Dinomite.net

A blog post by one of EQO’s founders, Jeff LaPorte, claims to have insight into the market share of the instant messaging services throughout the world. The graphic (PDF) shows the strength of competing IM services in different markets throughout the world. I immediately questioned the data because it showed AOL Instant Messenger with a narrow lead over MSN and Yahoo!, who share a close second. Personally, I’ve never used either of those services, nor do I know anyone who regularly does; is that because I’m a curmudgeon or the EQO data is inaccurate? I must side with the latter.

The crux of the issue is that EQO is an instant messaging application for mobile phones running Windows CE, so the data they have is for that subset of the IM user market, not representative of the market as a whole. So what can we glean from this data? People who have Windows-based smartphones prefer MSN worldwide, save for AIM’s slight edge in the United States. What can’t we learn from the data? The state of the instant message market as a whole. EQO’s usage data ignores other smartphone users, most notably the iPhone and users who don’t have smartphones or aren’t interested in IMing from their phones.

Much of the Electronic Frontier Foundation’s talk at DEFCON 16 concerned the temporary injunction that prevented MIT students from giving a talk on vulnerabilities of the MBTA farecard system. The injunction seems to say taht the only people prevented from disseminating the information about the system are the MIT students themselves and any cohorts in their research. One question that I head was which other systems may be vulnerable to the same RFID card attack and here’s the list from their slides:

• Boston
• London
• Netherlands
• Minneapolis
• South Korea
• Hong Kong
• Beijing
• Madrid
• Rio de Janeiro
• New Dehli
• Bangkok

More information is provided in their slides.

Put the Mechwarrior 2 CD into your Mac’s CD drive; it will show up as two mounted discs, one for the audio and one for the game data. The disc I have is for Mechwarrior 2 v1.1, so the game data is mounted as /Volumes/MECH2_V1_1; if you have a different version, the mount point will be slightly different and subsequent commands should be adjusted to match. To access the disc in DOSBox, you must mount it. Start DOSBox and type:

mount d /Volumes/MECH2_V1_1 -t cdrom -usecd 0

mount c ~/dosprog

d:
INSTALL.EXE

c:
cd c:\MECH2
MECH2.EXE

I doing some development involving Apache handlers which involves lots of restarting it and occasionally Apache shutting down uncleanly. After numerous restarts, the webserver refused to start printing the following in the error log:

[Wed Jun 25 18:57:11 2008] [emerg] (28)No space left on device: Couldn't create accept lock

After being confused because there certainly is enough disk space, I Googled for other possibilities and it turns out I had hit another limit, that of active semaphores. When some of my processes were killed, they died uncleanly and didn’t clean up their open semaphores, filling the maximum number available to my user. It’s easy to tell that this is the root of the issue; running ipcs -s showed a list of 95 open semaphores. Clearing them up is just as easy:

~$ for id in ipcs -s |awk '/USERNAME/ {print $2}'; do ipcrm -s $id; done

That prints the list of semaphores, grabs the ID ({print $2}) for lines containing USERNAME (your own or that which the webserver runs under) and then ipcrm’s those semaphores.

Side by side diffs are much more legible and useful than those in unified format or any other linear diff. By default, the svn diff command presents output in the unified format, though it has an option, --diff-cmd, which allows you to specify the program that will perform the diff. Passing vimdiff as the diff command doesn’t work as the options passed by svn diff are a bit complicated:

~/code/perllib/MG$ cat foo.sh
#!/bin/bash
echo $*
~/code/perllib/MG$ svn diff --diff-cmd foo.sh Proxy.pm
Index: Proxy.pm
-u -L Proxy.pm (revision 21095) -L Proxy.pm (working copy) .svn/text-base/Proxy.pm.svn-base /tmp/svndiff.tmp

Subversion is telling diff to output context (-u), show pretty names for the files (-L) and then the two files to diff. To make this work with vim, we simply need to cut out the extra options. To do so, I wrote a simple script:

\#!/bin/bash
shift 5
vimdiff "$@"

I put that in ~/bin/svnvimdiff and now I can do svn diff --diff-cmd ~/bin/svnvimdiff Proxy.pm and view the diff in vimdiff. Since this is a command I use often, I aliased it in my .bashrc:

alias svndiffvim='svn diff --diff-cmd ~/bin/svnvimdiff'

I’ve heard of doing this for a long time, but always figured it’d be a huge jump to put my home directory into a revision control system. For years I’ve scp’d files to a new machine when I moved in and when I made changes to my vimrc or ssh config, it was hell to propagate the changes to my other machines. No more is this an issue, since I’ve commited my home directory. The process is really quite trivial.

I first created a subversion repository on my Linux server. A special machine isn’t required, nor is root access; you just need somewhere that can be accessed with SSH from the intertubes.

~$ svnadmin create subversion

It’s as easy as that. To begin with, let’s create the standard subversion directory structure.

