2004-march.txt

acis说明文档

From zz_zxc123 at cims.sjtu.edu.cn  Mon Mar  1 08:53:11 2004
From: zz_zxc123 at cims.sjtu.edu.cn (zhang xuechang)
Date: Sun Feb 29 19:50:39 2004
Subject: [Acis-alliance] 
	Ask a function of computering the distance between a point and a
	surface
Message-ID: <003f01c3ff27$94e8ad70$c300a8c0@zxc>

Hi,everyone.
   Now i want to calculate the distance between  points and  surface.I think there is a api function to realize it.so please tell me it .

thanks,thanks.

zhangxuechang
2004.3.1
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.cs.columbia.edu/pipermail/acis-alliance/attachments/20040301/b4e86fbc/attachment.htm
From nil.wehr at softwehr-service.de  Mon Mar  1 11:56:06 2004
From: nil.wehr at softwehr-service.de ([e-mail])
Date: Mon Mar  1 05:59:01 2004
Subject: [Acis-alliance] [e-mail]
Message-ID: <5694195$[e-mail]>


Hello,


I want to create a hollow cylinder (C++ API).

Is it more easy to use boolean operations by using 2 cylinders or should
I create a revolution of a face.

Did anybody get experiance by creating objects like that?

Thank you for your support

Juergen Wehr
From nil.wehr at softwehr-service.de  Mon Mar  1 12:02:01 2004
From: nil.wehr at softwehr-service.de ([e-mail])
Date: Mon Mar  1 06:05:45 2004
Subject: [Acis-alliance] create complex BODY
Message-ID: <5694195$[e-mail]>


Hello,


I want to create complex bodies by using
vetices->SPApostion->edge->coedge->wire -> ... sweep straight along axis
-> BODY

I can create the wire by setting the first coedge (which points to the
following coedge ...).

But how do I use the sweep-algorithm to create BODY-objects?

If anybody could send an example (even simple geometry like box, made by
sweep) I would be(very) happy.

Also a link to an example would support myself.


Thank you for your support



Jeurgen Wehr
From kaufmann at ims.tuwien.ac.at  Mon Mar  1 13:53:03 2004
From: kaufmann at ims.tuwien.ac.at (Hannes Kaufmann)
Date: Mon Mar  1 07:53:41 2004
Subject: [Acis-alliance] create complex BODY
References: <5694195$[e-mail]>
Message-ID: <[e-mail]>

As far as I have seen it is not possible with ACIS to sweep whole bodys
around an axis. You can only sweep profile curves.
The sweeping does not seem to be very sophisticated. I don't even get a
valid result if I sweep a profile that intersects the axis around an
axis.

Hannes


[e-mail] wrote:
> 
> Hello,
> 
> I want to create complex bodies by using
> vetices->SPApostion->edge->coedge->wire -> ... sweep straight along axis
> -> BODY
> 
> I can create the wire by setting the first coedge (which points to the
> following coedge ...).
> 
> But how do I use the sweep-algorithm to create BODY-objects?
> 
> If anybody could send an example (even simple geometry like box, made by
> sweep) I would be(very) happy.
> 
> Also a link to an example would support myself.
> 
> Thank you for your support
> 
> Jeurgen Wehr
> _______________________________________________
> Acis-alliance mailing list
> [e-mail]
> http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance

-- 
    Hannes Kaufmann | Vienna University of Technology
         Researcher | Interactive Media Systems Group
T [phone] | F [phone]
           kaufmann @ ims.tuwien.ac.at
From alames at sandia.gov  Mon Mar  1 09:42:11 2004
From: alames at sandia.gov (Ames, Arlo)
Date: Mon Mar  1 11:42:57 2004
Subject: [Acis-alliance] [e-mail]
Message-ID: <[e-mail]>

The boolean operation is slower than creating a revolution, but creates the
minimal solid (i.e. 4 faces, 4 edges, 4 vertices).  Creating the revolution
will likely produce a seam, consisting of 4 edges, at the start of the
sweep.  It will result in a solid of 4 faces, 8 edges, 4 vertices. 

