Maplesoft Entertainment Inc. presents proudly the Worldwide Maple 11 Roulette
Vladimir Bondarenko
$1895.00 only!
http://webstore.maplesoft.com/
----------------------------------------
# TEST CASE one of many 1000s
restart; f := 1/(sqrt(1+z)*sqrt(z^5+1)):
fnormal(evalf(int(f,z=0..infinity),20));
----------------------------------------
# 1000 tests
----------------------------------------
OUTCOME TIMES
----------------------------------------
1.136079748 - 0.*I 908
1.136079748 - 7.919214241*I 85
-.9478494454 - 3.282497480*I 1
1.136079748 - 3.959607121*I 1
1.136079748 - 7.162996572*I 1
1.136079748 + 0.*I 2
2.257333722 + 1.849170171*I 1
2.479580518 - 5.619722874*I 1
----------------------------------------
$1895.00 only!
http://webstore.maplesoft.com/
----------------------------------------
Faites vos jeux!
Vladimir Bondarenko
VM and GEMM architect
Co-founder, CEO, Mathematical Director
http://www.cybertester.com/ Cyber Tester, LLC
http://maple.bug-list.org/ Maple Bugs Encyclopaedia
http://www.CAS-testing.org/ CAS Testing
doug....@informath.org
I executed the above fifty times. Contrary to your claim, I got the
same answer every time:
1.136079748-0.*I.
This was on Maple 11.01 (running under Windows XP, SP2).
JB
I am using a two year old Dell desktop with XP and maple 11.0
and I click on !!! in the tool bar to re-initiate a new result. That
produced the following results:
try#1 1.136079748-0.*I
try #2 same answer as try #1
try #3 4.182684160-2.969705340*I
try#4 same answer as try #1
try#5 same answer as try #1
try#6 same answer as try #1
try#7 1.136079748-7.919214241*I
try#8 same answer as try #1
It appears that the situation may have been resolved in the
update from Maple 11.0 to Maple 11.01 which I have not made
yet. Which version were you using Vladimir?
doug....@informath.org
I misunderstood what VB wrote. VB had included a restart; so I
restarted each time. But the restart should only be done once. I
just executed
fnormal(evalf(int(xpr,z=0..infinity),20));
several times and got different results--same as VB reported. Sorry
about the confusion!
VB, I think you do good work. But perhaps you could be more positive
in your style. Also, this discussion does not really belong in
alt.math.recreational.
Nick
<doug....@informath.org> wrote in message
news:1183963980.6...@c77g2000hse.googlegroups.com...
And sci.math?
Nick
Vladimir Bondarenko
$1895.00 only!
http://webstore.maplesoft.com/
---------------------------------------------------------------
# TEST CASE one of many 1000s
#
# Enjoy FOUR distinct outputs at random.
#
---------------------------------------------------------------
restart; evalf(int(ln(1+z^2)/(1+z^2+z^4), z= 0..infinity));
Error, (in assuming) when calling 'signum/... division by zero'
.3718816641+0.*I
.3718816648+0.*I
.3718816641-.737e-9*I
.3718816638+.5e-9*I
.3718816630+.329e-9*I
.3718816642-.37e-10*I
Error, (in assuming) when calling 'signum/... division by zero'
1.644934067
.3718816644+0.*I
.3718816634-.15e-8*I
1.644934067
Error, (in assuming) when calling 'signum/... division by zero'
.371881665+.4e-10*I
.3718816639-.18e-9*I
.3718816639+.11e-8*I
.3718816658-.397e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816661+.3e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816656-.137e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816647-.5e-9*I
.3718816632+0.*I
.3718816652-.10e-8*I
.3718816643+0.*I
.3718816639+0.*I
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
.3718816637+0.*I
.3718816647+0.*I
.3718816649+0.*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816644+.5e-9*I
.3718816647-.5e-9*I
.3718816644-.5e-9*I
.3718816643+.234e-9*I
.3718816653-.46e-10*I
.371881663-.14e-9*I
.3718816644-.37e-10*I
.3718816650-.32e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816641+.286e-9*I
.371881663-.23e-9*I
.371881663+.14e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816653+0.*I
.3718816649+.26e-10*I
.3718816628+.1774e-8*I
.3718816644-.4e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816651+0.*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816643+.1e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in signum) too many levels of recursion
Error, (in assuming) when calling 'signum/... division by zero'
.3718816640+.29e-10*I
Error, (in assuming) when calling 'signum/... division by zero'
.371881664-.24e-9*I
.3718816646-.537e-9*I
.3718816635-.6e-9*I
.3718816647-.1e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
.3718816653+0.*I
.3718816650+0.*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816637+.71e-10*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816661+.4e-9*I
Error, (in signum) too many levels of recursion
.3718816642+.2e-10*I
.3718816633+0.*I
.3718816642+0.*I
.3718816641-.4e-9*I
.3718816633-.337e-9*I
Error, (in signum) too many levels of recursion
.3718816634+0.*I
.3718816643-.129e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816635+.2e-9*I
.3718816632+0.*I
.371881663+.74e-9*I
.3718816654-.526e-9*I
.3718816641+0.*I
Error, (in assuming) when calling 'signum/... division by zero'
Error, (in assuming) when calling 'signum/... division by zero'
.3718816651+.1e-9*I
.3718816665+.12e-8*I
.3718816639-.1e-10*I
.3718816643+.26e-10*I
.3718816631+.329e-9*I
.3718816643+.1e-9*I
Error, (in assuming) when calling 'signum/... division by zero'
.3718816643-.1e-9*I
.3718816639+0.*I
---------------------------------------------------------------
$1895.00 only!
