diff --git a/gcompris.py b/gcompris.py
index 2b35bbc..4d1b507 100755
--- a/gcompris.py
+++ b/gcompris.py
@@ -1,471 +1,470 @@
-#!/usr/bin/python2
+#!/usr/bin/python3
 # -*- coding: utf-8 -*-
 
 import codecs
 import os
 import sys
 import jinja2
 import time
 import datetime
 from datetime import date
 import re
-from collections import OrderedDict
 import gettext
 from email import utils
+from importlib import reload
 
 reload(sys)
-sys.setdefaultencoding('utf-8')
 
 today = date.today()
 
 try:
     version = sys.argv[1]
 except:
-    print "Missing GCompris version"
+    print("Missing GCompris version")
     sys.exit(1)
 
 try:
     locale = sys.argv[2]
 except:
-    print "Missing GCompris locale"
+    print("Missing GCompris locale")
     sys.exit(1)
 
 try:
     localelist = sys.argv[3]
 except:
-    print "Missing GCompris list of locales"
+    print("Missing GCompris list of locales")
     sys.exit(1)
 
 try:
     gcomprisdir = sys.argv[4]
 except:
-    print "Missing GCompris installation directory, export GCOMPRIS_DIR as env var"
+    print("Missing GCompris installation directory, export GCOMPRIS_DIR as env var")
     sys.exit(1)
 
 # Load the proper locale catalog
 _ = None
 t = None
 def setLocale(locale):
     global _
     global t
     try:
         t = gettext.translation('gcompris', 'locale', languages=[locale])
     except:
         t = gettext.NullTranslations()
-    _ = t.ugettext
+    _ = t.gettext
 
 def formatDate(date):
     return date[0:4] + '-' + date[4:6] + '-' + date[6:8]
 
 #
 # Parse the config.c file in GCompris source code to get
 # the name of the given locale.
 #
 def getLocaleName(locale):
     result = locale
     try:
         with open(gcomprisdir + '/src/core/LanguageList.qml') as f:
             content = f.readlines()
             for line in content:
                 m = re.match('.*\"text\": \"(.*)\", \"locale\": \"(.*)\" }', line)
                 if m and m.group(2).startswith(locale):
                     result = m.group(1)
                     break
     except IOError as e:
         pass
 
     return result
 
 # Set the default locale
 setLocale(locale)
 
 #
 # Create locales [ [ locale, language ], ...]
 #
 locales = []
 language = ""
 for loca in localelist.split():
     # We want each country in its own language
     setLocale(loca)
     lang = _(getLocaleName( loca ))
     locales.append( [loca, lang] )
     if locale == loca:
         language = lang
 
 # Add en_US manually
 setLocale("en")
 locales.append( ['en', _(getLocaleName( 'en_US'))] )
 if language == '':
     language = 'English'
 
 # Back to the default locale
 setLocale(locale)
 
 locales = sorted(locales, key=lambda t: t[1])
 
 suffix = '-' + locale
 
 #
 # We don't have a translation of each manual. If we have it
 # we return it else we return the english one
 def getManual():
     manuals = {
         'en': 'wiki/Manual',
         'fr': 'wiki/Manuel',
         'de': 'wiki/Benutzerhandbuch',
         'es': 'wiki/Manual_es',
         'pt_BR': 'wiki/Manual_pt-BR',
         'he': u'wiki/מדריך_למשתמש',
         'ru': u'wiki/Руководство'
         }
     if locale in manuals:
         return manuals[locale]
     return manuals['en']
     
     
     
 descriptions = []
 
     
 def getBoards():
     '''create a list of activity infos as found in GCompris ActivityInfo.qml'''
 
     activity_dir = gcomprisdir + "/src/activities"
     for activity in os.listdir(activity_dir):
         # Skip unrelevant activities
         if activity == 'template' or \
            not os.path.isdir(activity_dir + "/" + activity):
             continue
         
         try:
             with open(activity_dir + "/" + activity + "/ActivityInfo.qml") as f:
                 content = f.readlines()
                 
                 description = ''
                 name = ''
                 title = ''
                 credit = ''
                 goal = ''
                 section = ''
                 author = ''
                 manual = ''
                 difficulty = ''
                 demo = ''
                 category = ''
                 prerequisite = ''
                 icon = ''
 
                 for line in content:
 
                     m = re.match('.*description:.*\"(.*)\"', line)
                     if m:
                         description =  m.group(1)
 
                     m = re.match('.*name:.*\"(.*)\"', line)
                     if m:
                         name = activity
                         icon = activity
                     
                     m = re.match('.*title:.*\"(.*)\"', line)
                     if m:
                         title = m.group(1)
                     
                     m = re.match('.*credit:.*\"(.*)\"', line)
                     if m:
                         credit = m.group(1)
                     
                     m = re.match('.*goal:.*\"(.*)\"', line)
                     if m:
                         goal = m.group(1)
                     
                     m = re.match('.*section:.*\"(.*)\"', line)
                     if m:
                         section = m.group(1)
                     
                     m = re.match('.*author:.*\"(.*)\"', line)
                     if m:
                         author = re.sub("<.*?>", "", m.group(1))+(' & Timothee Giet')
                     
                     m = re.match('.*manual:.*\"(.*)\"', line)
                     if m:
                         manual = m.group(1)
                     
                     m = re.match('.*difficulty:.*', line)
                     if m:
                         difficulty = (m.group(0)).replace('  difficulty: ', '')
                         difficulty = (difficulty).replace(' ', '')
                         
