Sunday, December 30, 2007

Extrapolation, calibration, truth

Context: See Tru'eng anewfocus going forwardmathematics.


Some comments (in earlier posts) were motivated by recent events in regard to our growing use of techniques based upon abstractions thanks to the ubiquity of computers and web servers.

One particular set of events dealt with frothing resulting from tranche-ing, as it's so easy to see this bubbly phenomenon in the financial areas that get divorced from proper evaluation methods (where is the science behind the grand intellectualisms?); but, one might think the problem exists too in the hard area of program management, where expectations and time are seemingly inversely related (one classic example is software engineering - at least, a plane can finally get to the air or not). Aside: remember that both expectations and time are heavily subjective despite all the apparatuses devoted to controlling the latter.

But, to be fair, during the whole of the 20th century, our observational selves have pushed out touch and measure far from our bodies and its senses. By the end of last century, there was very little that was not computationally supported. Did anyone in 1990 use slide rulers [sic] like those in 1960 (see Remarks), for instance? Metrology became a focal (pun) science. There is some type of artificial assist found in about any modern situation.

Even many metal workers, in 1990, sat back (figuratively speaking) and watched a machine work. Even though the semantics of the controlling devices were seemingly simple, the mathematics was not. Who would have thought that lowly geometry would now be the major subject that it has become with all the variants that have spawned (essentially topological movements)?

This subject will take a lot more attention. But, to limit the view for the moment, we could bring up the issues related to modeling such as we will see with simulation, virtual worlds, mixed-media situations (in terms of robotics), and the like. This is not a static set of things.

Any purely virtual event would not (does not, ..., remember the importance of handling quasi-empiricism) map well to anything corresponding to the world. Just look at the differences in geometry handling between the visually-based gaming world and the metrology-based world of CAx. There are situations where servos (which are and are driven by computational events) work very well (like stopping your car reliably). There are so many other examples as these types of things abound, again, as a non-static set.

The augmented reality work makes one think that real-world events and the computational are going to be more closely tied as an inevitable part of progress. The questions remain, though, about quasi-empirical issues and trust (the domain of truth engineering).

One thing to look at will be how we 'calibrate' between the world and the computer model, knowing full well that measurement itself will be model (and, for the most part, computationally) based. You see, even with a good calibration between the world and the computer, there will be computational steps that fill in (interpolate) or expand (extrapolate) thereby adding potential error due to several factors that we'll look at.


01/05/2015 -- Tru'eng, anew. This post is the most popular, of late, which raises some interesting questions. Eight years, ago? That was even before the Made-off revelations.

04/19/2011 -- We have to get back to the basics.

01/01/2011 -- We have four last posts of December under our belt.

11/21/2010 -- Three years ago, it was said: Computational foci raise miraculous need. Still applies.

08/20/2009 -- We'll switch blogs to look at technical economic issues.

12/18/2008 -- Well, things really fell apart in the 3rd Quarter of 2008. Of course, the tranche was only one factor. Others include the players and the games. Now, games include using mathematics erroneously, as in getting an aura from the use of derivatives (to be discussed further). We'll have to re-address the map/territory issue.

01/18/2008 -- Of course, 'slide rule' was meant and would have been appropriate with some changes to the phrasing.

'slide rulers' was a Jungian contraction (not dissimilar to the Freudian slip) where an event in the non-parallel (non-linear) collective unconsciousness collapsed out a concept into a little tidbit.

The topic was how the 20th century culminated in a whole slew of artificial media interposing themselves between us and reality (many of these measuring - rulers, so to speak). Of course, at the same time, the 20th-century thinkers pushed forth on us the idea that we cannot know reality anyway in the sense of having a dog's nose.

As well, it was not only sensing that was entrapped with our devices, thinking has been profoundly affected by the ubiquity of the computational (hence, truth engineering's necessity - we'll be looking at sufficiency issues as we go along) leading to several classes.
  • Simple user - not essentially ignorant, as anyone who used mathematics founded on computation runs into issues that are largely unresolved - some statisticians are prime examples.
  • Wizard - having access to the internals sufficient to effect behavior controls, seemingly doing so by magic, in many cases.
  • Magician - above the Wizard as a grasper of underlying theory and of the larger issues, yet subject to the same human failings.
  • Innocents - who can go around and not have to worry about this stuff.
So, who now uses a slide rule (not for measuring, but for calculating) which was ubiquitous in certain circles in 1960? Who can take any of our devices and from first principles recreate an analog? A friend of mine jokes that modern EEs are stock keepers (obtaining off the shelf circuity - not generating from scratch) rather than creative designers.

Of course, there are arguments for not re-inventing the wheel, as we seem to do perpetually.

Modified: 01/05/2015

No comments: