Saturday, August 11, 2007

Appeal of Abstractions

Everywhere we see greater facility at reducing things to a state that allows us to think that we understand them. One result is that a lot of little things, when disparately displayed, can look fairly complicated. We've learned rules and methods to handle separate items as well as wholes, mainly using computational analogs and tools.

That bit of evolution culminated in the modern gaming and visualization regimes where we successfully deal with complicated things that may or may not have counterpart in the real world. In the former sense, think of CAD/CAE/PLM in terms of products. For the latter, we have many examples of worlds that are without any natural map.

On the product side, think of a car that has been pulled apart. How much ground area would be covered? Also, in pieces, can the car function? That's the point. After we apply the reductionist step, how do we get things back together?

Of course, that we could put the car back together, if we knew how, is one point. How does one take an arbitrary set of items and collect them into a coherent whole that is functional?

Well, the difficulty is that pieces from different things may be collected together. As one would think, this would be a new entity. This is essentially the design context. But, 'would a design work?' is always the question. A bane of modern thinking is how truth (little t) cannot be used to integrate back pieces, we need more in the sense of simulation and test.

Now, think of the issue of a new design, especially of a complicated system with severe operational demands. Our improved computational prowess has led us to place strong reliance on types of modeling that may not be as solid (pun) as then seem. Somehow, we need strategies and methods related to maintaining the proper balance.

For now, we will ignore the influence on these issues of Truth (big T); there are many examples of related problems that truth engineering can address, but let's look at some examples.

Finance - the subprime issue results from a little bit of real property being layered with abstractions that map to money and that end up integrated into something with more value, at least from a market sense. One common problem in the market is that a bubble can arise from a lot of unfounded expansions. The Federal Reserve says that we cannot identify bubbles (no foresight is the claim); that leads to the strategy of trying to lessen falls like those recently seen by adding liquidity. Yet, truth engineering can show that the concept of intrinsic value can be used with somewhat of a good foundation. One common theme has been leveraging with things that are extraneous, such as borrowing, virtual transactions (selling something that is not there), etc.

Virtual worlds (including gaming) - a whole new realm of economies and environments (example is Second Life) is being developed upon the limited bits founded upon a set of servers communicating with databases and presentation devices that are not thin. Enough progress has been made in this regard that virtual value has been mapped to money in the 'first world' (considering that the virtual world is the second - how many of these can be stacked?). These types of creations are fairly new; computational progress will only make these more confounding; we don't understand their influence on the human psyche or on the social fabric. Yet, several potential roles for truth engineering will be discussed and demonstrated.

Design -the realm of new products deals with many types of abstractions that may not relate easily to each other. For instance, planning accumulates steps that hopefully map well to goals and necessary resources. Engineering and manufacturing have to define the product to meet requirements and then successfully build within a not small set of constraints. A recurring problem can be described under the 'earned value' umbrella with which project management has struggled. Too, 'analytics as truth' is still full of open issues. As risk increases, to error on the side of empirical tests might be a preferred way to go.

The list could go on. That the 20th century brought forth our abilities to deal with 'abstract nonsense' is something that we're still trying to understand. Opportunities abound for both study and application of knowledge.


01/22/2013 -- T-issues will migrate to issues of science and religion.

08/04/2012  -- Over five years, we had a lot of side trips. We'll try to focus more. BTW, Rumsfeld has recently had his say.

09/03/2009 -- Let's face it, folks, undecidability needs to be discussed and adopted in any complex situational setting, especially if computers are involved.

07/05/2009 -- This subject is central to understanding computerism.

05/18/2009 -- With things kicking up dust again, we'll need to re-address these issues related to truth, being, computation, et al.

01/27/2009 -- Now, a new day and way (the whole sky has changed) to consider these matters.

12/12/08 -- The financial mess turned out to be worse than thought possible. The 787 problems continued to loom large.

06/01/08 -- More analysis is being done on consequences related to bad financial choices, as usual adding interesting twists to the story.

01/12/08 --- Things are getting interested, both in the Design and Finance (analysis is progressing as it ought) realms.

Modified: 01/22/2013

No comments: