Re: [EGN] Numerical Accuracy
From: Edward G. Nilges (spinoza1111_at_yahoo.com)
Date: 09/14/04
- Next message: Edward G. Nilges: "Re: Zell Miller nailed the liberals last night"
- Previous message: Paul Lutus: "Re: comp.programming and the Hackers thread."
- In reply to: Randy Howard: "Re: [EGN] Numerical Accuracy"
- Next in thread: Randy Howard: "Re: [EGN] Numerical Accuracy"
- Reply: Randy Howard: "Re: [EGN] Numerical Accuracy"
- Reply: EMonk: "Re: [EGN] Numerical Accuracy"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: 13 Sep 2004 22:30:49 -0700
Randy Howard <randyhoward@FOOverizonBAR.net> wrote in message news:<MPG.1baf7c445205e70989a04@news.verizon.net>...
> In article <f5dda427.0409130250.4d34545a@posting.google.com>, spinoza1111
> @yahoo.com says...
> > Unfortunately, as we have learned, during the above period I learned C
> > well enough to (1) advise John Nash on C,
>
> Your description of that event shows that you gave incorrect advice,
> which has been pointed out to you before in this forum. Google archives
> also hold that discussion.
The advice worked because in fact the Microsoft compiler had a bug. I
am not interested in "Google archives" constructed by liars.
>
> > (2) teach classes in C at Princeton and
>
> It wasn't the first (or last) time a school had an incompetent teacher.
> The knowledge that you are not doing it any longer raises my confidence
> in that institution's ability to correct flaws.
I left the employment voluntarily for a better offer in 1992.
>
> > ... the major matter of the deficiencies of C,
> > including the fact that it cannot be optimized adequately and requires
> > the use of trivial and nasty tricks, which caused me to abandon C and
> > languages in the C family for subsequent development.
>
> Strange, just a few days ago you posted that anyone who did not know
> that the C optimizer *would* work adequately for your purposes was
> stupid. Which is it? The capabilities of the optimizer are not so
> much a language issue as a quality of implementation issue on the
> part of individual compilers. The language standard itself says
> precious little about optimizers. The fact is, it is probably NP
> to attempt to write an optimizer capable of unwinding Nilgewater
> spaghetti code to such a degree as it will perform adequately.
> I suspect your own compiler can't do it either, but I don't have
> your book. Perhaps you should post a link to the CD contents so
> that someone can try YOUR OWN CODE CONVENTION on your own compiler
> and see if it corrects such things. Maybe it takes a Nilges to
> optimize a Nilges.
You obviously didn't understand my point at all because you are
speaking on matters on which you aren't qualified. I've developed an
optimizer for a compiler: have you?
At any rate, if you were qualified to speak on compilers and their
optimization,you'd know that C cannot be optimized in the way a more
deterministic language without aliasing and without uninitialized
memory can be optimized.
But you are not qualified to speak on compilers and their
optimization. Merely being some sort of coder or manager who uses
compilers doesn't qualify you. Merely being able to create
"benchmarks" using compiler optimizations doesn't qualify you.
Had you any expertise in this area, you wouldn not have used "variable
hoisting" to refer to a questionable manual practice.
>
> > > You see Nilges, I can explain it for you, but I can't comprehend it for
> > > you. Therein lies the problem.
> >
> > You have posted systematic abuse based on the fact that Richard
> > Heathfield emotionally manipulated you
>
> Wrong. I read the early threads in which you participated here, and
> concluded after the course of a couple of weeks that you are a fake,
> a troll and a wannabe, despite your claimed long years in the industry,
> from which you seem to have absorbed so little as to be beyond belief.
This is an excellent example of psychological transfer because you did
not spend any time in research, nor did you contact me by email to
either complain, or to get clarification on issues that concerned you.
Instead, because you yourself are anxious about your own competence,
you have spent the last TWO YEARS in persistently trying to rid
yourself of your anxieties with the timelessness and lack of
development that is characteristic of the psychiatric patient.
>
> It is strange, because you do seem to be at least peripherally familiar
> with a wide range of topics, most of which are quite obscure in general,
> but have no concept at all of the simpler items. It's hard to explain
Such as the absolute need for loudmouths to be right all the time?
Righto.
> how this could happen, which is why I still suspect you are trolling.
>
> > benchmarks have been known to be falsifiable
>
> Perhaps, but not when *YOU* have the code and the compiler and can
> measure them for yourself that argument is rendered moot. That is just
> hand-waving, you know the truth, or could obtain it on your own in a couple
> of minutes, but choose not to admit it. That is precisely why people don't
> put up with your crap or take you seriously. You can *claim* it has to
> do with sociology, third-party manipulation, or sunspots all you
> like, but the truth is you are reviled because you are full of your
> own ego, with absolutely zero basis in fact for such a high opinion
> of yourself. You are either playing stupid from a position of
> real success, in which case you are not doing a good job of play-
> acting, because your natural holier-than-thou is coming through, or
> you are in fact a pretender, with an ego based upon zero basis whatsoever.
> Neither option is very complementary.
If you feel you are a fraud with a loud mouth, who has successfully
terrorized subordinates and women for years, deal with it. I am myself
confident enough in my own record to have in fact an antinominian view
of "programming" as a nonprofession.
>
> > meaningless compared to a
> > discursive and literary understanding of coding in the actual real
> > world, one you don't have.
>
> Talking about programming is not more authoritative than actually
> doing it. Measuring results is more authoritative than claiming
> results that *should* occur if some sociologist you read about 40
> years ago is correct about the relationship between ping-pong balls
> and ego castration in corporate America. I can assure you that the
> level of understanding of code you have exhibited here is nil. If
> you really do understand it, you have been very good about hiding
> such knowledge in public.
In fact, my level of understanding of C exceeded yours in 1992 after I
developed a business rules compiler in C and realized that there is a
better way...than worrying about memory leaks unnecessarily. Your
vanity is that of the old-fashioned assembler programmer who could
never admit a mistake, and you are thereby part of the problem set.
>
> > But you have been incapable of dialog on the relative merits
>
> ??? I have been pretty clear about it, and have in fact posted
> code which will actually compile, and will show, for your own
> examination if you actually had the care to know if your ridiculous
> claims were accurate. Calling the sky polka-dotted does not make
> it so.
>
> > you are incapable of admitting error.
>
> Do you compose these articles while looking at yourself in the mirror?
> 99% of your descriptions of others lately have been stunningly
> autobiographical. The above is exactly what has been said about
> you hundreds of times in the last year or so. As soon as you show me
> to be wrong, I'll admit it. When corrected on items here and in other
> forums, I admit it. Your track record is quite the opposite.
>
> > Although because of my disgust with you I haven't looked at your crap
> > code I do note that you simply can't dialog.
>
> Perhaps it was too complicated for you. Others have posted even
> simpler examples, which also show the results, but you haven't
> responded to them either. Even if you don't look at the code,
> multiple people have replicated the results on at least 3 different
> operating systems. I've since also performed them on NetWare and 10-yr-
> old AT&T SVR4 UNIX with cc and ancient hardware. The results are all
> uniformly opposite to your claims. That's far and away more
> authoritative than your asinine "if you are smart enough to turn on the
> optimizer" stupidity, particularly since it has been proven false on
> some of the most widely used compilers, processors and OS platforms
> in existence. If I have the time, I'll even try it on a TI MSP430
> embedded test board and see how the compiler for it behaves, but we
> already know the truth. You just can't see it.
>
I am already aware that there are a number of misfits who have come
here and who agree with each other:
> As you may not know, engineers and scientists actually perform experiments
> to verify such things. Psychologists claim stuff, and hope nobody
> argues against them, because there is no sound evidence to support them,
> just opinion.
The problem is you are neither an engineer nor a scientist, since you
are dealing with a social reality. The mistakes Dennis Ritchie made in
designing a language in 1971 are not the subject matter of any
"science" EXCEPT psychology, anthropology and sociology. Being skilled
at workarounds does NOT make you anything except a tinkerer, neither
and engineer nor a scientist.
>
> It's plain for all to see which is closer to the model of your behavior.
>
> > You haven't to my knowledge responded at all to the consideration that
> > the program in question was using a bounded input string and was more
> > concerned with clarity than with microefficiency.
>
> I have, but perhaps you do not remember. The most recent example of
> such a discussion is implicit in my use of a relatively short string
> within the "benchmark" code, which in addition is never modified
> during execution. Even with such considerations, the compiler still
> does not achieve what you claim will be achieved. It is obvious that
> you have never tried it yourself, but rather hope that it is true,
> but not being a programmer, you don't have the inquisitive nature to
> actually want to KNOW instead of BELIEVE. This is an intrinsic
> difference between philosophy and science.
No, the explanation is simpler. I have no interest in any
collaboration, direct or indirect, with people with serious and
empirically confirmable psychological problems in anger management.
>
> Further, first year programming courses teach students to not place
> calculations within loops which can be evaluated only once before
> (or after) the loop body. As such, such code is very commonplace in
> practice, and is encountered so often as to be idiomatic in multiple
> languages.
