Reply to CCogan: Waste and computer evolution

From: DNAunion@aol.com
Date: Mon Oct 02 2000 - 00:22:51 EDT

  • Next message: DNAunion@aol.com: "Re: Posting from other resources"

    >Ccogan: The evolutionary algorithm is *hugely* wasteful in design terms,
    precisely because it doesn't know what it's doing. And it doesn't know
    because it's not intelligent.

    > DNAunion: Waste itself is not a sign of lack of intelligence. Industries
    intentionally create huge amounts of waste everyday - are they not
    controlled by intelligent beings?

    >Ccogan: Thios sounds like equivocation of the word "waste".

    DNAunion: Okay, if you don't like "waste" meaning something along the lines
    of "surplus of something generally useful" (which is a common anti-ID
    argument - for example, "what about all the species that have gone extinct,
    or all those viable plant spores that don't participate in fertilization and
    just go to waste?"), then we can deal with "waste" meaning something along
    the lines of "surplus variants that serve no useful function".

    What about computer programmers, which are surely intelligently-designed:
    there is much "waste" in their "evolution". Programmers first code (we won't
    get into the conceptualization steps that proceed the first line of code
    being written) then attempt to compile their program. Any but the most
    trivial programs written from scratch will have compile-time errors that must
    be corrected, thus the programmer has to modify his/her code, then resubmit
    it to the compiler. This can take several iterations before a clean compile
    is achieved, resulting in waste (code statements that are useless and
    eliminated). After a clean compile, the program is run and any runtime
    errors (such as invalid paths or files that don't exist) must also be
    handled, which usually entails more modifications to the original code,
    resulting in more waste (elimination of, or modification to, code that served
    no useful function as originally written). Then, once a clean compile is
    obtained and no runtime errors are encountered, the system output must be
    validated since logic errors in the code would generate results, even if
    those results were invalid. Again, since most non-trivial programs written
    from scratch are bound to have logic errors (which are sometimes the hardest
    errors to track down and fix), then some more modifications must be made to
    the original source code - you guessed it: more waste. And even once all of
    these steps of debugging are followed and all is okay, then still more
    modifications are typically needed as a complete program/system is not
    written at once - a prototype might have be generated that gives the end user
    an idea of what the final system may look like, but it is useless until
    actually "fleshed out". In addition, individual modules have to interact
    with each other (passing and accepting the correct number and type of
    arguments and parameters), which can cause more modifications to be made (yet
    more instances of code that did not perform its function as originally
    written - i.e., waste). Not until the full final product is in place is the
    source code complete, and up until that point, a "whole buncha" wasted code
    would have been created (and subsequently eliminated of modified).

    In fact, on top of all of this, more wasted code can found within the final
    source code in the form of internal documentation: code that the computer
    completely ignores (it is useless and can be discarded without affecting the
    system).

    >DNAunion: Could you provide the example on which you base your claims? What
    exactly evolved in the virtual world by the evolutionary algorithm and where
    can one gain access to the guts of this algorithm?

    >Ccogan: http://www-lecb.ncifcrf.gov/~toms/paper/ev/

    DNAunion: I am familiar with this work and wrote up a short "problems with"
    post. But it is at ARN (note that the board rules here do not say it is
    "wrong" for a person to quote his/her OWN material from another site) and I
    cannot find it using the Search feature there (which I griped about while
    posting there) and the access to the actual board to do a post-by-post search
    is not possible. Since I don't feel like re-inventing the wheel, I will hold
    off on replying to this computer model until such time as I can once again
    gain access to my previously-posted material.

    >Ccogan: There is a good example of virtual silicon.
    http://www.newscientist.com/ns/971115/features.html

    DNAunion: Yes, I am aware of this experiment also. This one was directed
    evolution. A predetermined goal was setup before the process began, then
    each selection step referenced the final goal to determine which was the most
    fit. This "Evolution" was an example of a process governed by external
    teleology, something that evolution is nature does not possess.

    >Ccogan: … and thus does not need God or any other alleged non-natural

    >DNAunion: "Not needing something" is not the same as "something did not
    happen".

    >Ccogan: True but as Behe clearly stated we cannot conclude or infer design
    if a natural pathway could have existed.

     "We must also consuder the role of the laws of nature. The laws of nature
    can organize matter --for example, water flow can build up silt sufficiently
    to dam a portion of a river, forcing it to change its course. The most
    relevant laws are those of biological reproduction, mutation and natural
    selection. If a biological structure can be
     explained in terms of those natural laws, then we cannot conclude that it
    was designed." pp 203 Darwins Black Box

    DNAunion: Yes, but as Behe also clearly points out, not inferring design
    because the indicators of design are not strong enough might mean that we
    are overlooking the truth. Both he and Dembski state that their methods of
    detecting design can MISS design (but that the methods will not misclassify a
    non-designed object/event as design).

    >DNAunion: For example, let us hypothesize that 50 years from now, both
    intelligence and nature are shown to be capable of producing life from
    non-life. At that hypothetical time, if IDists and Creationists stated …

    >Ccogan: There is no evidence of intelligence to be able to produce life
    from non-life.

    DNAunion: I didn't say intelligence could produce life. Why not reread my
    statements to which you were replying, paying special attention to this part,
    "let us hypothesize that 50 years from now…".

    Second, intelligently-directed prebiotic experiments have far more success
    than hand-off, synthesis-in-the-whole, undirected prebiotic experiments ever
    have. If the current trend holds out (which logic would surely indicate)
    then an intelligently-directed method to create life from non-life will be
    found before a purely-natural method (if a purely-natural method is ever
    found).

    >Ccogan: Certainly there is no evidence of intelligence producing IC
    systems in biology for instance.

    DNAunion: But there is evidence of intelligent design creating biological
    things that nature never did - just look at genetic engineering and rational
    engineering in protein design.

    >DNAunion: [if in the future it was found that] "nature is not NEEDED to
    produce life", would you accept that life was designed? I seriously doubt it.

    >Ccogan: Design cannot exclude natural selection as the designer.

    DNAunion: And naturalists cannot exclude ETI's as the sufficient cause of
    life's appearance on Earth?

    >DNAunion: So why would you expect IDists and Creationists to accept now
    that life evolved purely naturally because intelligence is not "needed"?
    (And furthermore, I believe you still have not shown your basic premise to be
    true).

    >Ccogan: The problem is that ID'ers are trying to infer design through
    elimination.

    DNAunion: And the other side of the coin is that "naturalists" are trying to
    infer purely-natural causes for the origin of life by default: without
    evidence.



    This archive was generated by hypermail 2b29 : Mon Oct 02 2000 - 00:23:03 EDT