(Last line with error is just to demonstrate that without vectorize(), the error is the same as with vectorize(), so seems that vectorize() has no effect)

I suspect that the problem is because the Bessel function is defined in a IPluginLowLevelEvaluation plugin (StatisticalTools by Davide). This interface is afaik the most used among SMath plugin writers, so it would be desirable to extend vectorize() to work correctly with that kind of plugins.

Please excuse me if my reasoning is incorrect, and the real cause is something other.

Edited by moderator 23 April 2016 12:10:41(UTC)  | Reason: marked as solved

OH MY! Not again!







I only used any first custom function from my function list, just to demonstrate the problem. I could use e.g. CoolProp wrapper, or anything else.

Hello.


Everything you said is correct. I've just improved vectorize function to support functions from plug-ins.



Thank you!

Best regards.

Hi all, did not want to open a new vectorize() topic so posting my question here. From what I know about vectorize is that in order for it to evaluate you need to satisfy the following:

1. Arguments must be two matrices with matching dimensions
2. Operator must be applicable to scalar variables

If the above is correct, I might have detected an odd behaviour.

And as a side note: Andrey, please do not get discouraged by all the recent controversy surrounding vectorize(). It is a very useful function. My only suggestion would be to make sure that it consistently follows a simple list of logical rules, and to make this list available in the SMath release notes / documentation.

http://smath.info/cloud/sheet/xVqGb4rhZq

vectorize1.sm (10kb) downloaded 20 time(s).

Well, then I will add here, too



This minimal example is completely made up for demonstration only. No meaning in it other than to demonstrate a problem.

Here three pieces are relevant: if(), assignment of y, and eval(). Without any of these, it works.

Edited by user 12 April 2016 11:19:18(UTC)  | Reason: Not specified