Da ich schon seit Langem immer fleißig die .csv files zu den Rechnungen herunterladen hab ich schnell folgendes Python Script geschrieben das mir die Zählarbeit erleichtert (und Schreikrämpfe in Excel verhindert).

Als kleinen Bonus zählt es auch (etwas inexakt fürchte ich) die Homezone Minuten, daraus kann man ablesen was das Verlieren von Genion für Auswirkungen hat.

# parses .csv files from o2 germany billing
# counts talked minutes ('sprache' in 'ART') and sms sent ('sms' in 'TARIFGRUPPE')
# ignores international minutes/sms
import csv
import glob
import string

# <CONFIG>
mypath = '*.csv' # e.g. './mybills/*.csv'
mydelimiter = '|' # '|' aka "pipe" is default with o2
# </CONFIG>

if __name__ == "__main__":
    files = glob.glob(mypath)
    assert len(files) > 0, "Error! No Files found (check path)"

    totalminutes = 0
    totalsms = 0
    totalhomezone = 0
    for file in files:
        csvfile = open(file)
        reader = csv.DictReader(csvfile,delimiter=mydelimiter)
        smscount = 0
        timecount = dict(hours=0,minutes=0,seconds=0)
        homezonecount = 0
        for row in reader:
            if 'sprache' in string.lower(row['ART']):
                (h,m,s) = string.split(row['DAUER'], ':')
                timecount['hours'] += int(h)
                timecount['minutes'] += int(m)
                timecount['seconds'] += int(s)
                if 'homezone' in string.lower(row['TARIFGRUPPE']):
                    homezonecount += int(m) + (int(h)*60) + (int(s)/60)
            if 'sms' in string.lower(row['TARIFGRUPPE']):
                smscount = smscount + 1
        minutes = timecount['minutes'] + (timecount['hours']*60) + (timecount['seconds']/60)
        totalminutes += minutes
        totalsms += smscount
        totalhomezone += homezonecount
        print(file+' '+str(minutes)+' Minutes '+str(smscount)+' SMS ('+str(homezonecount)+' Minutes in Homezone)')
        assert minutes >= homezonecount, "Error! More minutes in homezone than total?"
    # the cool .format prints only work in python 2.6 ...
    #print('Total  :{0:4} Minutes and {1:4} SMS in {2} Months ({3:4} Minutes in Homezone)'.format(totalminutes,totalsms,len(files),totalhomezone))
    #print('Average:{0:4} Minutes and {1:4} SMS per Month'.format(totalminutes/len(files),totalsms/len(files)))
    # ... and since that isn't in debian stable yet... i added "old" prints
    print('Total   '+str(totalminutes)+' Minutes '+str(totalsms)+' SMS in '+str(len(files))+' Months ('+str(totalhomezone)+' Minutes in Homezone)')
    print('Average '+str(totalminutes/len(files))+' Minutes '+str(totalsms/len(files))+' SMS per Month')

Must see, really cool take on the topic. Complete with disturbing images of a north-european distopian, snowcovered town ;)

Far better than the twilight movie, which sucked (i liked the book though, i read all four).

<?xml version="1.0"?>
<xsl:stylesheet version="1.0"
  xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
  <!--
    ! Small XSLT prog that calculates the faculty of one or more
    ! numbers by recursively calling a named template
    ! numbers are expected to be in a tree:
    ! <nums><num>1</num><num>2</num>...</nums>
   -->

  <xsl:output method="text"/>

  <xsl:template match="num">
      <xsl:text>fak(</xsl:text>
      <xsl:value-of select="."/>
      <xsl:text>) is </xsl:text>
      <xsl:call-template name="fak">
          <xsl:with-param name="number" select="."/>
      </xsl:call-template>
  </xsl:template>

