I am working on yet another project which is using My Favourite Language®. This time, I am using the (rather new) TurboGears framework. TurboGears is an amalgamation of many different sub projects into a fairly lightweight web application foundation. They also provide an easy to follow tutorial which gives a pretty solid overview of the system. I highly recommend also going over the CherryPy documentation as well.

Following is a set of notes on the tutorial, and some help with fixing the errors I ran into.

Could it be true? While it sounds like the run up to a bad sci-fi movie, it appears that the cold side of the deep ocean currents that convery warm water back to the surface in Europe, may be slowing down substantially. These deep cold currents bring water to the warmer equatorial regions, where warm water is displaced, flowing back to European waters, resulting in unusually warm water for their distance from the equator. Especially troubling is Europes dependence on agriculture in the regions that would be hit the hardest by a dip in temperature.

This (huge) video shows a pretty good explanation of the entire system.

Incidentally, there is historical evidence that this has happened before. Natural climate change caused a "similar" plunge to low temperatures over a very short timespan. Evidently we should be careful of our own effects, since the climate is fragile enough without our interference...

One of the neat new applications I am getting the opportuniyt to play with (thanks to one of my clients), is OpenLaszlo. OpenLaszlo is a web application framework that provides an asynchronous web application framework which is not based on AJAX. OpenLaszlo is a Tomcat based Java app, which provides a very rich XML based (looks like html) programming language, which allows you to generate Flash based "Asynchronous Flash and XML" (AFAX?) applications. Some of the bonuses are better animation and graphical capabilities.

The drawback is the need to support a Tomcat install on top of your Apache instance, but that is not so bad, since the OpenLaszlo package is SO GOOD. There is even a beta quality IDE for eclipse! My current job is writing an embedded video player which will play movies described in an RSS feed. I have already written a bridge to allow Typo3 to generate and compile Laszlo scripts.

So we finally bit the bullet and purchased a "real" espresso machine. We had been suffering through far too many Tim Hortons (which are admittedly the best of the drip coffee purveyors), and decided that enough was enough.

The Silvia is a craftsperson's machine, one which I have to admit I do not really know how to use yet. It is unforgiving of poor technique, and requires a high quality grinder (which I do not have yet either) to get decent results. It does however generate some of the best coffee I have ever had when used correctly. We ran off several dozen shots over the last couple of days in order to test it, and practice. We learned some hard lessons about tamping and grind consistency.

The local coffee shop (Coffee King) was very helpful in supplying "throwaway" beans that we could use for practice roasting, as well as some really good house blend for the actual keeper espressos.

Now that I have the base equipment, I suppose that my geek side is going to take over. I am already contemplating a PID Loop modification...

Update: Helen and I have both gotten some pretty amazing shots out of the machine tonight. The timing and volume seem to be the issue. To be short, the perfect espresso shot seems to be a light tamp on slightly finely ground coffee, and 25-28 seconds to pour the full 2 oz shot.

Isn't it strange how your time seems to slip away on you? I had really planned on keeping this blog updated on a regular basis after getting back from Defcon, but I ended up going on several more business trips, and never really had the time.

Todo list for the next few days:

• Torrent server for RantMedia
• Fly to Toronto
• Build a massive CMS system
• Build a (potentially) even more massive POSTFIX install
• Fly to Dallas to deliver on the previous two projects
• "Learn some folks" about LDAP
• Update Cthuugle
• Finalize my "embedded CAN bus project"
• Go to someplace in the carribean and turn the brain OFF

P.S. Hello to Buck, Radek, Jason, and anybody else who thought I was dead ;)

Unfortunately, the hotel mistakenly billed me for $757US in room charges, which I will have to fight over.

I did however get some cool pix. The gallery is chock full of Defcon alumnus.

Maybe I should ask Major to give me a hand with my troubles ;)

I leave for Vegas in the morning, to run video/cable/networking stuff for Defcon. I am betting that I will have a blast. Going with me is Sparky, who will be handling some video tasks. Hopefully I will bring back a ton of good pics, and maybe and interview or two.

Just a quick set of updates:

• The Mini-itx board combo server for the channel scroller at defcon is working. Now I can use my laptop while I am there
• The avr-gcc crossdev AVR GCC package on gentoo is fux0r3d, but you can copy the avr-gcc binaries into /usr/bin to fix the problem. Done
• THe belligerator's accelerometer is now connected. See previous.
• I am working on a giant form generation project for a local client

The poor little belligerator has not made much progress as of late, but I have finally attached the head, and the rate sensor is connected. There is a little drift (around 1%) which I have been able to remove with a little clever code, but I have yet to set up the tilt sensor, or to code the kalmann filter.

I am also using a boot loader on this project, since I have finally tired of detonating avrisp programmers.

Next on the plan is setting up the PID loop, extending the torso (the mass moment of inertia is way too low right now)

Once this device is balancing, I will re-evaluate whether to continue with this small scale robot, or build something larger. Paradoxically, it is often cheaper to build big than small, and big allows for more interesting mistakes ;)

PyChannels now has a complete set of features. I am going to fix a couple of things so that the correct channel info displays, but the paragraph rendering and vertical scrolling is now completely debugged. Memory loss (the scourge of Python server coding) is also manageable right now.

Feel free to download the existing code and give it a try.