$ LANG=C time -p sort --buffer-size=128M -n -k1 -k2 test.info > /dev/null
real 92.58
user 86.04
sys 6.52
$ time -p sort --buffer-size=128M -n -k1 -k2 test.info > /dev/null
real 463.04
user 456.01
sys 6.64

P.S. grep and other locale sensitive utilities have similar issue.

BTW, if your disk is slow, --buffer-size is a parameter you might want to play with to optimize performance. By default sort uses very conservative buffer size while your seemingly used memory is actually just used for cache. [ref]

In the early 1950’s, Zacharias attempted to make a fountain of atoms by directing a thermal atomic beam upwards. [2]

[2] Knowledge of J. R. Zacharias’s work has been passed down in the oral tradition. …

ffmpeg -i input.avi -f mp4 -vcodec mpeg4 -b 1000kb \
-acodec aac -ab 192kb [-ac 2] output.mp4

“-ac 2” may not work if the input file is in mkv format with 5.1 aac audio (if you don’t convert 5.1 to 2.0, playing the file crashes your iPod/iPhone). Here’s a possible way to convert:

• Extract aac audio from the mkv file:

  ffmpeg -i input.mkv -vn -acodec copy audio.aac

• Downmix 5.1 to 2.0:

  faad -d audio.aac -o audio-2ch.wav

• Merge and convert to iPod format:

  ffmpeg -i input.mkv -i audio-2ch.wav -map 0.0:0 -map 1.0:1 -f mp4 \
  -vcodec mpeg4 -s 470x320 -b 1000kb -acodec aac -ab 192kb output.mp4

• The output of the last command gives a little more info of what I am doing here:

  Input #0, matroska, from 'input.mkv':
  Duration: 01:24:42.5, start: 0.000000, bitrate: N/A
  Stream #0.0: Video: h264, yuv420p, 704x480, 24.39 fps(r)
  Stream #0.1: Audio: aac, 48000 Hz, 5:1
  Input #1, wav, from 'audio-2ch.wav':
  Duration: 01:24:42.5, start: 0.000000, bitrate: 1536 kb/s
  Stream #1.0: Audio: pcm_s16le, 48000 Hz, stereo, 1536 kb/s
  Output #0, mp4, to 'output.mp4':
  Stream #0.0: Video: mpeg4, yuv420p, 470x320, q=2-31, 1000 kb/s, 24.39 fps(c)
  Stream #0.1: Audio: aac, 48000 Hz, stereo, 192 kb/s
  Stream mapping:
  Stream #0.0 -> #0.0
  Stream #1.0 -> #0.1

Video conversion can certainly be way fancier than what I am doing here, and faad is obviously not the best way to downmix audio from 5.1 to 2.0. You can also find quite a few iPod converters but personally I like command line. Anyway, happy converting! (ffmpeg ref)

[Also note that on MacOS, you may have to change "-acodec aac" to "-acodec libfaac".]

• Case 1: My colleague worked overnight, so I sent an email to him in the morning: “Great work, have some good sleep then.”. Gmail popped up a sponsored link: “Sleep Apnea Cure”. Oh come on, my colleague sleeps very well…
• Case 2: My friend did not show up in a scheduled meeting, and he emailed me: “Sorry I screwed up today….  Could we meet late afternoon tomorrow”. The Google ads beside this email was “Date Hot Stanford Girls!”. ok, ok, I guess Google knows my geolocation, and “screw up” does have another meaning…

• Q: Matlab desktop does not show up correctly.
  A: export MATLAB_JAVA=/usr/lib/jvm/java-6-sun/jre/ (requires Sun Java 6 Runtime)
• Q: Matlab windows disappear in taskbar (Window List applet).
  A: minimizing all windows (show desktop button) seems to bring all windows back into taskbar. Or you can try to export AWT_TOOLKIT=MToolkit
• Q: Printing in Matlab throws java exception.
  A: Go set all printers > Job Options > Orientation to something other than “Automatic Rotation”.

Links: 1, 2.