>
> In point of fact, NOT seeing it is confusing because of its rarity, as
> well as its demonstrably poor results. As such, your usage is less
> readable amongst professionals, not more so. Professionals also KNOW
> that such usage as you recommend will not achieve decent performance,
> EVEN WITH BOUNDED INPUT STRINGS, so they will immediately see it, and
> change it if found.
>
> This is exactly what happened to you, further proof that whatever ability
> you claim to have is uninformed and lacks depth and breadth of experience,
> regardless of how many years you have been trying to learn.
>
> This is essentially the old saw, you can have 30 years of experience, or
> you can have 1 year of experience 30 times repeated. Your postings, code
> and opinions indicate strongly the latter.
Again, you want me to play the game I have indeed seen played by
programmers for thirty years, in which in the absence of professional
solidarity, each one claims for himself to have professional
qualifications, the assertion of which is made, in the context of
economic competition, in a zero sum game.
I am not going to play that game because it is a game for psychiatric
patients, a sign of whose disorder is that they keep, like you,
returning to the source of the wound and trying to heal it, by
inappropriate means.
What bugs you is that without having to "prove myself" by "writing
code", my English prose makes it clear that I have experience in
programming that is wider, deeper, and was altogether more amusing,
more fun, than your experience, which probably consisted of abusing
subordinates with your big mouth.
My lighthearted prose and my willingness to sacrifice your
misunderstood efficiency for readability were in fact the sign of the
first-rate programmer, who regards programming properly, as writing
and communication.
It bothered you also that I could write to this ng while unlike you
maintaining a perspective and a sense of humor which you, in part
owing to excessively narrow culture (so narrow as to require you, at
the cost of *amour-propre* to "win" at all costs, because you live in
a Closed World).
You therefore concluded prematurely that because all the "great"
programmers you knew were fat and ugly misfits unable themselves to
write clearly unless like you they were angry at someone, that I must
be a fraud and a fake.
For is not a "programmer" a fatso beardo with anger management
problems?
>
> > You could have cleared up the misunderstanding by using email, as I
> > attempted myself (only to be accused of harassment despite the fact
> > that the use of email is recommended in these situations).
>
> That's a damn lie. You used email to attempt to blackmail me into
> not correcting your numerous other lies and mistakes here by the threat
> of legal action. It was laughable then, it remains laughable, and you
> are still drinking your own bath water on the topic. Since you continue
> to post almost daily references to me in threads I do not even participate
> in, publicly, in this forum, I see no reason for my comments in response
> to your BS to be done privately. You are a hypocrite sir, and a right
> foul one.
>
> > Your trolling has long since exceeded the level required for a civil
> > action and in a private network, it would be considered criminal in
> > nature and in the way it abuses the facility.
>
> Once again, the autobiographical nature of these statements is clear.
> It's okay to admit you are troll EGN, everybody knows already. It's
> not a secret.
>
> > I have a right to a reputation and a good name.
>
> No, you do not. What you do have is a right to be qualified on the
> items of which you choose to speak publicly, and if that is evidenced,
> you will *EARN* such a reputation and possibly also a "good name".
> Respect is not given, it is earned. To date, you have not done so in
> this forum, to put it lightly.
Boo hoo. Fools' approval stings: their honor stains.
>
> > You have no right to deal with your isolation and rage in
> > this deviant fashion with your big mouth.
>
> I have no isolation and rage, it is you that has admitted in this
> forum that you are lonely, and that such arguments give you attention
> that you need. Further, you have admitted that any interaction is
> better than none at all, so consider this therapy. The bill's in the
> mail.
>
> As far as me not having a right to speak, that is false.
You still have a big mouth.
You really should try reading your psychotic replies out loud to your
wife. She'd have you committed.
I need you to stop trolling and stop inciting this group troll. The
very language of "troll" is the language of human sacrifice in which
people can disavow their own dark side by scapegoating a convenient
individual, but the real trolling here is a group phenomenon like a
lynch mob.
You and your confederates will otherwise be targets of legal actions
including restraining orders.
- Next message: Edward G. Nilges: "Re: Zell Miller nailed the liberals last night"
- Previous message: Paul Lutus: "Re: comp.programming and the Hackers thread."
- In reply to: Randy Howard: "Re: [EGN] Numerical Accuracy"
- Next in thread: Randy Howard: "Re: [EGN] Numerical Accuracy"
- Reply: Randy Howard: "Re: [EGN] Numerical Accuracy"
- Reply: EMonk: "Re: [EGN] Numerical Accuracy"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
