Black Coffee (Deprecated)

Starbucks, fear McCafe

Over the Chinese New Year holiday we went to Great World City shopping center to watch The Last Samurai. Since we were early, we needed to kill some time. Since it was Chinese New Year, practically everything was closed. The only place to get coffee was McDonald's new venture, "McCafe." It sounds like a bad joke at first, but after reflection I realize that it's a particular lethal threat to other coffee places.

Why? Because no one else is better at Food Science and Standard Operational Procedures.

What's my biggest complaint about coffee places? Lousy coffee, of course...

• lukewarm latte milk
  
• weakly prepared espresso
  
• stale black coffee
  
• too much milk in the latte
  
• black coffee that's not piping hot
  
• "out of stock" black coffee
  and on and on an on

You have to get the preperation and repeatability right before you can even worry about whether your coffee bar is in the Super League of Coffee. Currently, I only recognize two Super League members, Peets, and an independent coffee bar in Alice Springs, Bar Doppio.

It's not like I'm a coffee martyr and don't goto Starbucks and Pacific Coffee and the others, it's just that my principle worry when I go in is not what their roast-of-the-day is, but whether the barrista is someone I remember as competent or the latest hired-on knucklehead.

That's why McCafe is such a threat to the Starbucks-tier of coffee bars. When is the last time you went to McDonalds worrying about who was working in the kitchen, concerned that you might get an undercooked burger, or too much ketchup, or forgotten pickles? Probably never. Why? Because McDonalds has devoted itself to consistency and standards. So if anyone knows how to take commodity employees and turn them into perfect barrista-drones, it's them. So no more will my first sip of the latte be filled with fear. I'll know it will be 165 degrees +/- 2 degrees. I'll know that the brewed black coffee is no more than 20 minutes old. I'll know that they'll rip out my drink reasonably quickly.

Starbucks isn't bad at that, it's just that McDonalds is better. They strike me as having similar ruthless efficiency as Walmart. As well, I don't think Starbucks has any unassailable "intangibles" that McCafe is somehow incapable of mimicking. Frankly, the interior of the McCafe was fine -- decent leather chairs, muted dark, tasteful interior, clean tables, decent crockery, etc. It was no worse or better than Starbucks.

Will McCafe succeed? I don't know. I could have written up this blog article from the point of view of how an MBA would look at McCafe's 'business proposition'. Enough others will do that. But from the "I just want a decent coffee on demand" angle, I'll say that if people can get over the "baggage" of buying a four-dollar coffee from a McDonald's-owned bar, they'll appreciate the consistency of their drink.

Do engage the truck

So you've probably seen footage on tv of an LAPD helicopter chasing a hapless guy through suburban backyards, illuminating the perp with a million-candlepower searchlight.

Now imagine that chopper is using super-resolved telescopic thermal imaging, the systems make the hovering platform absolutely steady, and this time the perp is carrying an RPG instead of a .38. Don't worry--that's what the 30mm exploding bullets are for!

The US Army gives you...... The Apache Gunship versus Fedayeen Saddam!

I also downloaded this movie from the eMule network. Here are a couple different links but for the same movie: first edonkey link and second edonkey link.

Heads Will Roll! oh, no they won't, this is NASA....

Decrepit bureaucracy NASA spent nearly a billion dollars sending the Spirit and Opportunity rovers to Mars. Presumably there would have been a lot of rigorous testing of the components and systems before they went?

So why didn't they discover that the flash file storage system crashes if there are too many filenames in the directories before they landed it on Mars?

So yeah, sure it's very cool that these smart NASA scientists can debug the system from Earth, can even fix it from Earth, but it is astounding the problem was ever allowed to get that far.

Geo::Coordinates::UTM

I needed to use the Geo::Coordinates::UTM module to convert between latitude/longitude coordinates and 'Cassini Solder Rectangular Spherical Projection' coordinates for a side-project of my Travel Blogging Software. I downloaded version 0.02 from CPAN.

Unfortunately it didn't work:

c:\Perl\gps>perl utm.pl
perl utm.pl
Unrecognized escape \Q passed through at c:/Perl/site/lib/Geo/Coordinates/UTM.pm line 167.
Found = in conditional, should be == at c:/Perl/site/lib/Geo/Coordinates/UTM.pm
line 167 (#1)
(W syntax) You said

if ($foo = 123)

when you meant

if ($foo == 123)

(or something like that).

Unrecognized escape \Q passed through at c:/Perl/site/lib/Geo/Coordinates/UTM.pm line 178.

The failure was occuring in the 'use Geo::Coordinates::UTM' statement, I never even get a chance to execute the code. Oh no!

I sent off an email to the module's author, Graham Crookham
asking about the error (I couldn't find anything on Google or USENET that helped).

The great news was that Graham responded before I even got home from work, sending me an updated version (0.03). Apparently, like a DNS update, it takes a while for the change to percolate into CPAN. For those in a hurry to get around this problem, I'll post the diff here:

diff ./site/lib/Geo/Coordinates/UTM.pm ./hand/Geo-Coordinates-UTM-0.03/UTM.pm
21c21
< $VERSION = '0.01';
---
> $VERSION = '0.03';
72c72
< my $long2=(($longitude + 180) - int (($longitude + 180)/360 ) *360 - 180);
---
> my $long2=(($longitude + 180) - int (($longitude + 180)/360 ) * 360 - 180);
167c167
< if (($valid_utm=~ tr/\Q$zone_letter//) = 0) {
---
> if ($valid_utm!~ /$zone_letter/) {
178c178
< if (("NPQRSTUVWX" =~ tr/\Q$zone_letter//) > 0) {
---
> if (("NPQRSTUVWX" =~ $zone_letter) > 0) {

hmmm. it says it updated the version from 0.01 to 0.03. When I used Perl Package Manager (ppm) it did specify that I was downloading 0.02. Not sure who is wrong. Regardless, once I applied the patch, everything works happily.

