SOFTWARE QUALITY IN UNBOUNDED NETWORKS

LECTURE NOTES

NOVEMBER 12, 1998

 

What is obvious is not necessarily true. When do you accept this or challenge this?

"The Internet is safe" "The Internet is secure" There is a perception of safety that is above reality. People must make independent judgements to validate the obvious.

Humans do not like change, but they like they benefits.

Unbounded networks – new context, new rules of the game. Must look to other disciplines to find techniques to solve the problem.

Rules of thumb in high tech areas – be cautious. Rapidly changing technologies change at different rates so you must make tradeoffs, but the tradeoffs change over time.

You must focus on representation before process.

Problem solving is a process that results in representation, so the representation must be right.

Absence of evidence is not evidence of absence.

People do not have all the information but you must make judgements. Problem may exist even though you have not identified them.

Is there extra terrestrial life? Current search methods based on the assumption that civilizations broadcast high-powered radio waves. Does not consider the move to low-powered wireless communication

Heisenburg uncertainty principal – so far as you are certain, you know less about reality; so far as you know about reality, the less certain you become.

In unbounded networks you must reason in abstractions. If you care about reality and have incomplete information, you must use abstrations.

RISK – measure of probability and severity of adverse effects. The field of risk is a philosophy so you cannot quantify risk.

SAFETY – the level of risk that is acceptable. This is different for each person.

RELIABILTY – does not deal with the consequences.

 

You cannot sell risk as cost v. benefits. Adverse effects can be the number of people killed.

If you graph benefits, you can choose optimal choices along the line or you can choose a set of points, so you cannot make choices in the middle.

Once you commit to a point, you are locked into it, so making a decision alters the future as well.

What is a bureaucrat? They minimize risk. The best ones do this the best. They make choices. Most of societal awards are based on how you avoid risk, not what risks you take.

So you must become a risk taker. Inovations come from risk.

Humans are adverse to change but they want the benefits from change. So there is a disparity between when a technology is invented and when it is accepted.

Notes from paper:

If your funding comes from the Department of Commerce, you must take risks. They were looking for break-through technologies. They were looking for providing interoperability without standards. Page 8 provides examples about Dr. Fisher’s ideas about the program.

Movement to a component based economy.

The Internet has changed some of the specifics, but not the goal.