The absolute fastest code, with minimal topology, is to construct the hollow
cylinder directly, by creating the appropriate faces.  This involves calls
through the direct interface (DI), which tends to annoy Spatial, as it is
the strongest limiter of freedom for their developers; the api interface has
constant size, but any minor interior change to the DI has to wait until the
next rev of the software, to avoid breaking everybody's code.  DI calls are
also the most expensive to write code for, as you have to do all, and I mean
ALL, the math.  And you have to be aware of all the topological
consideration, including the requirement for PCURVEs on your COEDGEs when
you have a model involving NURBS surfaces or curves.

I am continually curious that Spatial hasn't expanded their api interface to
the point where nobody needs the DI, (since it worries them so), but it
seems to not have happened yet..

If speed isn't a major factor, or if constructing the models is a small part
of your overall compuation, by all means use the APIs, including
construction and booleans.  If you don't care about topology and speed is
something of an issue, the revolution is good.  If absolute speed is your
problem, the DI will give you maximally small entities in minimum compute
time.

Arlo Ames
[e-mail]


-----Original Message-----
From: [e-mail] [mailto:[e-mail]]
Sent: Monday, March 01, 2004 3:56 AM
To: [e-mail]
Subject: [Acis-alliance] [e-mail]



Hello,


I want to create a hollow cylinder (C++ API).

Is it more easy to use boolean operations by using 2 cylinders or should
I create a revolution of a face.

Did anybody get experiance by creating objects like that?

Thank you for your support

Juergen Wehr
_______________________________________________
Acis-alliance mailing list
[e-mail]
http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance

From alames at sandia.gov  Mon Mar  1 09:46:42 2004
From: alames at sandia.gov (Ames, Arlo)
Date: Mon Mar  1 11:52:17 2004
Subject: [Acis-alliance] create complex BODY
Message-ID: <[e-mail]>

The sweeping algorithms are somewhat limited.  If you're rotating about an
axis, you need to trim the curves you're sweeping to one side of the axis.
As I recall (I'm getting older...), the curves you sweep have to lie in a
plane that contains the axis about which you're sweeping, but that
limitation might have been removed.

I suspect the example you seek is to be found in the documentation on
api_sweep_with_options.  I'm unsure, as I generally do my own construction,
having gotten tired early on with the limitations of the sweeping
algorithms, and being a control freak about model topology...

Arlo

-----Original Message-----
From: Hannes Kaufmann [mailto:[e-mail]]
Sent: Monday, March 01, 2004 5:53 AM
To: [e-mail]
Cc: [e-mail]
Subject: Re: [Acis-alliance] create complex BODY


As far as I have seen it is not possible with ACIS to sweep whole bodys
around an axis. You can only sweep profile curves.
The sweeping does not seem to be very sophisticated. I don't even get a
valid result if I sweep a profile that intersects the axis around an
axis.

Hannes


[e-mail] wrote:
> 
> Hello,
> 
> I want to create complex bodies by using
> vetices->SPApostion->edge->coedge->wire -> ... sweep straight along axis
> -> BODY
> 
> I can create the wire by setting the first coedge (which points to the
> following coedge ...).
> 
> But how do I use the sweep-algorithm to create BODY-objects?
> 
> If anybody could send an example (even simple geometry like box, made by
> sweep) I would be(very) happy.
> 
> Also a link to an example would support myself.
> 
> Thank you for your support
> 
> Jeurgen Wehr
> _______________________________________________
> Acis-alliance mailing list
> [e-mail]
> http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance

-- 
    Hannes Kaufmann | Vienna University of Technology
         Researcher | Interactive Media Systems Group
T [phone] | F [phone]
           kaufmann @ ims.tuwien.ac.at
_______________________________________________
Acis-alliance mailing list
[e-mail]
http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance

From kaufmann at ims.tuwien.ac.at  Mon Mar  1 18:04:40 2004
From: kaufmann at ims.tuwien.ac.at (Hannes Kaufmann)
Date: Mon Mar  1 12:04:33 2004
Subject: [Acis-alliance] create complex BODY
References: <[e-mail]>
Message-ID: <[e-mail]>


> The sweeping algorithms are somewhat limited.  If you're rotating 
I fully agree.

about an
> axis, you need to trim the curves you're sweeping to one side of the axis.
> As I recall (I'm getting older...), the curves you sweep have to lie in a
> plane that contains the axis about which you're sweeping, but that
> limitation might have been removed.