http://webstore.maplesoft.com/
---------------------------------------------------------------
quasi
<v...@cybertester.com> wrote:
>restart; evalf(int(ln(1+z^2)/(1+z^2+z^4), z= 0..infinity));
>
>Error, (in assuming) when calling 'signum/... division by zero'
Even if you replace infinity by 1 it acts badly.
evalf(int(ln(1+z^2)/(1+z^2+z^4), z= 0..1));
Maple apparently ignores the evalf and gives a rather scary looking
symbolic result.
However, the workaround is easy. The following commands both give
correct answers:
evalf(Int(ln(1+z^2)/(1+z^2+z^4), z= 0..1));
evalf(Int(ln(1+z^2)/(1+z^2+z^4), z= 0..infinity));
quasi
Vladimir Bondarenko
Q> the workaround is easy.
I am afraid, this is not a workaround at all. I used evalf() to
just produce a compact output showing that there is a menacing
disease -- disease the long liver! -- of bad, invalid, false
randomness in the Maple engine.
Maple engine's output depends on the Moon's phase. We came
to Maplesoft NOT for numbers but for formulas.
http://groups.google.com/group/sci.math.symbolic/msg/37999f6e326ba78d
Please let me add some clarifications.
Why millions persons use computer algebra systems?
Well, there are many reasons for it, but most probably it's
because, in contrast to piles of numbers showering from
THOUSANDS amateurish and professional numerical packages, a
computer algebra system can supply you with the highest math
aesthetics, one of the biggest math diamonds, with a - be in
awful fear, the numerical folks! - with Her Majesty FORMULA,
all hues of the rainbow over Her.
Not with a rank-and-file private, a pack of digits.
Thus, my message is about the full palsy in the symbolic mode
with maximize. Numeric calculation is trivial here, and it is
at the case neither here not there.
Let us imagine such a comparison.
You go to a shop (Maplesoft), and you see at the showcase a
yummy almond cake with custard and dusted with caster sugar,
the cat's whiskers (Maple, Help, ?maximize).
Then you enter the shop and ask the salesgirl the blonde job
about selling you this dainty bit, this very moment!
She agrees, and you hand her money (- Gotcha! - she adds
beneath her breath), and get the money payment receipt, and
a nice colorful parcel. Life *is* good! Purr-purr-purr...
Broiling with impatience, you order a cup of tea on the spot,
and are going to relish this wonder of higher cooking!!
You tear up the shrink-wrapper, and... and see... large dented
pieces of lump sugar fit to hammer nails in absence of a
hammer??!!
Strike me dead! - give you a cry - this is a piece of bamboozle!
Where on earth my pet yummy almond little cake with custard and
dusted with caster sugar?!
However, the belle gives you a chirp, Oh no, this sweet stuff
at our showcase is just a DEMO, this is why we put it into the
showcase (Maple Help, ?maximize). Actually, it does not work.
What we can offer you is the pack of lump sugar you already got.
- Okay, - say you hastily, - I want the refund.
Now the salesgirl sneers disdainfully.
Go to junior school, - she says between her teeth, - and learn
the ABC. Then, read our license agreement, carefully.
THERE IS NO REFUND. WE ARE NOT LIABLE FOR ANYTHING.
Take a load of this at last:
Maplesoft's guys are real programers.
1. Real programmers don't write specs.
2. Real programmers don't read manuals.
3. Reliance on a reference is a hallmark
of the novice and the coward.
4. Users should be grateful for whatever they get.
5. They are lucky to get any program at all.
Thanks for the purchase, don't let me keep you!
----------------------------------------------------------------
He that hath ears to hear, let him hear.
Best wishes,
Vladimir Bondarenko
http://www.cybertester.com/
http://maple.bug-list.org/
http://www.CAS-testing.org/
P.S.