2.66 Defends Me

A week or so ago I installed the MovableType version 2.66. We mocked its new features to block comment spam as poorly thought-out and easily thwarted.

I checked out my activity log today and saw that MovableType did manage to thwart two attacks.

2004.01.20 13:52:53 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:00 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:05 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:08 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:17 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:20 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:28 62.213.67.122 Throttled comment attempt from 62.213.67.122
2004.01.20 13:53:30 62.213.67.122 Throttled comment attempt from 62.213.67.122

from Russia

and

2004.01.22 21:33:43 204.215.204.81 Throttled comment attempt from 204.215.204.81

from Sprint

It didn't manage to thwart my mother double-clicking the comments submit button, but that's another story.

Oh for heaven's sakes.... I just saw this update:

Update: We've released 2.661, an update to 2.66, to fix a problem with invalid XHTML in the comment redirect script, along with using comment IDs instead of URLs to fix an open redirect problem.

This is tedious. I am not a big fan of updates every four days, especially when their upgrade process is so primitive.

Cautionary Tales of Secondhand Rolexes

Here's a true story I was told some time ago:

One of my friends is a cop. He'd borrowed some money from me and was having an impossible time covering all his debts. So instead of paying me back in cash, he gave me a Rolex.

Where had he got the Rolex? From the wrist of a dead person he'd come across during a crime-scene response.

What did I do with the watch? Kept it in my dresser drawer for years. One day my wife asked for a watch, so I gave her the Rolex and didn't tell her where it came from. She wore it for about three years till she discovered the watch's provenance from another friend.

She wasn't very happy and she made me pawn it.

The end.

A Pointless VaporDevice: War Driving Machine on my belt

With off-the-shelf equipment and, say, gnu/linux, could you build a handphone-sized device that could silently War Drive as you walked about? The device would have to:

• Assay the all the networks it comes across
  
• Know approximately the network's physical location
  
• Record all that data for retrieval later.
  
• Be inconspicuous (no larger than a typical modern handphone)
  

Attention Attention, Chinese New Year Has Arrived!

According to my Chinese wife, for the next 23:45 hours, I can neither swear nor work. Both would be very inauspicious things to do during the first day of Chinese New Year.

Apple iGPS

-----Original Message-----
From: Matthew Eldridge [mailto:graphics@eldridge.stanford.edu]
Sent: Monday, January 19, 2004 1:59 PM
To: Nils Blutig
Subject: RE: FW: Dates encoded in Garmin Waypoints?

Nils Blutig writes:
> you know who really needs to manufacture a handheld gps?
> Apple. They'd do it right.

Mmmmm. It would only speak AppleGPS protocol, but that wouldn't
matter, because it would only talk to a FireWire800 interface in a G5
MacIntosh.

But at least it would cost $800.

And be hermetically sealed so you could just throw it away once the
battery dies.

And it could use that ugly-as-shit-1984-nostalgia font they use on the
iPod.

Rally Car Videos

Another WRC crash video. Highlights include the silent footage a (probably Italian) driver berating his copilot after a crash, two racers oblivious to a major fire occuring in their backseat, and in-car footage of a rollover, after which the racers continue driving. Lowpoints include a heavy Queen soundtrack and no car audio.

Here's a very sharp highlights reel from last year's WRC series. Soundtrack is decent, cinematography is excellent, and there are a lot of great launching scenes. These guys take some nasty rollovers. I don't understand why their rollcage doesn't prevent the roof from caving in more -- every accident I see seems to crush the roof too close for comfort.

Came home from work tonight to watch a forty-minute video of the 2002 WRC at Calalunya Spain. Highlights include narraation in English, a lot of onboard camera action, and the top tier driving. (And also a copilot berating his driver after a crash, "How could you be so stupid!? I told you you twice")

Paris Dakar 2004 Footage Update: Good News/Bad News

Bad news first, during the last week the CNN and BBC coverage of the Dakar Rally has remained frustratingly slight. The good news is that I have found several other interesting sources.

eMule Network
I just started using the eMule P2P filesharing client this week. It's a really slick program, and deserves a good writeup, but for this article's purposes, I'll just say you can find (a few) Dakar videos on the network. I'm still downloading some. Unfortunately several that I've downloaded have turned out to be awful French comedies, either cartoons, or those stupid foam puppets that look like European politicians (zzzzzzz). However, I just downloaded the best Dakar video I've seen yet. Here's the eMule link for it. This link is meaningless unless you've installed an eDonkey-compatible client, like eMule.

Charlie Rauseo
Googling one night, I stumbled onto San Francisco DJ Rick Stuart's blog where he had posted a few notes on a buddy who was racing a bike in the Dakar Rally. He's been nice enough to continue updating me on Charlie Rauseo's adventure, including forwarding me an email from Charlie on how he eventually was forced out of the Dakar Competition. It's a great story just because it describes what this race is like for someone who is not a factory supported monster.

Thanks for following my progress in the Rally. I was certainly thinking of all of you watching on the web and on TV when I was out there.Here's the short version of what happened: (I'll write more in the next few weeks)

Everything went as planned in the European sections. I went slowly and did not crash.

In Morocco, I started to wick it up, but not really much more than a comfortable "trail riding with buddies" pace. Got a 41st when nothing went wrong on the first day. Then 2 finishes in the low 50s after getting spun around a bit in villages. Nothing major and I was feeling great. No physical troubles and RRUK (arranged by RallyConnex) was handling the minor bike problems very well.

Then, on the long stage, I felt really good. The piste was very fast, so I decided to turn up the heat. I passed bunches of riders with fancy corporate paintjobs and was feeling really good. Headed for my best stage finish yet. Then I broke a chain (the "unbreakable titanium one") Bummer. I fixed it, but it ended up a bit too short and too tight. So, I slowed down to save my countershaft and rear wheel bearings. I still finished 85th. Still in the hunt, and early enough to get good sleep and food, and repairs, before the next day. Up to this point, I had no big "offs," no damage to the bike, no particular drama at all.

