[MOAB-dev] ReadCGM question

[Tim Tautges]

I think this has less to do with merge tolerance than just plain tolerance issues with the underlying model, kind of 
like what you see when the end of a curve isn't at the same place as the start of a connected one.  So, I'd say print a 
limited number of warnings, or a list of curves showing the problem, and call it good.  One could argue here that if CGM 
isn't printing the warning, MOAB shouldn't, though maybe the facet-based algs are sensitive to this type of problem.

- tim

On 11/03/2010 04:59 PM, Jason Kraftcheck wrote:
> On 11/03/2010 03:59 PM, Steve Jackson wrote:
>>
>> On Nov 3, 2010, at 15:53 , Jason Kraftcheck wrote:
>>
>>> On 11/03/2010 03:43 PM, Steve Jackson wrote:
>>>
>>>>
>>>> * I think it would it be more correct to replace GEOMETRY_RESABS
>>>> with faceting_tol, i.e. the tolerance value that was used in calling
>>>> edge->get_graphics().  Does anyone disagree?
>>>
>>> Perhaps you could change it to be the merge tolerance.  Merging is
>>> (hopefully) the source of such warnings.
>>
>> You anticipated my next question, which was why this error was happening.
>> (I.e. do I blame the geometry, or blame the faceting engine?)
>
> My guess was the geometry, because I don't see how or why the faceting
> engine would introduce such an issue (other than a plain old bug.)
>
>> I don't
>> know of a way to get at the merge tolerance from the context of the
>> error, since merging is done to a geometry before ReadCGM gets access to
>
> After thinking about it for a bit I think my suggestion of using the merge
> tolerance was a poor one because Cubit only ever tracks (AFAIK) the merge
> tolerance that it will use for future merges.  What you'd actually need is
> the merge tolerance what was enforced during the merge of each entity (or
> whether a 'forced' merge was done, which would have bypassed tolerance checks.)
>
>> it.  But perhaps when this error is prevalent, it suggests that the
>> geometry was created with a bad choice of merging tolerance? ~S
>>
>
> It suggests that perhaps that GEOMETRY_RESABS doesn't accurately reflect the
> user's idea of what constitutes equivalence.  And emphasizes the fact that
> cubit is somewhat inconsistent in that regard (default merge tolerance is
> greater than GEOMETRY_RESABS).
>
> - jason
>
>

-- 
================================================================
"You will keep in perfect peace him whose mind is
   steadfast, because he trusts in you."               Isaiah 26:3

              Tim Tautges            Argonne National Laboratory
          (tautges at mcs.anl.gov)      (telecommuting from UW-Madison)
          phone: (608) 263-8485      1500 Engineering Dr.
            fax: (608) 263-4499      Madison, WI 53706