case of someone dying because he missed the right highway exit...

---