The next day, Stage 8, was rough. I crashed into a big hole in a sand wash early and hurt my ribs. At first I thought they were broken, but it was probably just a bad bruise. It still hurts pretty bad right now.

Then, I got a bit lost, like most everyone, and lost about an hour. The fast cars and trucks passed me. OK, I thought, today I just survive. I took 1000mg of Ibuprofin and just took it slowly. Then, about 150k from the end and a few hours before dark, I ran out of fuel. I was topped up in the morning, but some must have leaked out after my crash when the bike was upside down.

I flagged down every vehicle, but the bikes had none to spare, and all the cars and trucks are deisels. Hours later, the KTM Red Bull truck stops and gives me 20 liters. Plenty to make it to the next check, where there is gas. I'm going to be late and get a penalty, but I'm going to make it. Maybe I'd try for a good stage finish sometime later in the rally, but not that day.

In the dark and in the dunes, I catch a few slower riders. At one point, I'm leading Ennio, a French competitor #88, and Yuki, a Japanese woman #130, both on 660 KTMs like mine. An Organisation doctor in a Land Cruiser stops us and tells us to stick together for safety. So, I lead them through the darkness with my super-bright Baja Designs HID light. Ennio likes to fall and rest, and Yuki has trouble picking her bike up, so the going is slow and I help both them out a lot. My battery gets tired from starting and restarting (and maybe some other problems, I don't know), and they have to jump start me several times after our stops, but we continue. At one point, we sleep for an hour and a half because we, (mostly Yuki) are exhausted. At about 4 or 5 AM, we come to a difficult section where the trail is confusing. We try several options but can't find the right way.

Ennio's navigation equipment is smashed and does not work. My GPS has somehow lost its power from my battery, and Yuki's GPS plug is damaged. We fix Yuki's GPS with some of my tape and she rides ahead to find the trail. Then, I stall. She keeps going and Ennio stops to jump-start me. She never returns and Ennio and I are in trouble.

Half an hour later, a TV crew shows up and offers to show us the way. Great! We follow along, but I stall again in the deep sand and rocks and watch my chances fade as they drive off into the distance. I had never considered the possibility of not finishing that stage until then! That made no sense! I was strong and feeling good. My bike was, except for the battery, running great. But, there was no way of starting the bike and I was not going to make my start time. Several hours later, Patsy Quick and Clive Town came by and started me. We knew it was too late, but we continued until I fell again in the rocks, hurting my ribs so much that I could not lift the bike. I told them to continue without me.

Nearly a day later, the Sweeper Truck came to pick me up and take me into Tidjika. There were 13 bikes and 11 people on the truck. Some got helicopter rides. I wasn't done yet, and neither was Christo Aspeling from South Africa, so we repaired our bikes in Tidjika (I stole a battery from a crashed KTM) and we rode to Dakar. Took us 2 full days. 5 flights and 2 more days, and now I'm home in SF.

I feel like I have unfinished business in Africa.

Thanks for your support and well wishes! I'll write more, and post pictures, later.

Charlie Rauseo

Charlie sounds like a pretty decent guy, and Yuki [Tanaka?] sounds quite yucky. (Ditching your fellows in the middle of the desert? that's as bad as doing it in the mountains. Shameful.)

Google Answers

Someone responded to a Google Answers I posted last year. They pointed out that there are a lot more videos on the French side of the official website.

Mitsubishi Motors
Mitsubishi (who support the dominant drivers) have a nice site.

GPS-It v0.1

Added TDateTime parsing functionality to GPS-It.

Dragon Bento

Handphone picture of not-very-amused-at-all-Dragon-Lady after I brought up the hilarious "tuber-colon-sus" for the third time during our traditional Saturday afternoon bento box lunch at Kuriya.

TDateTime to Perl Date::Manip Date

OziExplorer outputs waypoints and tracks with dates and time stored as TDateTime objects. TDateTime objects are used by the semi-obscure Delphi language to encode dates. It's a real number which reflects number of days since the epoch. For whatever reason, their epoch is December 30, 1899.

GPS-It needs to understand the dates in these waypoint logs. So I settled on using the perl Date::Manip package to handle all my date arithmetic, and wrote a short subroutine to convert TDateTime into Date::Manip Date objects.

The function is written quite inefficiently for two reasons:
1) My perl is crap
2) I wanted to expose my conversion "logic" in case I am doing stupid things, like shifting by a day, etc.

Update... Perlmonger 31die refactored my stinky little routine into something half as long and twice as clever.

sub td2dmd {

# td2dmd -- TDateTime to Date::Manip Date converter
# takes a TDateTime and returns a Date::Manip date

# originally by Nils Blutig jan 17 2004
# thuggishly refactored by 31die jan 17 2004
# http://karavshin.org/blogs/black-coffee/archive/000807.html

my $raw_tdate = shift;

# whole number of days
my $tdate_days = int( $raw_tdate );

# number of seconds into the last day
# truncate to integer -- DateCalc doesn't take fractional seconds
my $tdate_secs = int( ( $raw_tdate - $tdate_days ) * 60 * 60 * 24 );

my $tdate_epoch = ParseDate( "12/30/1899" );

my $error;
my $date = DateCalc( $tdate_epoch,
"+ $tdate_days days $tdate_secs seconds", \$error );

die "td2dmd( $raw_tdate ) : $error, stopped" if ( $error );

return $date;
}

Ah-Ling

So Ah-Ling has been suffering this dry cold for ten days now. Apparently she was wracked with cough all night long, but since I went to bed after three and slumbered like a corpse, I didn't hear a thing. So just now she was hanging up laundry, coughing, and I said, "so what are we going to do about your health?"

"What do you mean about my health?"

"Well, you're coughing like you have tuberculosis."

