Here are some of the key dichotomies or spectra I face daily negotiating the inscribed ideologies (notably plural) of the text analytic technology that I manage: Transparent vs. Black Box Open vs. Closed Evolving vs. Static User-shaped vs. Inventor-shaped Collaborative vs. Individualistic Free vs. Fee-based Open Source Data vs. Proprietary Data General Purpose vs. Sole Purpose Mixed Method (pluralist) vs. Singularly Qual or Quant (purist or positivist) Inductive vs. Deductive Idiosyncratic Outlier-oriented vs. Central Tendency-oriented Measurement-focused vs. Discovery-focused Replication-friendly vs. Replication-hostile Scalable vs. Un-scalable Machine-centric vs. Human-centric System Lock-in vs. Interoperability There are probably many more. In each case there are debates, trade-offs, and ultimately choices that require software code and a wider architecture that does not easily fall to one side or the other of a particular instance of the listed spectra. The ideology of our code is all over this map as a result. One of the things I like most about being a technology inventor is talking at length with a huge ecosystem of very diverse, multi-cultural, multi-disciplinary, and multi-national users about what it is they think they want to do with text. Very often, we struggle to say no to a user with a good idea. A really good idea is usually repeated by successive users and becomes a trigger for a fork in later stages of technology development. Usually, these are utilitarian feature requests. However, they can also be deeper, almost seismic disruptions in my thinking that answer the basic question: "What the heck is this thing really good for?" This is an incredibly social process; a melting pot of assumptions, critiques, hopes, and dreams. The deeply plural nature of the ideologies reflects conscious selection, unconscious bias, accidents, and necessities like the stark reality of resource limits. Oh the things I would build with five more engineers just like the heroic one (Mark Hoy) who has been breathing hundreds of thousands of lines of code into my long dog walk-inspired visions since the summer of 2007. Some of the more scholarly sources of all these ideas are captured in an article Chi-Jung (Lucie) Lu and I wrote about 5 years ago while I was at Pitt just starting to develop the free, open-source kernal of what is now a sprawling mass of running commercial code tapped into the world's social media, survey data, public comments & email. You can read it online at: http://www.umass.edu/qdap/IJMRA.pdf "Rigor and flexibility in computer-based qualitative research: Introducing the Coding Analysis Toolkit" Volume 2, Issue 1, June 2008 INTERNATIONAL JOURNAL OF MULTIPLE RESEARCH APPROACHES ~Stu -- Dr. Stuart W. Shulmanhttp://people.umass.edu/stu Founder and CEO, Texifterhttp://texifter.com LinkedIn: http://www.linkedin.com/pub/stuart-shulman/10/351/899 Twitter: http://twitter.com/#!/StuartWShulman Director, QDAP-UMasshttp://www.umass.edu/qdap Editor Emeritus, JITPwww.jitp.net