| Kent Paul Dolan 2006-01-31, 7:59 am |
| Chris Smith wrote:
> Kent Paul Dolan <xanthian@well.com> wrote:
[color=darkred]
> Yes.
[color=darkred]
> No, it remains negligible.
My, my. You don't have much grasp of "asymptotic
complexity" even as a term in English, much less as
a concept in math, do you?
> You never actually reach infinity, of course.
Strangely, calculus, for example, does just that
quite regularly.
> When I said "but not until the program would have
> required infinite space anyway", that should be
> interpreted as "never."
Perhaps by you, but not by informed observers.
> The additional space requirement implied by a
> multiplier of 1.01 may grow very large indeed from
> the perspective of a small problem.
That doesn't parse as math.
> If the program would otherwise require 100 TB,
> then it's a whole terabyte, which is indeed quite
> expansive if it were required space on my desktop
> PC.
So, in your mind, "no, it remains negligible" is
compatible with "it's a whole terabyte", a file size
larger than most commercial disk drives? My, my.
> For an environment with 100 TB in storage already
> available, though, another terabyte is negligible.
You don't have much grasp of computing technology,
either. There are _lots_ of hardware lashups which
have huge base storage capacity, and relatively
tiny freely manipulable storage capacity. Systems
built on huge ROM chip immutable data stores, with
4Kbytes or so of RAM data space come to mind, and
are very common.
> The same thing applies, regardless of the fact
> that the additional storage approaches infinity.
So "the additional storage approaches infinity" and
"no, it remains negligible" are compatible concepts
in your mind? My, my.
> If it ever reached infinity, of course, we
> wouldn't have the tools to reason about it
"Reasoning about infinity" has been quite an
accepted part of math ever since Cantor categorized
the various infinities, but then, you don't seem to
have much math background either.
> ... but that's a direct consequence of the fact
> that it's nonsensical for the problem to become
> infinitely large.
Ah, but the behavior described by complexity theory
is about problems' asymptotic behavior when
approaching infinity, not about what they do once
they get there.
In the case of your "no, it remains negligible", it
is a well known theorem of calculus that for any
quantity that becomes asymptotically infinite, there
is _no_ value you can pre-specify in its range that
it will not exceed for some finite value in its
domain. So, if you concede that "one quintillion
bytes" is no longer "negligible", then you are
forced also to concede that an asymptotically
infinite storage increase will exceed one
quintillion bytes for some finite problem size N.
> This is, in fact, the central point of algorithmic
> complexity. Constant multiples are not
> communicated because they don't matter in any
> problem for which algorithmic complexity is an
> appropriate mechanism for considering the problem.
> If you really did have a machine with exactly
> 100.01 TB of storage and needed to solve a problem
> that requires 100 TB of input, you would be better
> off explaining that the algorithm can't use more
> than 10 GB of space, and abandoning the big-O
> nonsense.
Here we have the gist of the matter: you consider
big-O complexity expression to be "nonsense", and so
never bothered to understand what it means. Why,
then, do you insist on giving advice about matters
you do not understand, and insist on defending your
incorrect blathering in posting after posting,
despite being contradicted by several more
knowledgeable posters?
[color=darkred]
> You don't need to propose such a ludicrous idea.
Sorry, but "big ROM, small RAM" is a very common
computing environment for embedded computing
devices, so it is hardly "a ludicrous idea".
> I already gave a more realistic example earlier in
> this thread: a traditional compiler that reads and
> handles the source as it passes through the lexer
> without storing it. The disk is a much larger
> storage media than RAM, and I explained why an
> algorithm like that could reasonably talk about
> better than O(n) space. No one said that this
> matrix transpose algorithm meets that description.
You _still_ fail to concede the utility of
describing something as having O(1) complexity?
My, my.
SRR: An O(1) Time Complexity Packet Scheduler for
Flows in Multi-Service Packet Networks
Guo Chuanxiong
Inst. of Comm. Eng.
P.O.Box 110, 2 Biaoying, Yudao st.
Nanjing, 210016, China
xguo@ieee.org
SIGCOMM'01 August 27-31, 2001,
San Diego, California, USA.
http://portal.acm.org/ft_gateway. c...
OKEN%3D2222222
Aliquem: a novel DRR implementation to achieve
better latency and fairness at O (1) complexity
L Lenzini, E Mingozzi, G Stea - Quality of Service,
2002. Tenth IEEE International Workshop, 2002 -
ieeexplore.ieee.org
http://ieeexplore.ieee.org/xpl/abs_...umber%3D1006576
and so on for some 1,200,000 hits in Google
Scholar on "O(1) Complexity".
> Anyway, obviously you're being an ass. Don't
> bother.
How nice for you that you can continue to bray
insults and ignorance in the face of all evidence
and advice. Far be it from me to get between you and
your all hay diet.
> www.designacourse.com
> The Easiest Way To Train Anyone... Anywhere.
Sure, training is "easy" if it doesn't suffer a need
to be "correct".
> Chris Smith
> - Lead Software Developer/Technical Trainer
How  for your clients to have you as a trainer.
> MindIQ Corporation
Evidently, "MindIQ" is anther "O(1)" datum.
xanthian, still amused.
A few pertinent siggie quotes to rub salt in your
wounds:
==
karlt96@nox.nyx.net (Karl Thomas):
> Let's try this again.
> You seem to have the same problem that he had.
> Care to come up with something more technical
> than he was able to muster?
> Or do you just idolize Kent so much if he said
> the world was flat you would believe him?
Uh, let me put this to you gently,
if the world was really flat
he would be one of the people who knew.
Do you have a problem being instructed
by someone much smarter and
a hell of a lot more experienced than you?
-- Terry_Palfrey@mindlink.bc.ca
(Terry Palfrey)
==
Of course not.
Kent is about the most immature person for his
age and encyclopedic knowledge as (sic) one can
possibly imagine.
He is a freak of nature.
-- John P Sheehy <jsheehy@ix.netcom.com>
==
Diverse people have different skills.
Kent Paul Dolan's skill is being smart.
-- Mary Lee McGough <marylee@csufresno.com>
Founder and Leader of the FSU Poetry Jam
==
and so on for a few hundred sometimes grudging
bits of praise:
http://www.well.com/user/xanthian/bin/incl
|