"Choi!" she yelled [sort of a Chinese 'Knock-on-wood'], "saying I have 'toober-colon-us'" *phonetically spelled as she said it

"No," I explained, "I didn't accuse you of having a potato in your ass, I said you had a persistent dry cough."

I got my head smacked for that remark.

The Goatse Man is Dead; Long Live the Goatse Man

RogerW passes on the bad news that the Goatse Man is dead:

<@b> The domain goatse.cx has been found in violation of .cx AUP policies, http://www.nic.cx/policies/pdf/cx.AUP.pdf #5, page 7, and is therefore suspended.
<@r> shit, that sucks
*** joey (joey@brodels.gngsta.com) has joined nologin
<@s> yea i read, page 7 only talks about payment issues though
<@s> nothing about content
<@r> http://www.ircbible.destrukto-theater.nl/
<@b> ya
<@b> im confused too
<@s> i dunno what the #5 means
<@s> oh i see
<@s> Communication publication or distribution of adult or obscene content
<@s> or images by way of embedded links in unsolicited email, postings to
<@s> news groups, internet forums, notices to instant messaging programs,
<@s> where the internet user is not explicitly made aware that by clicking on
<@s> the link they would be directly exposed to adult or obscene content.
<@b> hah
<@b> he'll have to make a splash page
<@s> i already put the lawyer warning on there
<@p> hah
<@b> that amendment to thier AUP
<@b> is like 100% goatse
<@s> > Over the years we have received numerous complaints of this domain's
<@s> > content, but no person filee an AUP violation form against the
<@s> > domain. Recently the .cx board met and revised all .cx policies (December
<@s> > 2003). One of the .cx policies that has not changed is that each domain
<@s> > holder is required to review the policies every thirty days and make sure
<@s> > their domain is in compliance (Please read part 1, page 2 of
<@s> > http://www.nic.cx/policies/pdf/cx.registration.agreement.pdf).
<@s> >
<@s> > We do not review web sites and cannot ensure every domain holder is in
<@s> > compliance. But, if a domain is brought to our attention that fails to
<@s> > comply with our policies, we reserve the right to suspend the domain.
<@s> >
<@s> > I am unclear if you change the content, the suspension might be
<@s> > revoked. If you are considering this option, please send a note of inquiry
<@s> > to info@nic.cx.
<@s> >
<@s> > Best Wishes,
<@s> >
<@s> > Elaine Pruis

WRC Crash! Crash! Crash!

Just checked my eMule queue and saw that it completed downloading a thirty minute video of WRC crash footage. Turns out it's a Japanese production, so the announcer isn't much help. It starts out quite cheesy (1980's graphics titling, Hair Metal background music, and pretty lousy film quality) but even then it's entertaining. When I hear "WRC" I'm expecting $250,000 Peugot 206's and Lancer Evo's. Instead the lineup is mostly all the stinky cars you have to start out in when you begin Grand Turismo -- Toyota Truenos(*), obscure Opels, banal Escorts, drug-dealer Sciroccos, and tasteless 3-series.

Some of the stuff is quite ridiculous. Lots of rallys on snow roads, car flies off, crashes, crowd pushes it back on to the course, and it tears off again, sans windshield, or bumper, or whatever. I've only watched a fraction of the video and already I've seen two cars continuing driving after totally shredding the tire from the wheel.

Some of the race planners must be totally demented. The video showed three different massive roll-over, ejection accidents along the same stretch of the same race. Why? Well, you are flying along the road, start climbing a gentle grade while simultaneously making a gentle arcing turn. As you peak the hill, you launch off the ground, continue flying straight, which lands your left two wheels into a deep ditch running alongside a twelve foot high embankment. This nicely snags your fast-moving car and sets it tumbling end-over-end along two different axes.

But it's ok, because the drivers are demented too. One strange looking car plowed straight into an embankment, went almost straight onto its nose before landing again, crushing the roof and popping all the windows, but landing on four wheels. The crowd surged towards the drivers to help them escape before they were burnt up. You could even see the driver fussing about in the crushed cockpit. The thing was, he wasn't trying to escape the car, he was trying to sort himself out to get driving again. It was like something from the Terminator -- crushed car rights itself and continues driving!

And apparently the demented drivers and demented race directors directly leads to demented fans, as witnessed by the car with its front right wheel completely crushed into the engine compartment. A dozen fans jumped onto the back left of the car, to lift the dead corner off the road, and allow the destroyed rear-wheel drive car to somehow continue driving.

Finally to get a real taste of the dementia, they run a section of onboard cameras. You've got the co-pilot streaming driving dimensions like an auctioneer, the driver furiously double-clutch shifting and steering, but the thing is, it's at night... in rain... and you basically can't see anything. Then there's a sudden, huge bumping and shaking and you just hear someone moaning and the other one asking if he's alright as he furiously burns the car back onto the track and takes off again.

Unfortunately the movie truncates right as it's getting to the offroad desert races.

The video is called "Wrc.Rally.Crash.[osiolek.pl].MPG" (**) and is retrievable off the eMule network.

(*) I finally saw a Trueno on the road a few months ago! They DO exist, not just in video games. (It looked like a shitty old Datsun, basically)

(**) Here's another one which I haven't seen yet: WRC_RALLY_CRASHES.MPG

2.66....

Barely two days after upgrading to 2.65, Movable Type issues 2.66 in an attempt to deal with another wave of increasingly vicious spam-commenting/flooding.

One new feature:

Also in 2.66, we've changed the behavior of <$MTCommentAuthorLink$> to use redirects when linking to URLs given in comments. The goal of this is to defeat the PageRank boost given to spammers by posting in the comments on a weblog.

It doesn't appear to have worked retroactively, at least

GPS It v0.0

This weekend I installed Net::Blogger to help build some Travel Blogging software. I used some old, ugly crap I wrote last month that interfaced directly to MT on the karavshin server and combined it with some Net::Blogger example code from Jeremy Zawodny. That leaves me with a crude client-side program that takes a waypoints.csv file and converts it into individual blog items with embedded ICBM metadata. Then geourl.org adds the blog items to its searchable geographic database.