~$ svn co file://$home/subversion/ foo
~/foo$ cd foo
~/foo$ mkdir homedir
~/foo$ svn add homedir
~/foo$ svn commit
~/foo$ cd ..
~$ rm -rf foo

So now we have a repository with a homedir directory in the root; this is where all of the files from your home directory will be checked into the repository. To begin with, check out this new project into the root of your home directory.

~$ svn co file://$home/subversion/homedir .

Since there isn’t anything in there, nothing will actually be checked out, but subversion will setup source control on ., your home directory. So let’s add something to it. I have a .vimrc that ought to be identical across environments, so I’ll check that and my .vim support directory into the repository

~$ svn add .vim*
A         .vimrc
A         .vim/colors/drew.vim
A         .vim/plugin/feraltogglecommentify.vim
~$ svn commit -m "adding stuff for the first time!"

Easy as that, all of my vim configuration files are in subversion. Now for the awesome part. Like I mentioned, I have a number of machines and, things like vim configuration files should be the same on all systems. So now, I hop onto another machine and checkout the homedir project. First I remove the existing .vim* files since they will be replaced with the repository versions.

Caligula:~$ rm -rf .vim*
Caligula:~$ svn co svn+ssh://dinomite.net/home/dinomite/subversion/homedir .
A    .vim
A    .vim/colors
A    .vim/colors/drew.vim
A    .vim/plugin
A    .vim/plugin/feraltogglecommentify.vim
Updated to revision 1.

Bam. I’ve now got the exact same files on both machines. If I’m at the office and find some flash vim option that I want to use, I simply add it to my .vimrc, check it in with svn ci .vimrc and when I svn up from the other machines, they’ll get the new changes. In reality, I keep a bunch of stuff in my homedir subversion repository: .bashrc, .screenrc, bin directory and many others. Subversion helps me keep all of my config files in sync across a number of different machines that I use on a daily basis; if you use more than one system with any frequency, I highly suggest checking in your home directory.

Here-documents are something that, once you know about them, you wonder how you ever got along without. I first learned about here-docs in Python many years ago, but didn’t know their proper name. In Python, you define a here document with three single or double quotes:

foo = """
Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et
dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip 
ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu 
fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt 
mollit anim id est laborum.
"""

One day, when I wanted to put some large blocks of text amongst code, I searched for a way to do this in Perl and found the syntax:

my $foo =<< "EOF";
Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et 
dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip 
ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu 
fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt 
mollit anim id est laborum.
EOF

Note the use of quotes for the string-terminating sentinel (EOF, in this case, though it can be anything), which affects how the sting is interpreted. Just like with other strings in Perl, double-quotes mean that variables and escapes will be interpreted whereas they are ignored in single-quoted strings.

Anyone who pays attention to the car world from an efficiency perspective has no doubt heard of the popularity that diesel engines have in Europe, but lack in the United States. Diesels post significantly better distance per volume of fuel (miles per gallon, furlongs per peck) than their gasoline counterparts. I’ve always wondered how much of this disparity can be accounted for by diesel fuel’s higher energy content. A quick search turned up some statistics (See table B-4) produced by Oak Ridge National Laboratory’s Center for Transportation Analysis which lists a gallon of diesel as containing 38.6 MJ per liter and gasoline as posesing 34.8 MJ/L. Not that big of a difference; gasoline has 90% the energy content of diesel.

How is this reflected in the mileage of a real car? Well, let’s look at a Volkswagen Jetta which, until recently in the US, could be had with a 1.8 liter gasoline engine or a 1.9 liter diesel. The gas engine gives 19 MPG city, 27 MPG highway for an EPA combined cycle of 22 MPG (VW Jetta 1.8L gas). The diesel gets 28 MPG city, 39 MPG highway and 32 MPG combined (VW Jetta 1.9L diesel). Working from the combined numbers, the diesel powerplant achieves 45% better mileage with fuel that contains just over 11% more energy content. Not a bad deal at all.

The next question is how this all works out for the environment. Historically, diesels have produced much more environmentally destructive exhaust containing more NOx and particulates. Recently, however, diesel fuel in the United States has been switched to an ultra-low sulfur variety which, in turn, allows for particulate emission control technologies to be employed, reducing the airborne matter released by diesel engines. Do the changes mean significantly cleaner emissions? I’m not sure, because the stats are difficult to find. Perhaps I’ll find them later.

If I had this kind of cash, I would be of the same opinion:

Myhrvold’s friends, like Myhrvold, seemed to be of the opinion that there is no downside to having a CAT scanner, especially if you can get it for twenty-nine hundred dollars.

The New Yorker has a very interesting article by Malcolm Gladwell, author of The Tipping Point about Intellectual Ventures, a company that does nothing but think up inventions. They occasionally call sessions to which smart people are invited to talk and come up with patentable ideas, hand them to lawyers and then sell the patents.