You ARE clever. You can beat this girl's trade show,
but this is another story to be told later.
Vladimir Bondarenko
http://www.mapleprimes.com/forum/reproducibility
We do call these ordering problems. It can lead
to intermittent failure to produce any answer.
Sometimes it can lead to intermittently computing
an incorrect result.
Dave Linder
Mathematical Software, Maplesoft
-------------------------------------------------
$1895.00 only!
http://webstore.maplesoft.com/
-------------------------------------------------
# TEST CASE one of many 1000s
#
# Enjoy TWO distinct outputs at random.
#
restart: lcoeff(v*z - a*z);
-------------------------------------------------
# 1000 tests
-------------------------------------------------
OUTCOME TIMES
-------------------------------------------------
1 926
-1 74
-------------------------------------------------
$1895.00 only!
http://webstore.maplesoft.com/
Stein Arild Strømme
| restart: lcoeff(v*z - a*z);
|
| -------------------------------------------------
| # 1000 tests
| -------------------------------------------------
| OUTCOME TIMES
| -------------------------------------------------
| 1 926
| -1 74
So what? This is well documented in ?lcoeff:
- When x is not specified, the order of the indeterminates is given by indets (more
specifically, frontend(indets,[p],[`*`,`+`,`::`,constant,series,SDMPolynom,undefined]);).
In the multivariate case this ordering is session dependent.
SA
--
Stein Arild Strømme +47 55584825, +47 95801887
Universitetet i Bergen Fax: +47 55589672
Matematisk institutt, UiB http://math.uib.no/stromme/
Johs Brunsg 12, N-5008 BERGEN str...@math.uib.no
Vladimir Bondarenko
SAS> So what? This is well documented in ?lcoeff:
Yes, it is well documented. But what Dave Linder,
Maplesoft, claims
http://www.mapleprimes.com/forum/reproducibility
"We do call these ordering problems. It can lead
to intermittent failure to produce any answer.
Sometimes it can lead to intermittently computing
an incorrect result."
> Johs Brunsg 12, N-5008 BERGEN stro...@math.uib.no
Stein Arild Strømme
| On Jul 13, 6:12 am, stro...@math.uib.no (Stein Arild Strømme) writes:
|
| SAS> So what? This is well documented in ?lcoeff:
|
| Yes, it is well documented. But what Dave Linder,
| Maplesoft, claims
|
| http://www.mapleprimes.com/forum/reproducibility
|
| "We do call these ordering problems. It can lead
| to intermittent failure to produce any answer.
| Sometimes it can lead to intermittently computing
| an incorrect result."
You are misquoting Linder, he did not end the sentence that way. Here
is the full paragraph :
These kind of effects discussed here can result in different
computation paths being taken, as has been noted in this thread by
a few people. We do call these ordering problems. It can lead to
intermittent failure to produce any answer. Sometimes it can lead
to intermittently computing an incorrect result, which is far more
serious. Ordering problems are not dismissed out of hand as being
unimportant.
The pages puts forward a rather good explanation of the problem, in my
view.
--
Stein Arild Strømme +47 55584825, +47 95801887
Universitetet i Bergen Fax: +47 55589672
Matematisk institutt, UiB http://math.uib.no/stromme/
Johs Brunsg 12, N-5008 BERGEN str...@math.uib.no
Vladimir Bondarenko
SAS> The pages puts forward a rather good explanation
SAS> of the problem, in my view.
Yes, I agree with you. A nice discussion. In my view, especially
impressive comments belong to Prof. Jacques Carette who was for
long years a Waterloo Maple/Maplesoft top officer.
What's important, here Dave Linder admits in plain language that
there is a hereditary disease in ALL Maple versions (!) since at
least 1992 to 2007, that is during 25 years.
What Cyber Tester can add as QA engineers, over the last years
this false non-deterministic contagion propagates more and
more over Maple engine.
We are worried much as this could, being multiplied by other
unsolved development problems, gradually put the end to Maple
usefulness.
Best wishes,
Vladimir Bondarenko
VM and GEMM architect
Co-founder, CEO, Mathematical Director
http://www.cybertester.com/ Cyber Tester, LLC
http://maple.bug-list.org/ Maple Bugs Encyclopaedia
http://www.CAS-testing.org/ CAS Testing
> Johs Brunsg 12, N-5008 BERGEN stro...@math.uib.no
jmor...@idirect.com
>What's important, here Dave Linder admits in plain language that
>there is a hereditary disease in ALL Maple versions (!) since at
>least 1992 to 2007, that is during 25 years.
Now who needs an accurate math package? :)
rjf
....
>
> What's important, here Dave Linder admits in plain language that
> there is a hereditary disease in ALL Maple versions (!) since at
> least 1992 to 2007, that is during 25 years.