Incremental Improvements
I'm barely literate in perl and it shows. There are lots of code improvements necessary for gps-it, but the skeleton is there, and works.

Off the top of my head...

• Use a configuration file for all these constants
  
• Robust library for parsing and converting latitudes and longitudes
  

Shoot-and-GPS-It
The next interesting thing is code that takes a directory of digital photographs, strips out their EXIF time metadata, and then compares that to downloaded GPS track and waypoint data recorded while I was photographing. By corresponding time-stamps, Shoot-and-GPS-It should be able to make a blog item for each photo and annotate it with the nearest waypoint or nearest track location when the photograph was taken.

Simultaneous, Independent Discovery
It seems that the Research Guys at Microsoft have some similar ideas. I haven't tried out their application yet, I just discovered their site this evening and it appears to have been slashholed.

2.65 Upgrade

Upgraded to version 2.65 of MovableType. Vestiges of past hacking for GPS-It! subtly fucked something up in the XMLRPC interface, so I was getting intermittent server errors. This should fix that, as well as enjoy the upgrade's benefits, including incidentally, an XMLRPC security upgrade.

Only problem is that there are some hacks that I want to remain (mt-comments.cgi name, sizing on comments and some admin pane windows, etc) I'll have to go back by hand and re-do those changes. I made a backup copy of the 2.64 installation. Thank god for diff.

More features on Google , but some still missing

A new tranche of features available on Google... VIN numbers, area codes, flight numbers, and even UPC codes.

It's missing at least two things, though:

Currency Calculator
Like I can say "8 feet in meters" or "100 horsepower in kilowatts", I would like to be able to say, "25SGD in USD" or "1000JPY in RMB".

GPS Coordinates
Suprisingly, searching for specific gps locations doesn't give you anything special, like auto-generated maps or place information.

Twist
Then a variation on the above two... 25N 34' 244" in dd.mmmmm, for example -- there are several variations to GPS waypoint formats.

iBushed!

On average, it's a losing bet to ignore RogerW's product advice -- he knows more and researches better. Anyway, tempting fate, today I blew him off, spent $70SGD on an iTrip and was pleasantly suprised when it turned out to work great.

Overview
The iTrip is just a small module that snaps onto the top of the iPod, replacing the headphone jack. The module contains a small processor and a transmitting antenna. You load the iPod with a special playlist of songs labelled "107-1, 95-3, 102-5, etc" Each of those "songs" is just a series of tones that the iTrip's processor listens for as instructions to broadcast on a particular FM frequency. Any normal music is just broadcast through on the designated station. The transmission requires no extra power, it just bleeds it from the iPod.

Setup: First Attempt -- Car Sickness
I never have bothered to read the manual for the iPod, so I am sort of clumsy with it. I barely bothered to read the iTrip manual. I don't have a manual for our car radio, which suffers a terrible UI. So within fifteen minutes of driving to dinner, I was sick-to-my-stomach from trying to figure out how to get iTrip and the car radio to play and listen on the same frequency. It felt like I was trying to read in an astronaut training simulation. Nothing worked.

Setup: Second Attempt -- Total Victory
Came home and decided to work with the home stereo first. I read the five sentence, one-quarter-page manual (that's all it takes) and realized "Oh, so I hit the play button again after the iThink blinks rapidly." So that's what I did, set everything to 87.9, and it worked. No drama at all. The range is huge. My iPod was in my study and the stereo in the living room was still receiving it perfectly.

Next step was test it in the car. It worked fine there, too. The iPod is so diminuitive that I can stick it securely up near the sun visor. It seems happy to sit anywhere in the car and still works great. I turned the motor on to confirm that there were no interference problems.

I don't think the signal is powerful enough for the radio frequency scanner to pick up. In my case it seemed that I had to dial in the frequency with my radio first.

Review
I'm not an audiophile. If I was, I'd probably complain about diminished audio quality. But I'm not, so I am just ecstatic about the new possibilities this brings me -- lots of new music in the car, no more fussing with stupid CDs, decent music in the living room during dinner, and portable music jukebox when we goto others' parties. For me it's a great product.

In all fairness, as a manufacturer, when you are told, "you are going to make a product that bolts onto an Apple product," you should be afraid. When it's a product as aesthetically amazing as the Apple iPod, you should be very, very afraid. It's hard to appreciate what I mean until you've held an iPod in your hand -- everything absolutely flush, smooth, perfectly weighted, and blemish free. Against that, you can't help but note, "yeah, the iTrip is white like the iPod, and has a logo like a Mac font, but ewwww, look how the plastic pieces aren't perfectly flush, and that little bit there is a bit wavy, and how the rubber end piece is a bit loose." It really doesn't matter at all, the product works fine, but it unavoidably stands out.

I didn't go with the other solutions suggested in the earlier comments because I don't have a cassette deck or any sort of input jack. I had considered junking the existing stereo and replacing it with one that took DVDs or CDs of MP3's, but now that is totally unecessary. What a pity I didn't have this thing in Australia during our trip through the outback.

Net::Blogger on the FieldBox

Flailed till 4AM last night to get the Net::Blogger module installed on my fieldbox. What was involved?

ActivePerl
I ended up using Active State's ActivePerl 5.8.2 build 808 Windows AS package. I originally installed version 5.6.1 build 635 but in working out some build problems for Net::Blogger, I upgraded to 5.8.2. It probably wasn't necessary, but there was no reason to roll back once I did get everything working.

The easiest way to go would have been to use ActiveState's ppm (Perl Package Manager) to install Net::Blogger. Unfortuantely, Net::Blogger has not successfully built in their nightly process, so you have to use the CPAN module instead.

