Ohio IT Skills Podcast

Podcast-Patents and Open Source

Siddal on Open Source; Hollaar on Patent Proposal - Download the February 12, 2006 podcast. (Run time: 33:30)

This week's podcast addresses two key ways to develop and protect as well as disseminate new technologies.  The balance between dissemination and protection is one of the key issues in motivating inventors to create new technologies, and in getting new products and techniques into the hands of early adopters.

Dr. Scott Siddall, Assistant Provost of Denison University and CEO of the Longsight Group, discusses the nature of open source development - and the new business opportunities that are springing up around it.  All this in light of increasing acceptance in the world of higher education - and in the context of OLN's OSPILOT Project, intended to provide access to open source products for use by OLN members.  (Interview about 11 minutes)

Dr. Scott Siddall
Denison University, Granville OH

OLN's Open Source Pilot Program (OSPILOT)
The Longsight Group
Open Source Initiative
The Cathedral and the Bazaar by Eric Raymond

Then, we talk with Dr. Lee Hollaar, professor at the University of Utah and an expert in the US patent system, and a student of new approaches intended to improve the system and the flow of technology into open use.  He has worked with the US Congress on developing new legislation to enhance the operation of the US Patent and Trademark Office (USPTO). 

In the Feb issue of IEEE Spectrum, Dr. Hollaar published a "modest proposal" to establish a new class of patent that might smooth and speed up the process of obtaining protection - and hereby making inventors more comfortable with publishing details of their innovations, so that they and others can continue the process of incremental improvement and enhancement.  (Interview about 13 minutes)

Dr. Lee Hollaar, University of Utah, Salt Lake City UT
Digital Law site
Patents 2.0 (article in IEEE's Spectrum, Feb 2006)

(-- originally posted by Rich Bowers, Coordinator, Ohio IT Clearinghouse)

Podcast-Pattern Recognition, Dr. Martinez OSU

Download this week's podcast - An interview with Prof. Aleix Martinez of OSU about pattern matching techniques in large collections of data Run time: 23:30

Pattern recognition is a major issue in modern computing.  As more sensitive and higher volume data collection techniques accumulate more and more raw data - the problem of sifting true knowledge from it looms ever larger.  Professor Aleix Martinez and his team at the Computational Biology and Cognitive Science Lab of the Ohio State University has been working on the difficult underlying problems for pattern recognition.  For example - how does one know which data represent the "pattern" that leads to new information?  In a sense, Dr. Martinez and his group are focused first on the problem of "pattern detection" - and them with the problem of finding the right technique for recognizing that pattern within a specific data set.

Also, the work has some fascinating implications for researchers studying American Sign Language (ASL), used by and for the hearing impaired.  Turns out that facial expression is an integral part of ASL, and little is known about what contribution it makes.  Efforts to automate translation of ASL video to and from standard English are stymied until we have more understanding of that contribution.  Martinez' work, as it uses facial image data to resolve pattern matching techniques, may make some substantive advances to our understanding of ASL and how it really works.

-- Article: "Where are Linear Feature Extraction Methods Applicable" Aleix M. Martinez and Manil Zhu; IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol 27 N0 12, December 2005

(-- originally posted by Rich Bowers, Coordinator, Ohio IT Clearinghouse)