Re: isa strangeness

[Michael Goffioul]

On Mon, Aug 16, 2010 at 11:54 AM, Martin Helm <address@hidden> wrote:
> Am Montag, 16. August 2010, 12:40:31 schrieben Sie:
>> An example maybe...
>> Or just simply mentioning that the class argument is expected to be
>> a class name.
>>
>> Michael.
>>
> That was what came into my mind first, but then I thought there are already
> plenty of examples which can be seen by "type isa" (the tests at the end of
> the function).
> A new user is probably not aware that the source code can be shown with the
> command "type"?

Well, making implicit assumptions about what a user know/doesn't know in
documentation is not very good imo. This means that before being able to
know useful things about "isa", a user needs to know about "type" and that
there are (sometimes) embedded tests at the end of m-files.

IMHO documentation is made to document. I don't
see any problems in having verbose help strings with useful examples.
In this particular case, having an example as simple as "isa(x, 'double')"
would have clarified the ambiguity immediately.

Michael.