Neil, hope the lengthy discussion doesn't bog you down...maybe ignore for now.
Chris, I welcome further comments... I'll call this my pet topic for '03.
On Jul 3, 7:38am, Chris Pelkie wrote:
> Subject: Re: [opendx-users] Include/replace question
>
> Hi Sharon!
>
> The only thing I'd add or maybe clarify re (B) below is that Replace
> also serves as "Insert (even if not already present)", so there's no
> need to have an 'invalid' component in order to
> Replace[Insert](invalids) into a field.
True, Replace used in this case is actually an insert, and it might be
intuitive (although not to be taken for granted) that the thing that is being
inserted need not pre-exist in the destination field. But it's likely that for
Neil's case it is the source field (the mask) which has no invalid component to
"grab". I'm guessing that he's using Include:cull=1 to physically remove
subdomains of non-interest in the mask, but then trying to use the result of
that operation as a template to tag corresponding values in the "real" field
(Replace:invalid component). In doing a test of my own (using invalid
positions.. I assume same for connections), I only just now realized that with
cull=1 no invalid component is produced (makes sense - no need to refer to
those values later on if they're gone, but in all my work this specific item
has never come up...amazing). My surprise comes from DX not giving an error
when one tries to insert (Replace) something that doesn't itself exist. We're
not allowed to do much else on the non-existent (try Extracting the invalid
component here).
> Nevertheless, as we've both pointed out (to Neil), invalids are tricky. One
of > the most troubling things that has bitten us all (this is re (A) below),
is if > say "invalid connections" exist, but you then add or Replace ("invalid
> positions") into the field that aren't necessarily derived with reference to
> the "invalid connections". Weird things happen then.
I could list several other wierdnesses to look out for, if anyone's interested.
> On Wednesday, Jul 2, 2003, at 14:25 America/New_York, Sharon Gibson
> wrote:
>
> > (A) Note: You have to be careful if you have done any other Includes
> > on your field
> > prior to the masking.
> >
> > (B) I suspect that if
> > cull=1, you have no invalids to Replace (surprisingly, dx gives no
> > error).
> >
> >
> >
> >
> _______________________________
> Chris Pelkie
> Scientific Visualization Producer
> 622 Rhodes Hall, Cornell Theory Center
> Ithaca, NY 14853
>
> [ Attachment (text/enriched): 1184 bytes
> Character set: US-ASCII ]
>-- End of excerpt from Chris Pelkie
