On a positive note, I spent some time improving the client-server communication schema for the voiceServer and face applet over the last few weeks. The old method used a CGI-based polling technique that was rather slow and inefficient even with mod_perl. The new method uses a Perl POE-based server-side interface (POEfaceClient) to manage each of the connected face applets. This method greatly reduces the amount of handshaking needed to keep current with the voiceServer's cue stack. Another advantage that using shared memory provides is that backward compatablity is retained, so clients connecting through a firewall can still use the old interface method if needed.
The new client handshaking technique required re-writing several key areas of the face applet, so I re-named it zazaface2 so there won't be any caching problems with older browsers. The 2.01 release supports auto-reconnect on socket error, and can now handle server disconnect gracefully. I'll probably add an auto-fallback to CGI handshaking option in 2.02 and put a few of the options as applet parameters instead of hard-coding them.
The Tech received a rather sizable donation of Intel PRO/Wireless 5000 (802.11a and 802.11b) wireless LAN gear and is getting ready to install it. This opens up all kinds of options for Zaza, including real-time high-rate video and audio streaming. I have started the search for a ethernet-802.11a bridge that will alow use of both of the onboard computers, or alternatively a PCI card that has linux drivers available and an external entenna that can be positioned somewhere in the acrylic hood...

del.icio.us
Digg
Google bookmark
reddit
Simpy
StumbleUpon
Furl
Newsvine
Technorati
Tailrank