Re: Programmer's unpaid overtime.

From: Programmer Dude (Chris_at_Sonnack.com)
Date: 12/09/03 

Date: Tue, 09 Dec 2003 16:56:55 -0600

"Edward G. Nilges" wrote:

>> Why would NP behavior have anything to do with tokenizing input?
>
> HAHAHA

Ha, ha, ha, yourself. You've embarrassed yourself pretty badly
over this one. No dinner for you tonight.

> This is the reduction to absurdity of "relevance" for OF COURSE,
> nonpolynomially complete behavior has plenty to do with tokenizing
> input.

"nonpolynomially complete" ... HAHAHA!!

> A tokenizer algorithm has a formula for its efficiency.

Indeed. And it will likely be a linear one. For most tokenizers,
it *should* be O(N).

>> 2. These tests were not of your C attempt. That was so broken
>> that testing it would have humiliated it. The tests were for
>> your favored child, PowerString. *I* whipped up a couple of
>> quick versions that *all* blew the doors off your PS.
>
> oooooo I am so impressed.

The only thing that should impress you is your potential to improve,
grow and learn.

>> 3. Yours PowerString performs like a dog not because of a mere
>> flaw in the code, but because of deeply broken design decisions
>> that reflect an apparent lack of understanding of VB. For
>> example, I seem to recall a totally *unnecessary* string concat
>> inside an inner loop.
>
> I've laid this turkey of an issue to rest, but here goes.

Wrong then, wrong now...

> In older and less optimized versions of the now completely obsolete
> run-time interpreter for Visual Basic versions 6 and before,...

(RT Interpreter? VB has compiled to native code since, what, VB4?)

> ..occasionally, string concatenation inside a loop was marginally
> inefficient....

If you were anywhere close to being as smart as you believe you are,
you wouldn't need it explained that string cat operations--in any
language--on any platform--are **always** a speed issue and to be
avoided **if** **possible**.

That is, the requirement of allocating new space and moving the two
strings to it--whether you do it yourself (e.g. C) or the language
does it under the hood (e.g. VB)--is **always** significant enough
to consider during algorithm design.

The problem with a language, like VB, is that it allows less than ept
programmers to ignore these things.

> ...but typically in a way that is fully overshadowed by the end
> user's needs....

Pure Nigles BS at its best (or is that worst?). The user has no
"need" for ***unnecessary*** string cats.

> I have the very good habit, which you need to learn, "Programmer"
> "Dude", of writing self-verifiying code in the sense that inspection
> provides assurance of its behavior.

I prefer to do unit tests during development, regression tests as
the product builds, and "full scope" tests once it's working. That
eliminates the need for the code to verify itself constantly.

> Therefore I accept the modicum of inefficiency implicit in an inner
> concat,...

What you fail to recognize is that an improved design removes the need
and takes away none of your requirements. BS all you like, Ed. The
fact is, your algorithm design was naive. For someone who claims to
have been a working programmer for so long, it was embarrassing.

> This nonsense about it being a major crime to concat in a loop is an
> urban legend...

No, simple fact.

> It is at best a subcase of a more general rule, which is that "tight"
> loops should not contain "expensive" operations IF they are not
> necessary.

And as I've said over and over, it wasn't. I *showed* you how to
achive the same result--in all specifics--without it. All it takes
is an open mind, a little intelligence and a willingness to learn.

>>> Are you up to it? You might learn a thing or two. I will reply to
>>> any post that contains zero personal abuse.
>>
>> Fine, go for it. Let's see what you can do.
>
> This post contains personal abuse.

Hypocrite. I knew you'd weasel out of it. 

-- 
|_ CJSonnack <Chris@Sonnack.com> _____________| How's my programming? |
|_ http://www.Sonnack.com/ ___________________| Call: 1-800-DEV-NULL  |
|_____________________________________________|_______________________|