Personal Notes
A slingshot effect is predicted after corporate intranet development is essentially completed, to allow for the demand companies will have for sharing information between their intranets
Archie, gopher and Directory & Indexes will become more important
Encryption is very important for electronic commerce to develop
Snazzier features: IP telephone, RealAudio, AlphaWorld
Further research is estimating that internet will account for $327 Billion in 2002, or ~ 1% Gross World Product
Business is $8 Billion this year
Dell computer did $6 Million/day in computer sales durint Oct. & Nov. of 1997
GENERAL OBSERVATION:
Internet radio changes the natural previous audience structure; Currently 3000 stations on the net
Predict a mutual reinforcement among media
Jazzy Bioelectronic research
Development cycle:
Commerical <------ Privitization
| /\
| |
| |
\/ |
R & D --------> Partnership
Somes highlights illustrating internet-based collaboration
Major trend : Information --> Collaboration
Intangible value:
UM Distributed Library -- henry.ugli.umich.edu
to
"Computerists are trying to make something happen"
Computing Research For....?
Making computers do better the things they can't do at all or are rather
bad at.
35 years ago sorting was one example.
30 years ago public key encryption was.
Current areas are image retrieval, handwritting recognition, stenanography.
Theres a temptation to say: The problem is what the algorithm solves.
It's currently difficult to retrieve images based on description.
Trying to get computers to do something we're not good at is a slightly
fraught art.
Pricipally improvements to how one writes programs and such; make it
quicker to get them "right" enough
Showing that this program does that algorithm
Code that will work everywhere.
View this as an example of the "Succesive relaxation of assumptions"
The more the diversity of environment, the greater the requirements to be specific and inclusive.
Computers in large numbers
Coping with unreliability; Reasoning with uncertainty
Application of statistical-physics methods
Examination of statements such as "replication implies/equals reliability."
Not actually true -- it's related to the reason why processes die.
"Galactic Control": even if you could monitor the system, it
would be difficult to code this.
Also the problem of using limited local knowledge.
Large Deviation Theory is applicable to problems - buffer overflows and
queue failures.
May deal with such large numbers of computers that not everything will
work at once.
Also the question of how much can you do by measurement and inference.
Need to observe for How Long? to predict future system behavior.
Every analysis of networks has used statistical inference, techniques NOT representative of actual conditions.
A rather different sort of research.
Can't predict what people are going to find helpful.
We have to try things out and see.
Of the four, this sort gets away away from algorithms and mathematics.
Much of our present image of applications comes form Xerox PARC, and
is the outcome of alot of expensive research. Many tried, few chosen.
But mainly for COMPUTER people!
We're still "mining the research of the 70's "
And its done BY computer people to find a good environment to help them
do what they spend their time doing!
A PROBLEM: Not so easy when users aren't like the experimenters.
A methodological issue: Easier to imagine what's better for the user than
actually finding out
Leaves us with a huge nuber of options - navigating the choices can
render end-result ineffective.
Idea of tailoring it to the user's needs can become completely lost.
Reaching out beyond the local file server is the idea of a web browser. What will it be 10 years from now?
Ownership of intellectural ideas, etc. has been property of journals, now that is being questioned in academic circles.
Belief that digitizing ancient text will reduce interest in original is false - it's actually advertizing. Digitization increases demand for the original.
Future publishing evenues may stem from areas such as document spulliers rather than book suppliers.
Need to examine the property of digital coherence between separate digital media entities.
Collaborating will be about giving people more authentic experiences. There's the property of distance-independence. Potential for creating economies of scale. Higher product values which may create better learning environment than the classroom.
We're producing a society where the demand for specialized knowledge is increasing.
Significant shift from the basic laws of the universe to properties of complex systems.
Focus in academic community is sht to working on complex problems of all kinds and naturally interdisciplinary problems. The U is only one player among many.
Universities will shrink their core staff, but expand links with expert community - not on staff but associative
Transformation from culture of science to culture of research.
Using the economic diagram below:
Another way to increase the bottom line is given by ceasing investment. But net outflow cycles will eventually plummet.
How to get the far horizon into the bottom line
Simulations - network for peple to explore scenarios. Depends on ability to model and alot of people to do it.
Notion of options - the idea of making a bet on the future and paying for that bet. Institute lottery in terms of making bets on the future of companies, if payback is related to how the bet is.
Internet marks the end of Cartesion domination in our thinking - Connectionist replacing deductionist.
How the SAM parts are orginized into control networks is salient feature of different organizations, especially in biology.
How networked computation occurs <--> cells
Examine how networks evolve <-> relationship to biology
Autonomous agents' introduction into internet - observe the evolution here.
Important to understand logical organization rather than substantive deductionist parts.
Collaboration has been discussed here as a transient shift of intellectual resources rather than collaboration around a large fixed organization.
Leads to index of competencies to work in groups vs. individual creativity.
Conversion of tacit to explicit knowledge will be an active area in the future.
(Only a portion of discussion noted here, due to rescheduling.)
Disequilibirum economics important for consideing entrepreneurship on the internet.
In the Auto Industry
Where does the economy of scale come from? Optimum mnanufacturing plant size is ~200 000 workers; implies 350 total, which is less than what the world has.
So economy of scale is not in the plants, it is in product development.
What an engineer does these days is very different than previously.
Clue to the value added is the way you do the development.
It's moving very much to fundamental analyssis of material prperties, properties of combustions, sophisticated simulation of systms and processes for manufacturing vehicles.
Full crash test takes 6 - 8 hours on a big Cray computer.
Extremely many options. More reliable that performing actual crash. Result is larger diversity of products available today.
Shift in type of person who does engineering: 1/3 of engineers have Masters Degress. PhDs have doubled in 5 years.
Design-time has dropped by a factor of 3 - results from design being doen in a distributed way.
New barriers to entry that can be protected -- watch the people who come up with these ideas.
Entrepreneurs who can ride the wave of customers essentially writing their own RFP (request for proposal) will drive the new market.
Consumers will become co-creators withe the vendors. Those vendors who learn how to become "invited" will be the ones to watch.
The 1/2-life of success in a fast-moving transitory world is short. Learn to fail gracefully and how to move off positions of seeming power.
Entrepreneurial activities arose of of 3 worlds:
1) Material life
- anything is better than being a peasant
2) Market Economy
- market exchange rungs on credit, which drives the creation of capital
Examine the entrepreneur who bings technology to market - there exists a fundamental dichotomy between them and the established dominant capitalism.
3) Venture Capital
- Until the last generation - R& D has been funded by the government,
amateur investment angels, promotors and con men on the fringes
National Venture Capital Association has noted a 25-fold increase in money raised since its founding.
Information availability in a large company leads to breakdown of heirarchical authoriyt sturucture.
Intranet is the primary way af breaking down communication barriers in mature companies.