CygWin and other command line utilities
I installed Cygwin so that I could have access to a number of gnu/linux utilities used by perl to automatically retrieve modules from the CPAN repositories. (Things like wget, lynx, ncftpget, etc) Several of them are not included in the default installation of cygwin and you'll have to download them specifically.

I used the existing C:\windows\windows32\ftp.exe utility.

As well, I used Microsoft's Nmake.exe instead of GNU make, which does not work with the ActivePerl/CPAN combo. Apparently you will have to dig around for a copy of Nmake -- it's not default with windows anymore, and it wasn't easily found in Google. Best thing is to ask someone to send you a copy of it.

For whatever reason, I couldn't connect to ftp.perl.org last night. I had to give the CPAN module another mirror. The html mirror was a lot faster than the ftp mirror. Not sure if this applies to all mirrors or just the one I used.

CPAN
So when I

C:\> perl -MCPAN -e shell
cpan> install Net::Blogger

Perl::CPAN is supposed to download the Net::Blogger module, plus any other modules which it is dependent on. There turn out to be two, TermReadKey, and Error. My Make process couldn't build TermReadKey because the makefile wanted to compile part of it with 'cl', the Visual C++ compiler, which I don't have.

The solution as to use the ActiveState ppm (Perl Package Manager) to 'install TermReadKey'. This skips the step of having to locally compile TermReadKey.

Then when you run installl Net::Blogger, it will successfully make all the modules and dependencies.

Make test
There are some tests included in Net::Blogger's make file, but they're a bit dubious.

The Make process asks for some specifications about a 'valid Blogger API server' so that it can test the posting process. The test is very brittle. On my server, my user name is "Nils Blutig" -- the spaces horribly confused the test. So I made a one word account and tried again. It tried to post, but then claimed to have failed, although the error code it claimed karavshin returned was empty. When I checked the blog, I saw that in fact, it had successfully published. It wasn't obvious what to do besides forcing the make to install. Then I realized that during the interactive test period, one of the options [y/n] was "post to the server?" Select 'no', then it goes away happy and installs the module. (*)

(*) well almost happy -- during the interactive mode I got several complaints (shown in bold), but they seemed to be toothless.

t\00-basic....ok 3/6# Enable debugging output? [y/n] y # debugging is enabled # t\00-basic....ok 4/6# URI of a working Blogger API server ? http://karavshin.org/scgi-bin/mt/mt-xmlrpc.cgi # # Username ? maketestuser # # Please enter password: # # # App key (optional) ?

#
# Blog name ?
Beta Testing Blog
Use of uninitialized value in substitution (s///) at C:/Perl/site/lib/SOAP/Lite.
pm line 367, line 6.
Use of uninitialized value in substitution (s///) at C:/Perl/site/lib/SOAP/Lite.
pm line 367, line 6.
#
t\00-basic....ok 5/6# Please enter some text ?
another make test
#
# Publish this text? [y/n]
n
Use of uninitialized value in substitution (s///) at C:/Perl/site/lib/SOAP/Lite.
pm line 367, line 8.
Use of uninitialized value in substitution (s///) at C:/Perl/site/lib/SOAP/Lite.
pm line 367, line 8.
t\00-basic....ok
All tests successful.
Files=1, Tests=6, 36 wallclock secs ( 0.00 cusr + 0.00 csys = 0.00 CPU)
C:\WINDOWS\system32\nmake.EXE test -- OK
Running make install

iPod FM Transmitters

So on Roger's advice I installed iTunes to replace the uninpsiring MusicMatch software that came with my iPod. It was a lot better, and now I have all my music dumped to my iPod. Now the next question is, "what's the best way to play my iPod in the car?" There seem to be a lot of different iPod accessories that broadcast through FM channels in your car. Does anyone have any recommendations?

Someone recommended the idea for this a few days ago, and I think he was referring to the iTrip, which looks good. I neither have nor want a cassette player in a my car, and the transmitter pod looks smaller than any of the other devices I saw.

Cute....

I re-installed my Nikon Super Coolscan 4000 ED today. I found there was a firmware upgrade from 1.06 to 1.10. While preparing to install it, I found this caveat in the notes:

After starting the updater, do not perform any of the operations below until a message appears stating that the update is complete. Depending on their timing, any of these actions could damage the firmware, making it impossible to download it to the scanner: . Turning the scanner or host computer off . Disconnecting the interface cable . Performing any operation on the host computer, including moving the mouse . Connecting or disconnecting other IEEE 1394 devices or turning them on or off (note that some IEEE 1394 devices turn off automatically if no operations are performed for a specified period; such devices should be disconnected from the computer before the update begins)

Ridiculous. Of course the update worked fine. But a remark like that is such a cover-ass. So if it fails, I have to call them up, admit that I may have bumped the mouse, and then they declare themselves admonished of responsibility? If they write such brittle software, they should be criminally liable.

Dakar 2004 Race Photos

originally posted January 6, 2004

Finally! I found some semi-decent Dakar Rally race coverage. At least they have some decent photos. Still lacking any meaningful commentary. This is the sort of event that just screams for a long eigh-page magazine article of the style of Smithsonian, Atlantic, or Sports Illustrated. I just have never seen one so far.

Trucks!
Here's a (dutch (unfortunately)) page on monster T4 trucks from the Netherlands. Goddamn thing weighs 9000kg powered by a 12.6 diesel engine that pumps 552 kW!

Ah! there is an "english" link at the bottom of the page where you can read some engrish translation some of the pages. In particular, follow the "Dakar Diary" link for a decent summary of each day's events. (at least compared to anything else I've found so far)

Update! Jan 9-10th 2004

Eurosport is providing the most coverage I've seen yet on the Dakar. Haven't managed to see any of their videos yet -- I keep getting 'server busy' errors. Ah! Finally discovered why -- you need the dreaded RealAudio player(*) to see the movie fragments These clips are extremely small, short, and low bandwidth, but I guess they're better than the alternative--nothing. Note that many of the more-recently posted clips throw "file not found" errors, so try the earlier ones. Most of it's the same fleeting footage you can see on CNN and BBC.

