Comments for Hugh Perkins's tech blog

Comment on Simple samba configuration to let windows users see your files at home by Hugh

Hugh — Thu, 10 Nov 2011 17:00:03 +0000

Note to self: remember to enable the directory in apparmor

Comment on virtualbox + lvm: running virtualbox against lvm partitions by opotonil

opotonil — Sun, 16 Oct 2011 19:32:17 +0000

With qemu-kvm you can use a lvm as user, I think it will work with virtualbox (I am using Gentoo):

# cat /etc/udev/rules.d/99-custom.rules
ENV{DM_NAME}==”vg-virtual”, ACTION==”change”, GROUP=”kvm”

PS: change vg-virtual by vg0-centos and kvm for… ¿user?

Comment on Python, jinja, print and unicode… by Carey Underwood

Carey Underwood — Tue, 11 Oct 2011 03:24:12 +0000

A unicode object is not UTF-8. It’s just abstract codepoints. UTF-8 is a particular way of encoding unicode codepoints into bytes (a str). It’s exactly analogous to encoding objects as json: the json isn’t the objects, it’s just a particular encoding of those objects into bytes.

>>> json.dumps([1,2,3])
‘[1, 2, 3]‘ #a bytestring encoding of the list

>>> json.dumps(u”áṡḑḟ”)
‘”\\u00e1\\u1e61\\u1e11\\u1e1f”‘ #a bytestring encoding of a unicode string

>>> u”áṡḑḟ”.encode(‘utf-8′)
‘\xc3\xa1\xe1\xb9\xa1\xe1\xb8\x91\xe1\xb8\x9f’ # a different bytestring encoding

Comment on Statistical profiling in Java: easy, fast, and useful by Kari’s World » Blog Archive » Analyzing thread time usage

Kari’s World » Blog Archive » Analyzing thread time usage — Thu, 02 Jun 2011 09:16:10 +0000

[...] References: Got idea from Statistical profiling in Java: easy, fast, and useful [...]

Comment on virtualbox + lvm: running virtualbox against lvm partitions by Chuck Payne

Chuck Payne — Thu, 11 Nov 2010 23:30:40 +0000

The one thing you might want to let people know, is user has to be a member of the group disk as well. If not you will get an error.

Comment on Estimating number of words known in mandarin (or any language really) by John Biesnecker

John Biesnecker — Sat, 09 Oct 2010 21:47:26 +0000

It’s been a long time since my last political stats class taught me this stuff, but this little tool http://www.surveysystem.com/sscalc.htm says that you are pretty close, and that for the 2-character word test 50 questions gives you a 10.67% accuracy at 95% confidence, which isn’t too shabby, especially if you run it repeatedly to smooth out any real outliers.

Comment on Estimating number of words known in mandarin (or any language really) by hughperkins

hughperkins — Sat, 09 Oct 2010 13:14:14 +0000

Thanks!

I’m not sure of the exact statistical certainty level. I’m hoping someone else can calculate that for me My gut feeling is that 50 words should get within 10%, maybe 5%, of the actual number. I’m going to run it once a week or so, and see how that goes. Draw a graph over time, as you are saying.

Comment on Estimating number of words known in mandarin (or any language really) by John Biesnecker

John Biesnecker — Sat, 09 Oct 2010 12:51:52 +0000

This is pretty cool. You’d need to run it several times (dozens of times?) to get decent results, though, right? If you’re only picking 0.33% of the words in the dictionary, then the chances of those words being uncharacteristically hard (or easy) is pretty high. Seems like if you tested yourself every day and ran a best-fit line down the results, though, you’d have a decent measure of your progress.

Comment on Bluetooth and eee-control on eeepc debian by Mike

Mike — Sun, 26 Sep 2010 14:45:17 +0000

While trying to run this, I get an error:

in
import ioport
ImportError: No module named ioport

Any ideas ?

Comment on Converting CantoFish dictionary for stardict by Fab

Fab — Thu, 23 Sep 2010 20:19:06 +0000

hughperkins – many thanks!

I had to make the tabfile an excecutable first – this is maybe too obvious for skilled users but nevertheless I finally managed to create the dictionary files and import it into stardict! I’ve learned a lot…thanks again for sharing the valuable code to make it happen!