>
> What Cyber Tester can add as QA engineers, over the last years
> this false non-deterministic contagion propagates more and
> more over Maple engine.
>
1. Thanks for pointing us to the MaplePrimes group, which clearly
has improved since I last visited it. It demonstrates rather
conclusively that people who know about Maple or want to know about
Maple can find more authoritative and informative data there. By
reading VB's posts here you might think the Maple technical people are
deaf; what you learn by visiting MaplePrimes is that they are quite
responsive. They just don't care to respond here, and who can blame
them, as long as VB tries to get in the record books for most content-
free irrelevant cross-postings.
2. Regarding the non-determinism of ordering. This has been an
issue discussed from the earliest days of the Maple design. Anyone
who writes a script that depends on ordering of terms should be
careful to SORT the terms before picking out (sight unseen) the 3rd
term of a result. There is an excellent discussion of the matter in
the MaplePrimes llink above. I would like to add only one point. I
have, several times, come up with algorithms for different tasks that
were substantially faster than the standard (in Mathematica, Maxima),
based on a novelty (for them) like hashcoding.
But Maple sometimes was faster yet. Why? Because it already used
hashing but did not have to sort the terms. Mathematica and Maxima
generally view a sum of 10,000 terms as something that must be
"properly" ordered to be considered simplified. Maple just blasts the
sum out using its internal order. Asking Maple to sort it properly
considerably slows down the process, so Maple is no longer fast. Of
course the original Maple sum IS not randomly ordered. It is just that
the ordering predicate is determined by a mechanism that is mostly
opaque to the user. Defenders of Maple have had to deal with this
issue since the original design, and seem resistant to changing it.
Mathematica also has what appears to be non-deterministic behavior,
apparently in some cases from caching results, and in other cases from
re-evaluating expressions because they were allocated space on the
same page of memory as others. It is unfortunate that in the
marketplace for software, speed tends to be the most prominent measure
of quality.
Vladimir Bondarenko
writes:
JM> Now who needs an accurate math package? :)
A wildly bold hypothesis, The CAS customers??!!
On Jul 13, 9:31 am, "jmorr...@idirect.com" <jmorr...@idirect.com>
wrote:
Axel Vogt
want to understand him, I find it really funny how VB handles
communication:
first he uses commands in obviously false manner and then talks
about something quite different ... like a trilling troll ...
Vladimir Bondarenko
Why don't you see thousands Maple defects?
Why don't you comment this regression bug?
http://groups.google.com/group/sci.math.symbolic/msg/7140b8f4d772f3b1?hl=en&
Maple 11> evalf(Int(1/(z^exp(I)+z^Pi), z=1..infinity));
-.5000000000
?
Best wishes,
Vladimir Bondarenko
Axel Vogt
> You see my way of quoting and comment it.
>
> Why don't you see thousands Maple defects?
May be because I handle Maple more adequate?
Not everything you post is a bug and as others already said:
there are lots of repetitions.
And more personal: because there are other interesting things
as well (besides a job ...)
> Why don't you comment this regression bug?
>
> http://groups.google.com/group/sci.math.symbolic/msg/7140b8f4d772f3b1?hl=en&
>
> Maple 11> evalf(Int(1/(z^exp(I)+z^Pi), z=1..infinity));
>
> -.5000000000
>
> ?
Done.
Vladimir Bondarenko
http://www.mapleprimes.com/forum/reproducibility
We do call these ordering problems. It can lead
to intermittent failure to produce any answer.
Sometimes it can lead to intermittently computing
an incorrect result.
Dave Linder
Mathematical Software, Maplesoft
----------------------------------------------------------------
$1895.00 only!
http://webstore.maplesoft.com/
----------------------------------------------------------------
# TEST CASE one of many 1000s
#
# Enjoy TWO distinct outputs at random.
#
#
restart; int(1/(z^4+1/sqrt(z^2+1)+1), z= -infinity..infinity);
limit((sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R^5...
-limit((-sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R...
limit((sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R^5...
0
-limit((-sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R...
limit((sum(_R*ln(1/103*(-103+54317952*_R^7*z+1366560*_R^5*z+7...
-limit((-sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R...
limit((sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R^5...
0
0
0
0
limit((sum(_R*(-ln(103)+ln((-103+54317952*_R^7*z+1366560*_R^5...
0
limit((sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R^5...
0
0
limit((sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R^5...
limit((sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R^5...
-limit((-sum(-_R*(ln(103)-ln((-103+54317952*_R^7*z+1366560*_R...
----------------------------------------------------------------
$1895.00 only!
http://webstore.maplesoft.com/
----------------------------------------------------------------