Yes, that limitation has been removed. The curve can be a real 3d curve,
the curve can even "go around" the axis. But if the curve intersects the
axis, api_sweep_with_options does not return a result anymore.
Another "interesting" limitation: If the curve is a planar curve or
assume the points of a space curve are nearly in a plane and the axis of
revolution is nearly orhogonal to that plane - no result is returned by
the sweep option. I could understand it if plane and axis were really
normal to each other but api_sweep_... stops working as soon as they get
close to normal.

> I suspect the example you seek is to be found in the documentation on
> api_sweep_with_options.  I'm unsure, as I generally do my own construction,
> having gotten tired early on with the limitations of the sweeping
> algorithms, and being a control freak about model topology...

That sounds to be very wise :)

Hannes


> 
> Arlo
> 
> -----Original Message-----
> From: Hannes Kaufmann [mailto:[e-mail]]
> Sent: Monday, March 01, 2004 5:53 AM
> To: [e-mail]
> Cc: [e-mail]
> Subject: Re: [Acis-alliance] create complex BODY
> 
> As far as I have seen it is not possible with ACIS to sweep whole bodys
> around an axis. You can only sweep profile curves.
> The sweeping does not seem to be very sophisticated. I don't even get a
> valid result if I sweep a profile that intersects the axis around an
> axis.
> 
> Hannes
> 
> [e-mail] wrote:
> >
> > Hello,
> >
> > I want to create complex bodies by using
> > vetices->SPApostion->edge->coedge->wire -> ... sweep straight along axis
> > -> BODY
> >
> > I can create the wire by setting the first coedge (which points to the
> > following coedge ...).
> >
> > But how do I use the sweep-algorithm to create BODY-objects?
> >
> > If anybody could send an example (even simple geometry like box, made by
> > sweep) I would be(very) happy.
> >
> > Also a link to an example would support myself.
> >
> > Thank you for your support
> >
> > Jeurgen Wehr
> > _______________________________________________
> > Acis-alliance mailing list
> > [e-mail]
> > http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance
> 
> --
>     Hannes Kaufmann | Vienna University of Technology
>          Researcher | Interactive Media Systems Group
> T [phone] | F [phone]
>            kaufmann @ ims.tuwien.ac.at
> _______________________________________________
> Acis-alliance mailing list
> [e-mail]
> http://lists.cs.columbia.edu/mailman/listinfo/acis-alliance

-- 
    Hannes Kaufmann | Vienna University of Technology
         Researcher | Interactive Media Systems Group
T [phone] | F [phone]
           kaufmann @ ims.tuwien.ac.at
From alames at sandia.gov  Mon Mar  1 11:07:59 2004
From: alames at sandia.gov (Ames, Arlo)
Date: Mon Mar  1 13:08:46 2004
Subject: [Acis-alliance] create complex BODY
Message-ID: <[e-mail]>

You'd have to ask Spatial for the "real" answer as to why the
nearly-orthogonal case is disallowed, but I suspect it's about surface
numerics.  There are some really interesting problems in floating point
representation of geometric entities as "simple" as conical surfaces.  Try
looking at the numerics of surface-surface intersections for cones having a
half-angle of < .5 degrees or greater than 89.5 degrees.  In both cases 32
bit double precision just plain isn't enough to get anything like reasonable
results.  I met a man almost 2 decades ago who had spent *years* working on
approaches to good plane-cone intersections in the dark corners of steep and
shallow cones.  Such surfaces are one of my standard ways of embarrassing
CAD vendors when they're trying to make a sale.  You can always get a
modeller to crash, long before you get to anything as complicated as a
NURBS, with these cases.

So, maybe that's a good reason for *using* the api's.  They sometimes keep
us from shooting ourselves in the foot, when we fools rush into areas where
angels fear to tread.

-----Original Message-----
From: Hannes Kaufmann [mailto:[e-mail]]
Sent: Monday, March 01, 2004 10:05 AM
To: Ames, Arlo; ACIS List
Subject: Re: [Acis-alliance] create complex BODY



> The sweeping algorithms are somewhat limited.  If you're rotating 
I fully agree.

about an
> axis, you need to trim the curves you're sweeping to one side of the axis.
> As I recall (I'm getting older...), the curves you sweep have to lie in a
> plane that contains the axis about which you're sweeping, but that
> limitation might have been removed.