(*) If you're forced to install RealAudio, the trick is to specify a "custom install" and then just say NO to every option they list -- no file associations, no toolbars, no icons, no nothing-at-all. RealAudio InstallerWizard needs to be treated like a detainee at Guantanomo-bay-- manacles, three guards, and a shotgun if it makes a sneaky move.

(**) Bugmaster, regarding RealAudio, says:

Screw that. Just get the Real Alternative codec (http://home.hccnet.nl/h.edskes/mirror.htm), and then watch real video in whatever player you like. I prefer Media Player Classic.

Volkswagen has a semi-decent site dedicated to their Touareg entry.

Unfortunately I won't be able to follow the Speed Channel's marathon coverage of the Dakar.

Burger King customers told: 'You are too fat to have a Whopper'

Police believe teenage pranksters are hacking into the wireless frequency of a US Burger King drive-through speaker to tell potential customers they are too fat for fast food.

Policeman Gerry Scherlink said the pranksters told one customer who had just placed an order: "You don't need a couple of Whoppers. You are too fat. Pull ahead."

Typical Worker Safety in Singapore

Maintenance worker perches his flimsy 5' high aluminum ladder on top of two bar-height cafe tables to reach a light bulb.

MoveableType Pro: Vaporware?

Promised to arrive summer 2003, MoveableType Pro seems forever elusive. My vaporware cerci, finely tuned from eighteen months squandered at ArsDigita, are quivering at the same frequencies.

Why? MoveableType Pro has, at least a few times now, silently slipped scheduled. Its feature set is very vaguely defined. The only news seems to be feature enhancements to TypePad and, lately, services to support ATOM RSS format.

None of these points themselves specifically sound unreasonable, but collectively they give me the same vibrations I felt with others' projects that were eventually scrapped, formally or otherwise.

It's getting irritating that nothing decisive is being announced about MoveableType Pro. There are things I dislike about 2.65, but I haven't fixed them because it always seems like, "well, MT Pro will be out in a month, so why invest the time now."

It would be nice if they could just release their best guess at a schedule and feature set. If it slips, it slips, but at least it's some concrete data we can use to make decisions today. The uncertainty is becoming a bit grueling.

Dakar Multimedia Coverage Disappoints Again

It drives me nuts that NASA can send back better pictures from Mars than the press corp covering the 2004 Dakar Rally manages to do from Africa.

Look around on Google News or CNN. All you get is a dry press-release and no photos. Such a pity.

I saw a snatch of video coverage from the Prologue held in France last week on a flooded track made worse by snowfall. It showed a massive Tata truck taking a tight turn and snorting wto angry black tubes of exhaust as it downshifted. Just brilliant stuff. I really wish I could find a better news feed. I had the same complaints last year.

This one goes out to all the HP Apologists out there...

For the second time in as many years, my HP 970 CXI color deskjet printer has died.

The first time it was some sort of electrical issue. This time the print head decided that the "home" position was roughly seven feet to the right of the printer.

Away it goes to HP Servicing once again.

Considering what a light duty cycle that printer endures, I can't say I find the HP products particularly reliable. The HP inkjet it replaced (500-something series) also died of Sudden Death. It's a shame it wasn't better -- it has a very slick duplex printing function for printing back and front of pages.

Update January 19th 2004 Got the printer back, 20$ later, and it apparently works fine now. The service receipt was sort of garbled: "REPAIR SUMMARY: Reseated kicker-unit monitor & tested ok, user to aware this. Product had obsolete"

Anyway, so what, and a $20SGD repair is cheap.

My new shop and office

Mercifully, our lease expires in April. When we left San Francisco, we sold 20-30% of the stuff in our full, but comfortable, 3000 sq ft. house. When we arrived in Singapore, we didn't find a 2100 (3000 * 0.7) sq ft house, we found something more like 1500 sq feet condominium.

We kept our promise not to buy much stuff for...hmmmm... the first six months? Now we (aka I) are back to the old habit of buying lots of books, lots of random computer and electronic equipment, bicycles, and lots of other rubbish. Basically my house is something like living on the USS Springfield, but with less space, and a fierce female captain.

So as I said, soon I'll be liberated, and can find a sufficiently big house. I've been idly daydreaming about my new study/office/shop. Ideally it would be a huge, open studio. Unfortunately Singapore has the nasty habit of taking a house and subdividing it into too many rooms. This condo, for instance, has the square footage to accomodate two decent-size bedrooms. Instead, there are three, one of them being ridiculously small. So I may be unable to find a house with a luxuriously large study room. In that case, I may have to settle for two smaller rooms.

Today I tried to brainstorm the functionality my new shop needs to have. I am tendering for clever ideas.

New Shop-room feature wishlist

1. Lighting

1.1. No outside glare on the monitors

1.2. Good reading light

2.1. UPS for the things that need it

2.2. No extension-cord tangles and lots of power-points

2.3. Computers and their wiring accessible, not falling behind desks

2.4. Wall-mounted speakers

3.1. L-shaped computer desk so I can use both laptop and desktop by only rotating 90 degrees

3.2. Peripherals space

3.2.1. Scanner

3.2.2. Two printers

3.2.3. DSL Modem

3.2.4. Two monitors

3.2.5. Router

3.2.6. Many docking stations (iPod, Exilim, DSL, etc)

3.3. Writing/paperwork/file-processing desk

3.3.1. Doesn?t need to be big, just tidy

3.4. ?Shop? table for projects and eating

4.1. Paper files

4.2. Manuals and product literature

4.3. Stationery

5.1. Tools and hardware

5.2. Bike equipment

5.3. Camping equipment

5.4. Small electronic equipment (GPS, etc)

5.5. Camera Gear (needs a dry-box)

5.6. Computer rubbish (old cards, cables, components)

6.1. Binder size

6.2. Manual size

6.3. Magazine size

6.4. Fiction/softcover size

6.5. Oversized books

The 'C' in Centrino stands for 'Sucks'

My Toshiba Tecra M1 has an Intel PRO/Wireless 2100 Land 3B Mini Adaptor... a Centrino Wireless card.