                     
                     m = re.match('.*demo:.*', line)
                     if m:
                         demo = (m.group(0)).replace('  demo: ', '')
                         
                     m = re.match('.*type:.*\"(.*)\"', line)
                     if m:
                         category = m.group(1)
                         
                     m = re.match('.*prerequisite:.*\"(.*)\"', line)
                     if m:
                         prerequisite = m.group(1)
                         
 
                     infos = {'description':description, 
                              'name':name, 
                              'title':title, 
                              'credit':credit,
                              'goal':goal, 
                              'section':section,
                              'author':author,
                              'manual':manual,
                              'difficulty':difficulty,
                              'demo':demo, 
                              'type':category,
                              'prerequisite':prerequisite,
                              'icon':icon}
                 
                 descriptions.append(infos)    
 
         except IOError as e:
             pass
     
     return descriptions
 
 # /a/b /a => 1 0
 # /a/b /c => 2 1
 # /a/b /a/b => 0 0
 # /a '' =>  1 0
 # /a/b/c/d /a/b/e/f => 2 2
 # /a /a/b/c => 0 2
 #
 # Return (open, close)
 #
 def sectionDiff(old, new):
     if old == new:
         return (0, 0)
 
     olds = old.split("/")
     news = new.split("/")
 
     closes = 0
     opens = 0
 
     for i in range(0, max(len(olds), len(news)) ):
         try:
             if olds[i] == news[i]:
                 continue
             else:
                 closes += 1
                 opens += 1
         except:
             if len(olds) > i:
                 closes += 1
             else:
                 opens += 1
 
     return (opens, closes)
 
 templateLoader = jinja2.FileSystemLoader( searchpath="." )
 templateEnv = jinja2.Environment( loader=templateLoader,
                                   extensions=['jinja2.ext.i18n'],
                                   trim_blocks=True,
                                   lstrip_blocks=True)
-templateEnv.install_gettext_callables(t.ugettext, t.ungettext, newstyle=True)
+templateEnv.install_gettext_callables(t.gettext, t.ngettext, newstyle=True)
 
 # Specify any input variables to the template as a dictionary.
 templateVars = {
     "locale" : locale,
     "suffix" : suffix,
     "language" : language,
     "vocabularyActivity" : _("Enrich your vocabulary"),
     "revision_date" : today.strftime("%Y-%m-%d"),
     "current_year": today.strftime("%Y"),
     "version": version,
     "news": [],
     "screenshots": [],
     "screenshotsmenu": [],
     "locales": locales,
     "manual": getManual(),
     "manualTranslation": _("Manual"),
     "license_info": _("This software is a GNU Package and is released under the GNU General Public License"),
     "feed": []
     }
 
 # Use this filter in template when you know the text comes from the
 # software part and should not be added to the web site po file.
 def trans(text):
     return _(text)
 templateEnv.filters['trans'] = trans
 
 #
 # Build a map of all news file to proceed for our locale
 #
 filenames = {}
 for filename in os.listdir("news"):
     if not filename.endswith(".html"):
         continue
 
     # If a news is found with a -<LOCALE> suffix before .html
     # It supercede a news without a such suffix (english one).
     filename_noext = filename.split('.')[0]
     try:
         (dat, loc) = filename_noext.split('-')
         if locale == loc:
             filenames[dat] = filename
     except:
         if not filename_noext in filenames:
             filenames[filename_noext] = filename
 
 for filename in sorted(filenames, reverse=True):
     filename = filenames[filename]
     templateOneNews = templateEnv.get_template("news/" + filename)
     templateVars['newsDate'] = formatDate(filename)
     templateVars['fileName'] = filename
     templateVars["news"].append(templateOneNews.render(templateVars))
 
     # read file to get the title, not sure if it is doable using jinja
     lines = ""
     with open ("news/" + filename, 'rt') as in_file:
         for line in in_file:
             lines += line.rstrip('\n')
     rgx = re.compile('set title = \'(?P<name>[^{}]+)\'')
     variable_names = {match.group('name') for match in rgx.finditer(lines)}
 
     dateRFC822 = utils.formatdate(time.mktime(datetime.datetime.strptime(templateVars['newsDate'], "%Y-%m-%d").timetuple()))
     currentFeed = {"dateRFC822": dateRFC822, "date": templateVars['newsDate'], "title": next(iter(variable_names))}
 
     templateVars["feed"].append(currentFeed)
 
 templateNews = templateEnv.get_template("template/news.html")
 outputNewsText = templateNews.render(templateVars)
 
 templateNewsAll = templateEnv.get_template("template/newsall.html")
 outputNewsAllText = templateNewsAll.render(templateVars)
 
 templateFeedAll = templateEnv.get_template("template/feed.xml")
 outputFeedAllText = templateFeedAll.render(templateVars)
 
 #
 # Get the board list and make some adaptations
 #
 boards = getBoards()
 #print boards
 
 for screenshot in boards:
     if screenshot['name'] == 'menu':
         screenshot['name'] = 'root'
         screenshot['section'] = 'administration'
         screenshot['type'] = 'root menu'
         screenshot['difficulty'] = "0"
 
     if screenshot['name'] == "login":
         screenshot['section'] = 'administration'
 
 boards = sorted(boards, key=lambda t: t['section']+' '+t['name'])
 
 # Reorder root, administration and login boards
 #boards[0], boards[1], boards[2] = boards[2], boards[0], boards[1]
 
 #
 # Now process the board list
 #
 templateScreenshot = templateEnv.get_template("template/screenshot.html")
 previousSection = ''
 depth = 0
 
 for screenshot in boards:
     if screenshot['section'] == '/experimental' or screenshot['name'] == 'experimental':
         continue
 
     #section = screenshot['section']
     #if screenshot['type'] == 'root menu':
         #section += "/" + screenshot['name']
 
     #(opens, closes) = sectionDiff(previousSection, section)
     #depth += opens - closes
     #if closes:
         #templateVars["screenshots"].append("</div>" * closes)
     #if opens:
         #templateVars["screenshots"].append("<div class='row screenshot" + str(depth) + "'>" * opens)
 
     #previousSection = section
 
 
     #screenshot['author'] = re.sub(r" \(.*\)", "", screenshot['author'])
     if screenshot['goal']:
         screenshot['goal'] = _(screenshot['goal']).replace('\n', '<br/>')
     if screenshot['prerequisite']:
         screenshot['prerequisite'] = _(screenshot['prerequisite']).replace('\n', '<br/>')
     if screenshot['manual']:
         screenshot['manual'] = _(screenshot['manual']).replace('\n', '<br/>')
     if screenshot['credit']:
         screenshot['credit'] = _(screenshot['credit']).replace('\n', '<br/>')
     if screenshot['title']:
         screenshot['title'] = _(screenshot['title'])
     if screenshot['description'] and screenshot['description'] != "":
         screenshot['description'] = _(screenshot['description'])
     templateVars["screenshot"] = screenshot
     screenshot["depth"] = depth
     templateVars["depth"] = depth
     templateVars["screenshots"].append(templateScreenshot.render( templateVars ))
 
     # Create the screenshot menu
     templateVars["screenshotsmenu"].append(screenshot)
 
 (opens, closes) = sectionDiff(previousSection, '')
 templateVars["screenshots"].append("</div>" * closes)
 
 templateScreenshots = templateEnv.get_template("template/screenshots.html")
 outputScreenshotsText = templateScreenshots.render( templateVars )
 
 # Count the number of activities
 demo_activities = 0
 total_activities = 0
 for screenshot in boards:
     if screenshot['name'] != 'root':
         total_activities += 1
         if screenshot['demo'] == "true":
             demo_activities += 1
 
 templateVars['total_activities'] = total_activities
 templateVars['demo_activities'] = demo_activities
 
 templateBuy = templateEnv.get_template("template/buy.html")
 outputBuyText = templateBuy.render(templateVars)
 
 templateDownloads = templateEnv.get_template("template/downloads.html")
 outputDownloadsText = templateDownloads.render(templateVars)
 
 templateIndex = templateEnv.get_template("template/index.html")
 outputIndexText = templateIndex.render(templateVars)
 
 with codecs.open('index' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputIndexText )
 
 with codecs.open('news' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputNewsText )
 
 with codecs.open('newsall' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputNewsAllText )
 
 with codecs.open('screenshots' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputScreenshotsText )
 
 with codecs.open('buy' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputBuyText )
 
 with codecs.open('downloads' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputDownloadsText )
 
 with codecs.open('feed' + suffix + '.xml', 'w', encoding='utf8') as f:
     f.write( outputFeedAllText )
 
 template = templateEnv.get_template("template/download_macosx.html")
 outputText = template.render(templateVars)
 with codecs.open('download_macosx' + suffix + '.html', 'w', encoding='utf8') as f:
     f.write( outputText )
 
 if suffix == '-en':
     template = templateEnv.get_template("template/404.html")
     outputText = template.render(templateVars)
     with codecs.open('404.html', 'w', encoding='utf8') as f:
         f.write(outputText)