Tuesday, January 25, 2005

TAG and the world

Somewhat off my normal topic, I recently attended a Technology Association of Georgia event (Georgia Technology Summit – in which my current company made top 10 Georgia innovative companies, and my previous company was in the top 40) – but I digress – anyway, hear Jeffrey Rosensweig (Associate Dean for Corporate Relations at the Goizueta Business School of Emory University) speak. His site is www.globalguru.com and he has a number of interesting presentations available for download.

He has an interesting global perspective on things and does a lot of demographics, aging, fertility, etc. analysis factored against global trend in dollar strength, gdp (national and global), trade deficits, and many others, which shed some light on a couple things for me. First, based on his projections, India will surpass China as the #1 population around 2030 – due to the fertility rate, population statistics, and the fact that there are constraints on China’s population growth. Second was how dark (http://www.globalguru.com/request.php?21) the world is in general and where the population centers are. Makes you feel rather puny. A very good speaker and somebody who would be a wonderful professor to have…

Also got to witness Ted Turner (as he received the Captain of Innovation award). It was inspiring to hear him talk about the early days of CNN and how he persevered through many challenges. Beyond what he has already done, he has now taken on several global causes:

1) Environment and our ability to sustain our world

2) Alternative fuel (related to #1, but other cause/effects as well)

3) Nuclear armament destruction – why do we need these weapons?

There were some other items, but these were the key things that came out for me.

Kipp

Wednesday, January 05, 2005

What's on a CTO's top book shelf?

Thanks to my friend and colleague, Mitchell Friedman, I have what I consider a top shelf top shelf at work.  The list of books includes:

 

  • A. K. Dewdney, The Turing Omnibus
  • Terrence W. Pratt & Marving V. Zelkowitz, Programming Languages – Design and Implementation (Third Edition)
  • Edward A. Feigenbaum & Julian Feldman, Computers and Thought
  • Alec Sharp & Patrick McDermott, Workflow Modeling
  • Wil van der Aalst & Kees van Hee, Workflow Management – Models, Methods, and Systems
  • Christopher Alexander, The Timeless Way of Building
  • Christopher Alexander, et al., The Oregon Experiment
  • Christopher Alexander, et al., A Pattern Language
  • Harvard Business Review, Managing the Value Chain
  • Tom Mochal & Jeff Mochal, Lessons in Project Management
  • Claus Heinrich with Bob Betts, Adapt or Die – Transforming your Supply Chain into an Adaptive Business Network
  • DiBona, Ockman & Stone, Open Sources Voices from the Open Source Revolution
  • Eric S. Raymond, The Cathedral & The Bazaar
  • Frank Rose, West of Eden – the end of innocence at Apple Computer
  • David A. Schmaltz – The Blind Men and the Elephant
  • Peter Schwartz, The Art of the Long View
  • Eric Raymond, The New Hacker’s Dictionary
  • Douglas R. Hofstadter & Daniel C. Dennett, The Mind’s I
  • Douglas R. Hofstadter, Metamagical Themas: Questing for Essence of Mind and Pattern
  • Douglas R. Hofstadter, Godel, Escher, Bach (actually, this is on my shelf by my bed…but it would look good here!)
  • Christiane Fellbaum, WordNet – An Electronic Lexical Database
  • Philip N. Johnson-Laird, The Computer and the Mind – An Introduction to Cognitive Science
  • Douglas E. Comer & David L Stevens, Internetworking with TCP/IP – Volumes I, II, and III
  • Edward Tufte, The Visual Display of Quantitative Information
  • Edward Tufte, Envisioning Information
  • Edward Tufte, Visual Explanations

 

My second shelf is not nearly as lively, consisting primarily of linux, java, algorithm and compiler books…

 

Feel free to suggest additional books that should be on my list…I may have them elsewhere!

 

Kipp

 

Tuesday, January 04, 2005

Aspect Oriented Information?

My last post was on languages for handling information.  As I considered further, perhaps we can lean on some work in the computer languages with respect to aspect oriented programming…

 

AOP provides facilities for what are called crosscutting concerns – these are concerns that cross more than one ‘module’.  Things such as security, transaction integrity, and logging generally fall into this camp.  Are there similar concerns in information management that can be extracted from the information itself and handled as aspects of the information?

 

Perhaps…some examples might be:

 

-          access control

-          creation time

-          author

-          reliability

-          recency

 

As such, the information can be treated as a stream and the ‘aspects’ can be injected/tested/verified/etc. as an aspect of the information rather than muddying up the information stream itself.  Would this help?  I think in some cases it would…as part of my query, I offer up my ‘aspect parameters’ as well as my ‘information parameters’.  Or, this could be done as a second level filter on the returned information (except for the access control part of the query).

 

To some extent, existing search engines do this on our behalf – they sort by recency, relevancy, etc. and some base their results on your own personal track record.  Can I have more control over this?  Would it make sense?  How would you implement it, or does this truly reside in the information gatherer/disseminators such as Google?  Can the information providers/producers provide better indications of the information that can be leveraged by other tools downstream?

 

To this last, I believe the answer is yes – there is a lot of information lossage along the way from producer to consumer, and we can do better.

 

Kipp

 

Monday, January 03, 2005

Information Annotation

Just spent some time catching up on (well, at least made a dent) on my backlog of magazines. A Queue mag had a series of articles on computer languages. New languages they said, come about for a couple of reasons, and I paraphrase:

1) Better performance (generally order of magnitude in order to displace an old one)
2) Better abstraction/programmer efficiency
3) New application/usage (web)
4) Specialized usage (parallel)

There are also incremental improvements that have been and continue to be made on existing languages as we learn how to do things better.

Anyway, this set of articles got me to thinking about the language of information and how we handle, manipulate and specify information. I'm not really familiar with this area, but what are the available languages for information specification? I can think of several, loosely:

1) Databases - in these the meta information is stored implicitly and explicitly in the database system itself
2) HTML - information embedded in structure, some implicit information can be retrieved
3) SGML/XML - more information stored in the markup including some semantics
4) RDF et al. - additional relationships and semantic information included with the information itself

I'm sure I'm missing a lot so I'll keep digging. Nonetheless, these all seem woefully inadequate and insufficient to enable real reuse of information. There are so many attributes of the information that are perhaps orthogonal to the actual information and yet are required to truly scale and make a usable infostructure.

Is a new language of information needed? Or can we use existing systems to facilitate these needs? Things like access, ownership, authorship, timeliness, expiration, accuracy, etc. It seems that we need additional mechanisms to create and maintain this meta information beyond what currently exists.

Kipp




__________________________________
Do you Yahoo!?
Send holiday email and support a worthy cause. Do good.
http://celebrity.mail.yahoo.com