comp.lang.idl-pvwave archive: archive » Re: Optional parameters

Home » Public Forums » archive » Re: Optional parameters

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: Optional parameters [message #38191 is a reply to message #38190]

Tue, 24 February 2004 13:45

Pavel Romashkin
Messages: 166
Registered: April 1999

Senior Member

And what exactly would be the purpose of such sequence of calls, other
than actually trying to *define* foo in the first function call? If I
didn't need foo re-defined, I'd not even mention it on the first line.
So this *is* the case for 3 copy-paste calls, I guess :-)
I never tried to use keywords for this sort of creative purposes. If I
have a large dataset I don't want to re-compute, I keep it in a pointer
(or object, allrigh, David? :-) and I try to always know when it gets
passed into a function to avoid redefinition and other mishaps.
After all, how is this behavior different than simply

foo = 100 ; foo is now defined, too...

If we did want to get funny I suppose we could try

a=myfunc(test,FUNNY_VARIABLE=[foo]) ; foo didn't change

but if foo is BIG, we'd get a dramatic slowdown.
Pavel

JD Smith wrote:
>
> I'd offer another real life example where read-write keyword variables
> can get you into trouble. When you call a function several times in a
> row, expecting the same results each time, but in actuality the
> keyword variable is defined after the first invocation, which can
> change the function's behavior:
>
> a=myfunc(test,FUNNY_VARIABLE=foo) ; foo undefined
> a=myfunc(test,FUNNY_VARIABLE=foo) ; foo defined
> a=myfunc(test,FUNNY_VARIABLE=foo) ; foo defined
>
> For this reason, I try to make keywords either exclusively read
> (i.e. pass in a value), or exclusively write (i.e. return a value),
> and not both. It is rather convenient, though, to append a keyword
> for the sole purpose of keeping some expensive-to-compute data around
> between calls to the function: this is usually where I get tripped up.
>
> JD

Report message to a moderator

[Message index]

		Re: Optional parameters By: Pavel Romashkin on Wed, 25 February 2004 10:21
		Re: Optional parameters By: David Fanning on Wed, 25 February 2004 09:26
		Re: Optional parameters By: Craig Markwardt on Wed, 25 February 2004 08:48
		Re: Optional parameters By: David Fanning on Wed, 25 February 2004 07:08
		Re: Optional parameters By: Paul Van Delst[1] on Wed, 25 February 2004 06:49
		Re: Optional parameters By: Craig Markwardt on Tue, 24 February 2004 21:50
		Re: Optional parameters By: JD Smith on Tue, 24 February 2004 14:03
		Re: Optional parameters By: R.G. Stockwell on Tue, 24 February 2004 14:02
		Re: Optional parameters By: Pavel Romashkin on Tue, 24 February 2004 13:45
		Re: Optional parameters By: JD Smith on Tue, 24 February 2004 12:13
		Re: Optional parameters By: Pavel Romashkin on Tue, 24 February 2004 09:59
		Re: Optional parameters By: David Fanning on Tue, 24 February 2004 09:08
		Re: Optional parameters By: David Fanning on Tue, 24 February 2004 08:31
		Re: Optional parameters By: JD Smith on Tue, 24 February 2004 08:10
		Re: Optional parameters By: David Fanning on Mon, 23 February 2004 17:39
		Re: Optional parameters By: David Fanning on Mon, 23 February 2004 17:05

Previous Topic:	Re: pointer and structure stuff ...
Next Topic:	Re: Something similar to a "Widget_Container" ?

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Tue Dec 02 23:54:29 PST 2025

Total time taken to generate the page: 3.21405 seconds