comp.lang.idl-pvwave archive: archive » Re: Is a dynamically sized pointer array object component possible?

Home » Public Forums » archive » Re: Is a dynamically sized pointer array object component possible?

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

Re: Is a dynamically sized pointer array object component possible? [message #66620 is a reply to message #66525]

Fri, 22 May 2009 10:05

Paul Van Delst[1]
Messages: 1157
Registered: April 2002

Senior Member

mgalloy wrote:
>
> Another way to do it would be for "::allocate, n" to just create a
> pointer to a single vector of size TOTAL(n) and also save the n array,
> then for "::get, i, j" to use TOTAL(n, /CUMULATIVE) to find the correct
> value(s). I'm not sure that would be simpler, but it would eliminate the
> double dereference.

Ah, yes. That is actually how I do it now and is the reason I am changing it. For some
larger context, the problem is when we have satellite instrument channel responses that
contain significant regions of zero response (e.g. like microwave instruments where the
channel response can straddle atmospheric absorption lines). Lumping the response
altogether exposes our integrator to numerical problems and, since we use the first moment
of the response curves as the channel central frequencies, that can eventually bugger up
our radiance calculations.

I'm going with the pointer to a pointer array approach and just deal with the double
dereference. As someone paraphrased Voltaire in a meeting I was at recently: perfection is
the enemy of good enough. :o)

cheers,

paulv

Report message to a moderator

[Message index]

		Re: Is a dynamically sized pointer array object component possible? By: Michael Galloy on Fri, 22 May 2009 09:45
		Re: Is a dynamically sized pointer array object component possible? By: Paul Van Delst[1] on Fri, 22 May 2009 06:30
		Re: Is a dynamically sized pointer array object component possible? By: Michael Galloy on Thu, 21 May 2009 15:21
		Re: Is a dynamically sized pointer array object component possible? By: Paul Van Delst[1] on Fri, 22 May 2009 10:05

Previous Topic:	OT: Re: HDF vs HDF-EOS?
Next Topic:	Re: FOR loops and efficiency

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Sat Oct 11 19:20:48 PDT 2025

Total time taken to generate the page: 0.64179 seconds