Thursday, April 1, 2010

Newton/Einstein/Doe

Dead-end framework. It would not have become a framework if it were not in some way useful. Curious: the dead-end becomes more catastrophic as a function of framework's longevity. Counter-intuitive in the sense that longevity is often associated with success.

The mechanisms we put in place to anticipate the dead-end cliff are highly speculative. One benefit of a framework is that it can be taken for granted. Many of its beneficial properties are emergent. This exacerbates the fall off the cliff, though. This ramp and fall pattern is inherent in the abstraction model of progress. Patterns are adopted often because they 'feel' right. Meaning no criticism but there are risks.

As we point out in many contexts here, naming procedural or other types of operations is key to their utility. But when a framework fails, its handle, some kind of name is already widely dispersed into everyday processes and synthesized into new processes.

A name consisting of many characters is difficult to use flexibly. It tends to require interpretation anew because of its complexity. Such a name is much easier to discard, if it is no longer useful (it describes something that has reached its dead-end) the chances it has been etched in stone is lower. After a short while, it won't be missed.

A name consisting of a few characters tends to be much easier to re-use. If what it describes has proven highly useful over a long period, the chances it is etched in long term memory is much higher. Discarding the pointer and that which it points to will correspondingly be more difficult.

Now for the instance that caused consideration of the foregoing. A large system affecting many. Management of it is a named study. Quite apart from the content of the study is a name for it that, perhaps, contains a 'color' hinting its eventual meltdown, cliff, the dead-end.

The name then is, possibly, a reflection of a fundamental flaw in the development of the field. The trick then is the diplomacy to approach someone who is likely to understand and develop something to replace it worthy of a name so that it can propagate …

The recursion did not escape him. Determinism and probability, fascinating soup mates.

No comments:

Post a Comment