Kickstart my PCR!
My pals Tito Jankowski and Josh Perfetto have been working for the last, oh, nine months or so on designs for an Open Hardware thermocycler -- basically a Xerox machine for DNA. They've finished their first working prototype, and have set up a Kickstarter project to fund the process of turning this into a full working device that you'll be able to buy for less than $400, or build all by yourself with parts you can easily obtain online. If they can get to $6000, this will happen.
I'm particularly interested in this because its software will be the second real-world demonstration of some of my theoretical work. Some of you might remember Dejector, the "kills SQL injection dead" library I built back in 2005 (and have been really slack about keeping current, though it really needs a serious rearchitecturing). Dejector uses a technique I call "restricted sublanguages" to make sure that SQL queries which don't fit into a very limited (programmer-specified) subset of all possible queries -- that is to say, queries which have had a malicious clause injected into them -- are rejected before they get near the database. The OpenPCR machine is a networked device; you'll be able to plug it into your router and configure a PCR run via a webpage, rather than having to key instructions in on a tiny little keypad. It'll also log data for you (which you can also view in a browser) and, if you want, report results to you over Twitter or SMS.
All this fancy web stuff will be made possible -- and secure! -- through a restricted sublanguage of HTTP which I will be implementing for the AVR series of microcontrollers. (We're actually starting with an Arduino, but we might move to pure AVR by the time we're done.) Your contribution will help go toward making that happen, along with tools for generating custom restricted HTTP sublanguages for other embedded devices. (Networked lab tools are cool; networked lab tools that get hacked to pump out Twitter-spam, not so much.)
If you can spare a few bucks, please kick something in, and please signal boost anywhere you can think of. Thanks!
I'm particularly interested in this because its software will be the second real-world demonstration of some of my theoretical work. Some of you might remember Dejector, the "kills SQL injection dead" library I built back in 2005 (and have been really slack about keeping current, though it really needs a serious rearchitecturing). Dejector uses a technique I call "restricted sublanguages" to make sure that SQL queries which don't fit into a very limited (programmer-specified) subset of all possible queries -- that is to say, queries which have had a malicious clause injected into them -- are rejected before they get near the database. The OpenPCR machine is a networked device; you'll be able to plug it into your router and configure a PCR run via a webpage, rather than having to key instructions in on a tiny little keypad. It'll also log data for you (which you can also view in a browser) and, if you want, report results to you over Twitter or SMS.
All this fancy web stuff will be made possible -- and secure! -- through a restricted sublanguage of HTTP which I will be implementing for the AVR series of microcontrollers. (We're actually starting with an Arduino, but we might move to pure AVR by the time we're done.) Your contribution will help go toward making that happen, along with tools for generating custom restricted HTTP sublanguages for other embedded devices. (Networked lab tools are cool; networked lab tools that get hacked to pump out Twitter-spam, not so much.)
If you can spare a few bucks, please kick something in, and please signal boost anywhere you can think of. Thanks!