Technically, nowhere.  I say that because I can't consider this project to be at a 1.0 state until I'm automatically controlling and monitoring the dehumidifier.  Even without Twitter integration (which I think I've figured out - you can follow the kiln at @SuburbanKiln) and data logging to the SD card and display of the data via a web browser (which I've got partly figured out), I'm still not at a point where I'm leaving the kiln to run 24x7.  So what remains?  Here's the list:

- My #1 concern, you ask?  Fire.  The fact is, I'm dealing with a dehumidifier that I've modified into a state that it was never intended to be in, I'm controlling mains voltage with a microprocessor, and I'm pulling a heavy current load through the Arduino.  I'm really paranoid about having a massive fuel source (the boards) right next to a highly 'modified' electronics stack.  I'm thinking I'm going to put a smoke alarm in the kiln itself.  Then, at least I'll have a few seconds notice before my house goes up in flames....

- The code isn't mature.  I've done coding in various languages, but C is not my language of choice.  I'm poking around in forums, on the Arduino.cc site, in some old books I have, and anywhere else I can get my hands on, and the code is slowly getting there.  But it still needs work.  Here are the tasks I need to implement:

  * Integrate the RTC.  I currently have a DS1302 clock running, and I'd really like to use that for timing, rather than the millis() functionality.  

  * SD card data logging.  I've managed to get the kiln on Twitter, which is obviously a function of my success in getting a web connection.  However, I'd really like to do data logging onto the micro SD card that is on the Ethernet board.  I've read all over about how this is possible, I just haven't done it yet.  The thing is, the Arduino environment is really hard to do troubleshooting on.  Ideally, you would change one thing and then make sure it ran.  However, the one change takes a while to implement, as the compilation and upload can take 30 seconds or so.  My tendency is to make many changes, and then see how they work.  When they don't work as planned (a virtual inevitability), I'm not sure which change caused the issue.  I'll get through it, I'm sure.

  * Web server with SD card data.  Once the data is logged into a datafile, I'd like to be able to hit the Ethernet card with a client request and have it dump out the file.  Again, this doesn't appear to be hard to do, but I need to figure it out, and it'll be one slow step at a time.

  * Transition to MEGA.  I was complaining to a friend about the limited programming space on my Duemilanove, and he asked what other options there were.  I mentioned that I have a MEGA 2560 that has many times the space, but that would require me to remap any pins, etc.  He told me to man up and do it.  He's right.  First, I realize that there are likely many, many things that I'm doing in my code that could be optimized; however, in my case, I'm not looking for optimized code.  I'm looking for dry wood.  Long term, maybe I'll work to refine the code, but for now, it's all about velocity.  Anyway, the transition to a MEGA gives me a LOT more flexibility, especially to implement a method to directly measure the water content of the wood.  I'll be working it this weekend.