  <xsl:template name="fak">
      <xsl:param name="number"/>
      <xsl:choose>
          <xsl:when test="$number = 1">
              <xsl:value-of select="$number"/>
          </xsl:when>
          <xsl:otherwise>
              <xsl:variable name="recursefak">
                  <xsl:call-template name="fak">
                      <xsl:with-param name="number" select="$number - 1"/>
                  </xsl:call-template>
              </xsl:variable>
              <xsl:value-of select="$number * $recursefak"/>
          </xsl:otherwise>
      </xsl:choose>
  </xsl:template>
</xsl:stylesheet>

Expected XML looks like this (with DTD even! ;P):

<?xml version="1.0" ?>
<!DOCTYPE numbers [
  <!ELEMENT nums (num+)>
  <!ELEMENT num (#PCDATA)>
]>
<nums>
  <num>3</num>
  <num>12</num>
</nums>

XKCD 528 <img alt=“Disclaimer: I have not actually tried the beta yet. I hear it’s quite pleasant and hardly Hitler-y at all.”>

ps) i saw the (leaked) beta in vmware and it actually was quite nice ;)
pps) the horde on webspace (no shell) guide will be delayed until my host decides to fix his PHP installation (imap_open() is broken…)

Part 2 The Database
Since without shell access we can’t use Horde’s setup (./scripts/setup.php) to do this for us we have to manually set up the database for Horde.

This step depends on your webhosters so i’ll just make it quick:
First create a database with your webhost, then find: hostname (often locahost), username, password, type (often tcp) and databasename for that database.

Now your webhoster probably offers some kind of management frontend (mostly phpmyadmin). There you navigate to your database and select the “SQL” tab.

In the editbox you will find there you have to paste some of the lines from the .sql file provided with Horde that matches your database. I use mysql so i used ./scripts/sql/create.mysql.sql.

Out host already created the database, the user etc. so we only need lines that create tables. In my file this would be lines 52 -> end (look for the first CREATE TABLE statement). Copy&Paste and press Submit. This should have given you no errors and created ~18 tables in your database.

In the next step we will create a basic configuration that allows Horde to run so we can do the rest of the setup from within it.

This experiment will most likely fail because i don’t have shell access and the php installation is probably missing vital modules .. but what else is there to do on a saturday evening

Part1: Horde is f***ing huge!


du -sch .
108M total


I’ll start by making horde a little slimer…
This will give me all the folders with locales other than german or english (cd to horde first)

find -iname ??_?? -type d | grep -v de_DE | grep -v en_US | grep -v en_GB

and this will delete them

find -iname ??_?? -type d | grep -v de_DE | grep -v en_US | grep -v en_GB | xargs rm -r


The same for the “.po” files:

find -iname ??_??.po -type f | grep -v de_DE | grep -v en_US | grep -v en_GB | xargs rm -r



du -sch .
51M total

Half the size, nice :)

An onomatopoeical adjective based on the sound emitted when something is “oh so tasty” (either through hunger or flavorological value) that one gnaws through it without regard to cleanliness or etiquette.

These Prices are in German Marks, which back then was maybe half to two thirds of a dollar (date is in MM/YY form and the comma marks the decimal point, so first entry reads 62 Mark and 3 Pfennig)

07/98: 62,03
08/98: 190,66
09/98: 193,22
10/98: 174,92
11/98: 176,06
12/98: 231,59
01/99: 179,12
02/99: 207,84
03/99: 237,94
04/99: 238,27
05/99: 226,31
06/99: 137,04

Back then you had 56kbit with a modem or 64kbit with ISDN later and you had to make every minute count ;) All of this was to play Ultima Online in the evenings and i was still in school back then … kind of an expensive hobby :)
Comparing this to today’s DSL market really makes you cry ;)
You get 16mbit down 1mbit DSL up for 20 euros flat or even 50mbit down 10mbit up VDSL for 70 euros flat. Kids today have it way to easy! :)

It worked fine, currently I set it to the newest tag, but in future I might switch to the trunk, maybe then finally something breakes that I can then fix ;)

Title is once again inspired by a song I currently like, which was used as the music in this lovely “hooping” video: