try to explain what wave will be: its email, im, chat, collaboration in one sleek interface. and its all in the browser with heavy usage of bleeding edge technology like the not yet finished html 5. in wave the basic thing is a “wave”. consider it like a im/email-conversation on steroids. if you now google mail you should be familiar with its great conversation view. that is showing emails with similar subject grouped together. although the ones you send (i thnk this is the most compelling feature of gmail and i havent found this done right in any other mail client). in wave this is an important feature too. everything is grouped in conversations – sorry “waves”. but whats the difference? the difference is, that every member of the wave can edit the wave on every point in time. meaning you can go back to the initial post and add your message there or append it to the end. all changes can be replayed using a cool history playback.

for a nerd like me it looked all very cool. but i think it is very complex. perhaps too complex for big audience. i think google is aiming for everybody…but the usual user will be overwhelmed by the possibilities. to grasp the hole flow of a conversation over time could be very hard despite the playback. so for now i cant see email going away soon. a very interesating point they didnt mention in the presentation is the spam-problem. also since i am using gmail for a long time now my spam-problems are nearly vanished i think that is one of the most annoying things in email. could they find a way to eliminate that with wave?

but one of the best things with wave is the open source approach. they will open all stuff they already have. the protocol will be open and there will be a sophisticated plugin architecture to add many features (as they showed twitter, bug tracking, chess, sudoku). and what i understand one can host a wave server by himself so the most privacy issues for business could be solved with this as wave wont send any stuff to google that is intranet.

so i am very excited to see how it will be. but it seems we need to wait for the end of the year to see it live.

apt-get install trac

after that trac is installed in /usr/share/trac. to create your trac-environment you have to call the admin console program from trac:

trac-admin /var/www/trac initenv

where /var/www/trac is the place where your website will go and initenv is the command to tell trac-admin to initialise a new environment. next i setup apache to serve the new site. there are a bunch of possibilities to let apache server trac: cgi, fastcgi, mod_python etc. i tried cgi and fastcgi first but the mod_python version worked at last. so i need to install mod_phyton:

apt-get install libapache-mod-python

ok got some tips for the site config out of the net. here is my virtualhost-config part:

# Trac Configuration
<VirtualHost *>
ServerName wop.dotob.de
DocumentRoot /var/www/trac/
<Location / >
SetHandler mod_python
PythonHandler trac.web.modpython_frontend
PythonInterpreter main_interpreter
PythonOption TracEnv /var/www/trac/
PythonOption TracUriRoot /
AuthType Basic
AuthName “trac”
# Use the SVN password file.
AuthUserFile /var/svn/.dav_svn.passwd
Require valid-user
</Location>
</VirtualHost>

sorry i assumed you have set your subversion on the same server an it is already up but i wont tell how this is done here. thats said: the /var/svn/.dav_svn.passwd is my subversion authentication file. i use the svn authentication for trac too.

well now trac is up and running. you can use trac-admin, which is a great commandline tool with tab-completion and all to config stuff like when you create a ticket which level of severity are displayed or what the name of your milestones is. what i suggest also is to set some of the user permissions so you can change more stuff via the webinterface.

another thing i wanted is to let trac look into subversion commits an retrieve some information from my commit messages. so assume i have a ticket #5 and i check some fix for that. now i can write “fix #5 introduce new style in xaml” in the commit message and trac will close the ticket #5 with the svn version mentioned and the commit message attached. so how is this done:

subversion uses hook files to let users run scripts. so you go to your subversion repository (assume /var/svn/src) into the directory hooks. rename the post-commit.tmpl to post-commit. than that file is executed by subversion. now you need to insert the code from trac:

REPOS="$1"
REV="$2"
LOG=`svnlook log -r $REV $REPOS`
AUTHOR=`svnlook author -r $REV $REPOS`
TRAC_ENV='/var/www/woptrac'
/usr/bin/python /usr/share/doc/trac/contrib/trac-post-commit-hook -p "$TRAC_ENV" -r "$REV" -m "$LOG" -u "$AUTHOR"

i found some tutorials where the LOG part was missing but i couldnt get it to work without it. another thing that cost me a lot of time: when you write the commit it didnt work for me to write: “Fixes: #5″. i think the : between the command Fiexs and the #5 was not considered by the scripts regex.

i am curious about using trac. will tell how it behaves.

this is  a memory dump of the things we where talking about at the lixhot-boys meeting today.

-caltodo, sync todo from iphone to your mac
-lacie 5big, raid network drive
-qnap 439pro, raid network drive
-raid, wikipedia entry
-fotolia, buy photos online
-istockphoto, buy photos online
-tweetdeck, nice adobe air powered twitter client
-twitterific, iphone twitter client
-blue, beautiful vista twitter client
-scotts twitter intro
-evernote, multiplatform note organizer
-dropbox, online storage
-subversion, nice version control system
-mozy, unlimited online backup
-keynote, presentations on the mac
-wptouch, wordpress iphone plugin
-claas tajtes-blog, business and political oriented german blog

was a nice evening. had some kölsch and a good talk. hope to see you soon again.