NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/sys/amiga/programmer/7260

Text File | 1996-08-05 | 5.3 KB | 105 lines

Path: news.NetVision.net.il!news
From: Jack <[e-mail]>
Newsgroups: comp.sys.amiga.programmer
Subject: Re: 680X0 -> PPC translator?
Date: Sat, 13 Apr 1996 10:17:25 -0700
Organization: NetVision LTD.
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]> <[e-mail]> <[e-mail]>
NNTP-Posting-Host: ts006p16.pop4a.netvision.net.il
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Mailer: Mozilla 2.01 (Win16; I)

Mathew Hendry wrote:
>
> Jack ([e-mail]) wrote:
> : Mathew Hendry wrote:
> : >
> : > Jack ([e-mail]) wrote:
> : > : Mans Engman wrote:
> : > : > Once you think a bit, it's easy to show (prove!) that static code<->data
> : > : > distinction can't be determined by an algorithm. It is what theorists call
> : > : > an "undecidable" problem. Granted, for a "real" computer with a finite amount
> : > : > of memory/indata it can be "solved" by brute-force search, but this is
> : > : > not a practical approach.
> : > :
> : > : many problems aren't solvable with today's technology, but it doesn't mean
> : > : they're not solvable with other yet-to-devised methods, now do they??
> : >
> : > Some may be solved by new algorithms, but no finite number of algorithms can
> : > solve all problems. You surely can't be suggesting that your static translator
> : > will contain an infinite number of algorithms - or one infinitely large one.
> : >
> :
> : hell no, the algorithm is definitly not infinite otherwise i wouldn't suggest it
>
> In that case, it can't work in all cases. There is an inherent indeterminacy
> between code and data which cannot be completely resolved by ANY algorithm
> which you may come up with. This can be proved using a variant of G÷del's
> theorem formulated by Alan Turing.

by all means, prove me wrong. be concrete and don't generalize one theorem on this case, please.
code and data can easily be distinguished, if a series of words makes sense as a code section it's
probably code and you can make sure it is by deferring its translation until it is called from some
place else or it simply makes too much sense to be random data. for example: valid jump/calls, references
valid addresses, code instruction sequence maintained w/o being interrupted by some data word which
doesn't make sense as code. what you claim suggests that not any algorithm can solve ALL situations
of a given problem, like how many algorithms exist for adding 2 numbers. there's obviously one.
your inability to deal with more complex problems prevents you from seeing a possible solution.
just as human can do manual translation of the code so can an algorithm.

>
> :                                  your claim is based on the fact that some thing have yet to find
> : their earthly solution, but not all secrets have been uncovered yet, the inability to solve something
> : doesn't make it impossible to solve, there's always some way or another to solve things even in the
> : most indirect and mysterious ways, you can't just claim they're impossible to solve just because you
> : still hav'nt found any workable solution.
>
> Sorry, no matter how many "indirect and mysterious ways" you invent, the
> problem cannot be solved algorithmically for all cases. That is a mathematical
> truth.

oh is that so, you speak as if ALL of the math secrets are discovered to you. well you DON'T know
that for sure. have you tried ALL possible solutions to solve one of your undecidable problems.
i'm not eliminating the impossible, oh no, some things are definitly impossible even to imagine.
but you simply have no CONCRETE basis for your claim. as all the others have already submitted their
ideas of impossible getaway situations which have successfully been dealt with by yours truely,
i suggest you do the same to undermine my theory.

>
> : > In any case, we ARE talking about today's technology...
> :
> : i'm also talking about today's technology.
>
> Then you can't win.
>
> : > By the way, what happens if your static translator tries to translate a
> : > program containing bugs? Presumably your algorithm is sophisticated enough to
> : > recognise bugs when it sees them? God help us if you throw AMosaic at it...
> : >
> :
> : it's not trying to solve bugs, it'll translate a buggy program any day.
>
> Great, try translating a program which in some circumstances attempts to
> execute some of its own data. You immediately have a problem - is that portion
> of the program code or is it data? You cannot decide, for sometimes it appears
> to be data, and sometimes it appears to be code. This indeterminacy remains
> even if the program contains no bugs, and remember, nearly all useful programs
> DO contain bugs.

read above paragraphs for hints of solving this simple situation.

>
> You may choose to work around such problems by storing indeterminate program
> fragments in their original form in the translated binary. But these residual
> untranslated pieces will have to be handled on the fly when running the
> translated program. In that case, you no longer have a static translator -
> you have an emulator.
>

no no no, no further run-time analysis is necessary. only static translation.

Avi Lev.