if x is 1m, the loop fails, if x=1, it succeeds. Any idea, what is going on? How to explain?


unit-range.sm (4 КиБ) скачан 38 раз(а).

Wrote

Any idea, what is going on? How to explain?

Old style works OK.

vectorize also works independently of previous values of x. It is just the above example of implicit loops. I am not searching for alternatives but want to have robust implicit loops. They contribute to readable documents. Closer to pencil and paper handcalc than to algorithmic pseudocode.

when you assign 1m to x, it doesn't accept to be vectorized with range function afterwards.
making x a vector firstly, the problem is gone.
my diagnosis is; data type error. similar to C++, Basic, etc

regards

Wrote

when you assign 1m to x, it doesn't accept to be vectorized with range function afterwards.
making x a vector firstly, the problem is gone.
my diagnosis is; data type error. similar to C++, Basic, etc

regards

Sorry, but the problem is that the presence or absence of the unit in the existing value of x should not matter regardless what you are going to assign to x afterwards.

The failing implicit loop doesn't have any units involved, I would expect the loop to cleanly overwrite whatever pre-existing scalar value. I am well aware that pre-existing matrix values may proagate to the result if the dimensions are bigger than what is adressed in the loop.

Also, if there is some problem which can't be fixed, I'd vote for a better error message. In the given case, "Cannot calculate" will leave most users in frustration and dispair.

BTW, the first statement in your example just crushes the vectorize() function. It doesn't affect the actual value of y.

yeah Martin you are right a out my vectorize solution.
we expect to overwrite the values but somehow it doesn't.
we can use clear(x) before the new assignment temporarily, while this is inspected and corrected if necessary.

regards

ps: if x is a matrix then it seems this problem disappears also.

Hi. THis also fails too. Actually, it is a short form for the function NameVec in mathcad.



Best regards.
Alvaro.

Wrote

I'd vote for a better error message. In the given case, "Cannot calculate" will leave most users in frustration and dispair.

"can not calculate" because "syntax error"
"can not calculate" also appears in 6179 and 5346
resulting from more than a single "syntax error"

Thank you we'll check it. I agree that the behavior should be the same regardless if there is an unit.

Wrote

Hi. THis also fails too.

There's no indexing in your LHS; the syntax you were looking for is the following:

Wrote

"can not calculate" because "syntax error"
"can not calculate" also appears in 6179 and 5346 resulting from more than a single "syntax error"

This is because implicit assignment loops by vectors are available since SMath Studio 0.99.6638, in previous versions it's invalid syntax.

Wrote

This is because implicit assignment loops by vectors are available since SMath Studio 0.99.6638, in previous versions it's invalid syntax.

Yes Davide, missing for local implicit previous 6638.
It turns red some of Ber7 & Alvaro stuff.
For the essentials, I don't miss local implicit loop.

Cheers ... Jean

Wrote

Wrote

Hi. THis also fails too.

There's no indexing in your LHS; the syntax you were looking for is the following:
...

Oh, yes, thanks. That's correct.

Best regards.
Alvaro.