The idea of Centrino is that it has advanced power management -- it only spends enough electricity to keep a connection with your wireless access point. Ok. Great.

So last week I spent time testing out my access at the various Internet hotspots around my office. I could connect to Singtel and Starhub just fine. I couldn't connect to Airwarp at all, however. My Intel PROSet utility told me I was connected, but I was never allocated an IP from their DHCP server.

We went back and forth on this for a while with no success. I did some more research myself and found that very, very frequently turning off this "Advanced Power Management" and turning power to 100% solved the problems. Yep. It sure does. The problem is totally repeatable.

Ok, so that seemed like a fix, although not a very nice one, wasting some of the Centrino's advantage.

So today I decided to install a wireless network at my house. Presently I have an Efficient Networks SpeedStream DSL modem that connects to my PC via a USB port. Then my PC shares an Internet Connection (ICS) through it's ethernet card to my laptop, connected via its ethernet card.

What I wanted was a wireless router that my PC could plug into and my laptop could wirelessly connect to.

Since most routers can only talk to the DSL modem via an ethernet cable, I had to toss out my USB DSL modem and buy one with an ethernet jack instead. I ended up buying a Aztech DSL305E(B) ADSL Ethernet Bridge Modem and a NetGear MR814 Cable/DSL Wireless Router (version 1, this is not MR814v2).

The Modem is garden variety and Singnet (my DSL provider) approved. The Netgear Router is quite popular, and quite common.

First I installed the Aztech. There isn't much installing it. Since the router talks directly to it, there is no need for any setup. All I did was plug it into power and phone. I did notice that the "DSL" light kept blinking -- it never stayed on. This seemed wrong. After some tinkering I realized that the noise filter that came included with the modem seemed to be interfering. Once I removed it, and plugged the cable modem straight into the wall, it worked fine and kept a solid green light.

Next the Router. It's simple too -- give it power, connect it to the Aztech and another ethernet connection to the PC. Configuration is done via a web interface to a specific IP address the router listens to. It's accessible by both wireless and wired connections. I started out by configuring it with my PC. The configuration is very easy. Basically you just tell it your name/password with Singnet. The rest is roughly automatic.

One thing that does annoy me, though, is that my PC doesn't seem to immediately get access to the web. It seems to take some time to get an IP and DNS servers allocated to it. I am not sure if I had used the wizard included on the CD (which I didn't use) if some sort of "Connect to Internet" icon would have been created that I could use to force an immediate connection to the internet. Any ideas?

But that turns out to be the least of my irritations. The wireless situation is fucking infuriating.

Essentially I am having the exact same problem I had with the airwarp network. I ostensibly connect to it fine, with a nice strong signal, but the router and laptop absolutely refuse to exchange an IP and DNS address to use. I have been through every fix I can think of or find.

Drivers and Firmware.

Hacker C^4I

> -----Original Message-----
> From: Matthew Eldridge [mailto:elronhubbard@Graphics.Stanford.EDU]
> Sent: Friday, January 02, 2004 6:30 AM
> To: Nils Blutig
> Subject: fools errand
>
>
> So I'm still seeing a few messages a day regarding the computer
> breakins at Stanford. Seems so utterly hopeless. How do they
> breakin? Well, the compromise a single machine, and then they have
> internal access, and then they start compromising other machines with
> a local attack (from a user logged into the machine) which is much
> easier than a remote attack.

Given that the trick is to capitalize on your initial exploit as quickly as possible, it seems like a motivated hacker needs a C^4I assistant to help you keep your battle tempo after the initial success.

So if I was a hacker, instead of worrying about finding the next obscure exploit that might compromise the three servers in existence with some particular build of four different programs, I'd be working on this exploit framework.

What would it do? It would have a library of all the known exploits and how they can be fired off. So once you score the initial intrusion, it probes the system and looks for signatures (file versions, directory structurs, permissions, etc) that it can then suggest the ten most likely-to-succeed/powerful follow-up attacks. Then the hacker decides which branch(es) to follow, tries those exploits, and re-probes the system again.

The framework gives you the power of thoroughness (can quickly examine the whole exploit corpus), speed (automate as much as possible), and maybe even intelligent filter and 'look ahead'. What do the last two mean?

Filtering would be something like, hacker says to himself, "I think the sysadmin will detect attacks on sendmail quickly, so avoid initiating any sendmail-based exploits." Lookahead would be something like an understanding, "I know that exploits 3, 7, and 14 will work, but if I fire exploit 7, it cripple other things which will preclude a number of follow-on exploits, so let's not do that.

I don't follow this sort of stuff, so probably all this already exists and it's probably far more clever anyway.

Whee, I'm getting carried away like Neil Stephenson now.

Unsecured Wireless: The Driver becomes the Hunted?

I've been enjoying my new Toshiba Tecra M1 laptop. This evening I was experimenting with one of its powerful components, its Centrino wireless card. I live in a condominium complex which is also adjacent to an industrial estate. When I went war driving in my living room I found at least a few networks, including one or two which were unsecured. From these I got an IP address and could surf the internet happily.

As I was browsing "My Network Places" I remarked on the large number of computers exposed to me. It started me wondering, "is my laptop secure in this environment? Am I falling for a honeypot by using this network?" Really, I don't know. I need to figure out "how do I secure my XP client computer from Shadow of Death when I am using a randomly-discovered wireless network" ?

Is it as simple as just turning on my firewall? Is there something I should do to hide the very existence of my client PC on that network? What else should I be concerned about?

I sort of know the answer, yes, there are lots of things to worry about. Anyone want to suggest the few fixes and precautions that stem eighty-percent of the danger?