Subject:
|
Re: *** LDView v0.7 Released ***
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Mon, 4 Sep 2000 00:13:52 GMT
|
Viewed:
|
1060 times
|
| |
|
|
Travis Cobbs wrote...
> "Rui Martins" wrote in message
> > On Fri, 1 Sep 2000, Travis Cobbs wrote:
> >
> > Well, for usability sake, we should correct what we can, but this is a
> > dangerous feature, if we don't issue at least some warnings about it.
> > Because this will lead the authors to get sloppy (please don't take me
> > wrong here, its a Human factor) and even worse, if the autocorrecting
> > software is used to check parts for voting, then a lot of errors may get
> > by us, without beeing detected.
>
> Good point. I think I'll add an option in LDView to enable/disable
> warnings/errors during model loading. I currently just ignore all errors
> and problems.
I have also had second thoughts about the autocorrecting in L3Lab,
when it is used for part inspection[1]. L3P and L3Lab share
the input routines, which does several corrections.
Some corrections are imperative to have POV work at all!
L3Lab and L3P normally have the warning level set to 0, complaining
only about outright errors. However, "l3p -check" sets the warning
level to 1 showing all the autocorrected errors.
A full list of possible warnings and errors and their levels
can be found in l3p.txt (comes with l3p.exe)
> Due to the fact that certain problems are debatable, I may
> even do fine-grained warning enabling/disabling. In addition to the
> outright errors such as missing sub-parts and DAT-file errors, the following
> come to mind:
>
> 1. Concave quads
I don't think concave quads present any problems.
And then, IIRC LDraw renders concave quads as the bounding triangle!
> 2. Non-planar quads
> 3. Quads with points in criss-cross order (This happens fairly often in the
> parts library, and may not be considered a problem. This warning should be
> able to be disabled independently of the rest.)
> 4. Degenerate quads with two or three identical points (should have been a
> triangle or line)
> 5. Degenerate triangles with two identical points (Should have been a line.
> I haven't noticed any of these)
> 6. Degenerate quads, triangles or lines with all points identical (I haven't
> noticed any of these.)
L3P detects 2-6.
> I'm not sure #3 above is really a problem
Well, not if you deal with it! But LDraw importers/exporters/coverters/viewers
that don't, have a problem. And it make it harder (impossible) to BFC it.
> , but I do spend a fair number of
> CPU cycles detecting and correcting for it, since my solution for detecting
> the problem involved checking to see if any of the normals for a quad point
> in opposite directions. If anyone has a faster and/or easier way of
> detecting this, I'd welcome it. The minifig arms have a particularly large
> number of quads with criss-crossed point ordering.
You can see how I handle autocorrecting in L3P and L3Lab:
Get L3Input.cpp from ldglitesrc0_6pre7.zip at
http://www.geocities.com/SiliconValley/File/8624/
> I don't currently have any plans to turn LDView into a tool
> geared toward parts authors; quite frankly L3Lab does an excellent job of
> that already. On the other hand, the lighting model of LDView does make
> certain surface imperfections more obvious. If you have multiple triangles
> making up a facet which are almost--but not quite--co-planar, it can
> sometimes be more obvious in LDView than in the other viewers.
Right, your lighting model is dazzling.
/Lars
[1] Also remember to turn off View/BFC!
|
|
Message is in Reply To:
 | | Re: *** LDView v0.7 Released ***
|
| "Rui Martins" <Rui.Martins@link.pt> wrote in message news:Pine.GSU.4.10.1...0@is-sv... (...) Because (...) its a (...) check (...) detected. Good point. I think I'll add an option in LDView to enable/disable warnings/errors during model loading. I (...) (24 years ago, 1-Sep-00, to lugnet.cad.dev)
|
9 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