To convert a Microsoft Visio document to EPS without losing scalable vector format (and without installing 3rd-party software…):

1. Install Microsoft Save as PDF or XPS Add-in for 2007 Microsoft Office programs (only available for Office 2007?).
2. Save Visio document as PDF in Visio 2007.
3. Convert PDF to Postscript (.ps) with pdf2ps utility.
4. Open .ps file with gsview, and select File -> PS to EPS, be sure to check “Automatically calculate Bounding Box”.

Now you have an EPS file to be embedded into your LaTeX document. Many other converting methods exist, see also Visio and EPS.

Here are some updates (edited on 2009.01.30):

1. In theory, you should be able to open that saved PDF directly in gsview and convert to EPS, but gsview either crashes or refuses to accept the PDF.
2. In some cases, e.g. when you compile LaTeX directly to PDF using pdflatex, you just need a PDF figure with correct size. In that case, you can set the document size in Visio: File -> Page Setup -> Page Size (tab) -> Size to fit drawing contents, and then save the Visio document to PDF, which will then be in correct size ready to be embedded into LaTeX.
3. To create a Postscript file, you don’t necessarily need to save as PDF and convert to .ps using pdf2ps. All you need to do is to use a postscript printer (most modern printers are, including most PDF writers) and print to file. That would create a postscript file which can then be passed to gsview and converted to EPS by adding a bounding box. Note: this is in fact a generic way of converting anything to (vector-format) EPS/PDF with bounding box.

1. Stop/restrict web service.
2. Dump the database:

   $ mysqldump --default-character-set=latin1 --databases wordpress > m.sql

3. Change encoding settings in m.sql, including table default encoding and client communication encoding:

   $ replace "CHARSET=latin1" "CHARSET=utf8" \
       "SET NAMES latin1" "SET NAMES utf8" < m.sql > m2.sql

4. Import the “converted” sql:

   $ mysql < m2.sql

5. Add the following lines to WordPress wp-config.php

   define('DB_CHARSET', 'utf8');
   define('DB_COLLATE', '');

6. Reopen web service. That’s it.

Note: For future convenience, you might as well change the database default encoding to utf8.

Ref to WordPress documentation for details: here and here.

I installed CHDK on my Canon A570 IS camera and shot a few test pictures to find out defective pixels of my camera. To my surprise, this 7.1 mega-pixel CCD sensor has couple thousands defective pixels. The following picture represents the raw output from CCD sensor when I shot a 15 sec picture with ridiculously bright light source: (The green border is just showing image boundary, not from the actual picture.)

The raw file is 3152 columns by 2340 rows, 10 bits per pixel, so the file size is 3152 x 2340 x 10 bits = 9219600 bytes. The first few columns and rows are used for background subtraction (masked pixels, their values are usually around 30-40 on this 2¹⁰ = 1024 scale), so the active area is actually 3072 x 2304 pixels. The above picture is RAW CCD output down scaled to 0-255 and converted to 8-bit gray PNG (i.e., before demosaicing/gamma correction/any other process taken place). As you can see, there’re a lot (9205) of 0-valued defective pixels in the active area. There is no defective pixel in the masked area.

When I shot with faster shutter speed, I have less (2991) defective pixels. The following was shot with 1/250 sec shutter speed, showing CCD value 0 as black (defective pixels), 1-40 as gray (masked pixels), and 40-267 as white. No pixel has value more than 267.

I shot a lot of RAW pictures and found the number of defective pixels is either 9205 or 2991. Generally speaking, the slower the shutter speed, the more chance I get 9205 defective pixels. But I did have 2991 defective pixels with 1/30 sec picture, and 9205 defective pixels with 1/250 sec picture, with all other settings the same. So I have no idea what determines the defective pixel pattern. Also what’s interesting is those 2991 defective pixels are subset of the 9205 defective pixels.

If I go ahead and demosaic, combining RGB channels together, This is part of the image I get: (You can see defective pixels in different color channels.)

Instead, most RAW format processors find those 0-valued (sub)pixels and interpolate over with neighbor pixels. By doing that, here’s what I get: (white balance etc. was tuned a little bit as well)

Nicely white, no annoying dots any more, right? This “cheating” process is done inside the camera when saving JPEG file, and also done in most RAW format conversion. DNG specification states that:

Defective pixels should be mapped out (interpolated over) before the raw data is stored as DNG.

Enjoy your camera with 0.1% defective pixels!