Schrodinger - SMath

#1 Опубликовано: 01.10.2016 13:02:10

Jean Giraud

983 сообщений из 6866 понравились пользователям.

Группа: User

Smath does not "floor" n/2 in Hermite.
You can force: Floor(n/2)

Jean

Maths Schrodinger.sm (26 КиБ) скачан 66 раз(а).

1 пользователям понравился этот пост

frapuano 02.10.2016 06:36:00

#2 Опубликовано: 01.10.2016 15:08:28

NDTM Amarasekera

130 сообщений из 352 понравились пользователям.

Группа: User

Integration error in 98.6103??

Look within!... The secret is inside you. Best Regards Eng. NDTM Amarasekera - Sri Lanka

#3 Опубликовано: 01.10.2016 21:43:29

Jean Giraud

983 сообщений из 6866 понравились пользователям.

Группа: User

That's interesting for the designers team.
"k" is very back to Hermite summation !
98.6103 breaks the chain code.
Other versions ???

#4 Опубликовано: 03.10.2016 17:48:05

Andrey Ivashov

2269 сообщений из 3729 понравились пользователям.

Группа: Super Administrator

Wrote
Integration error in 98.6103??

I've checked it.

The reason is that 0^0 is undefined in latest versions of SMath Studio.

So when program tries to calculate integral of initial equation, it has to calculate (2*x)^{n-k}, where x=0, n=4 and k=4.
That is why error occurred. Will think what can be done here.

Regards.

#5 Опубликовано: 03.10.2016 19:44:13

Mike Kaganski

184 сообщений из 434 понравились пользователям.

Группа: User

Andrey,

Please take note of another side of this problem: the error message "k - not defined" is completely irrelevant to the real cause of the error. In general, in SMath it's too often that the correct error message isn't propagated from the code that knows what has actually happened to the UI.

While passing through exception handlers, the message could get modified. If it's impossible to choose the most relevant place where the problem may be explained best, I suggest to keep all the messages from the most low level in a hierarchical message, to present the user most complete information. There could be an UI option that controls this behaviour.

Could you please put this to your schedule?

С уважением, Михаил Каганский

1 пользователям понравился этот пост

sergio 04.10.2016 04:36:00

#6 Опубликовано: 04.10.2016 16:09:13

Andrey Ivashov

2269 сообщений из 3729 понравились пользователям.

Группа: Super Administrator

Wrote
Please take note of another side of this problem

Good approach in theory, but not applicable in real life (at least in current real life

).
0^0 error happens on sum() function processing and not actually means anything for int() who initiated calculation. int() decides that its body cannot be calculated on first attempt and simply eats exception to return int() itself (with all its body) for further actions. But the next action is numeric calculation, which proceeded by numeric engine which knows nothing about summation and integration, so it just stumbles on the first undefined variable (k) and throws this error to user.
You'll say "bad design"? Yes, most likely. But it cannot be fixed fast and simple... unfortunately.

Regards.

#7 Опубликовано: 04.10.2016 22:05:12

Jean Giraud

983 сообщений из 6866 понравились пользователям.

Группа: User

I understand what you are saying Mike about error messages not chaining
from the first error code. On that, Mathcad/Mathsoft was slightly better.
Often, in Smath 5346, going upward goes right to the error code.

The Hermite formula is totally correct. Here the sanity with my bible.
What's lovely with Smath 5346, at opening a work sheet it defaults to
symbolic. Thus, compose the formula, up it plots and resumes the
remaining of the project.

Jean