autocoding and the new MS book regarding software factories
timrueAT at mindspringDOT.com
Sun Aug 15 22:19:03 CEST 2004
Roman Suzi wrote:
> On Sun, 15 Aug 2004, threeseas wrote:
>>(CLI #1) + (GUI) + (IPC) = the three primary .....
> Well, there is no scientific foundation and in fact science tells us that
> autocoding is not possible for arbitrary tasks. I believe, it is possible to
> create autocoding (automatically generated) software for any little enough
> problem domain, maybe, say, for writing GUI. However there is theoretical
> abyss between modern computers and humans. It is called formalisation.
> Only programmer's brain is capable to cross this abyss formalising any
> non-trivial novel task.
What science? Certainly not the science of abstraction physics.
You seemed to snip what you didn't want to consider .... sorta like what
those who don't want to see my presence on a python IRC channel, ban
me.... But denial is no proof against something, but only proof of the
choice of ignorance by those in denial.
from another direction -- http://www.codegeneration.net/generators.php
Software development is still quite young. Like mathmatics via the roman
numeral system, very limited. But then came along the hindu-arabic
decimal system with its "nothing can have value" zero place holder that
cause every thing to change, including the droping of status for expert
accountants in roman numeral mathmatics.
The science of roman numeral mathmatics tell us that the complexity of
calculation required to create a computer (as we call it today and in
hindsight - for it was beyond even imagination at that point) is simply
not possible to achieve.... go figure..
But if you are looking for a singularity or magic formula or holy grail,
forget it, it simply doesn't exist. But what does exist is:
Programming is the act of automating complexity, typically made up of
simpler parts (in comparison to the sum product of the automation) and
done so in order to make the use and re-use of the automation easier for
the user(s) of the automation. This is a recursive act, from automaing
some assembly language code into a higher level language function all
the way up to a typical user scripting (perhaps via some user action
recorder/playback application)so as to make some task of theirs easier
Its clear to me that Artificial Intelligence is.. well nothing is
naturally that stupid... an illusion of simply automating enough,
perhaps in a highly dynamic manner, to present the illusion of
intelligence. Its simply a matter of the foundation nature of the
technology... binary transistors... like a two deminsional creature
being totally incapable of comprehending a three deminsional being or
Soooo..... to use an analogy we simply need to put away our roman
numeral mentality and its inability to comprehend that nothing can have
value. To drop the psuedo science of programming and get to reality with
the understanding of abstraction physics.
> And for trivial task we already have an interface:
> $ grep -i word file | sort | uniq | wc
> Saying the above, I must admit that more tasks become trivial as new
> instruments arrive. So, programming users (CP4A anybody?) can
> solve their 90% of their tasks by almost mechanical combination
> of existing tools.
> I imaging Year 2015:
> - by your command
> GvR: (standing at his computer)
> - computer, make Python 3.0 release. The grammar I sketched
> on the paper in the scanner. Please guess semantics.
> If in doubt post to c.l.p a question. If c.l.p in doubt make a
> poll. If in doubt of poll results make another poll on how to
> interpret poll results.
> - YES, SIR!
> - Good.
> (after half an our)
> - I need your help
> - yes?
> - please affirm you want to pay $10k for the Microsoft license
> and also Python 3 violates 20 patents.
> - Oh no!... (calling sponsors)
> (after half an hour)
> - Python 3 ready. Total cost $21k.
> (typing "python3" at his terminal)
> - Oh... But this is #D...
> - under given constraints this is the best solution. C.l.p poll
> shows 52% of people support this as it is familiar to them.
> (upset, playing with his time machine controls)
> (1 minute later)
> - there is another option
> - What?
> - freeware aka OSS
> Sincerely yours, Roman Suzi
Amusing.... Like writing a Buck Rodgers script as how the space program
will look like, huh?
I noticed you put alot of faith in the patent sysytem to be able to
maintain its command over father physics and mother nature, or at least
to convince humans it has such power. Perhaps someone should simply just
write out a patent application for anti-gravity and the patent office
grant it so we can all finally just have it..... huh?
But here is what is really going to happen:
you walk off a cliff and gravity will make you fall down. Likewise, you
try and put false constraints on what is naturally possible and its a
sure thing that father physics and mother nature will expose you for
your falseness, making you fall down (metaphorically speaking)
That is what is going to happen to the patent system in regards to
software. Especially when abstraction physics are established and found
to be natural and with a physical element that cannot be denied (patents
give the right to exclude) anyone. The absolute unpatentable nature of
abstraction physics and commonality of solutions via the then obvious
non-novel abstraction creation and manipulation mechanics.
from one side --- autocoding, from the other side --- code generators...
Those in the middle trying to hold on to what they want to believe is
non-automatable are going to get ....... exposed... perhaps die of
The python programming language itself has made quite a bit that was
previously more complex to impliment, easier to use and reuse, via
automation of that complexity into modules, functions, classes, etc...
I have also found the eric3 IDE exactly what I was looking for in having
the ability to trace or follow thru the execution of python code and
variables it includes.
The only drawback I find with python is its required code indentation,
but that might be turned into an advantage in developing or configuring
an autocoding tool set for it. As it is perhaps easy to follow where an
autocoder shifted inward or outward in code depth/detail.
At any rate, the fall of the patent system in regards to software and
the establishment of genuines abstraction physics.... Of course OSS is
the way to go. its the ship that is truely capable of sailing around the
world and proving that the proprietary tech of the flat earth simply is
just a denial of what is, for only the benefit of a few. Together
openly, we can do a great deal more, and that is what IS. (tunnel vision
greed or wide scope benefits?)
maybe I've just given MS a few hints..??? shrug...(a matter of denial ... ;)
More information about the Python-list