Subject:
|
Re: Latest BFC Spec?
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Sat, 22 Mar 2003 00:26:30 GMT
|
Viewed:
|
2455 times
|
| |
|
|
In lugnet.cad.dev, Kyle McDonald wrote:
> I saved this thread for later reading, and I've got a few questions
> now. Sorry if I'm resurrecting something everyone thought was dead :D
I thought it was dead 2 (or 3?) years ago. ;)
> Once the parts library is fully BFC certified, will there ever
> be a need to use CLIP or NOCLIP?
That's a pretty big 'once' you've got there.
One problem I've had, revisiting old part files and applying BFC -- once
you look at a file, you're also tempted to fix any/all errors you find in
the file. And if you don't fix them, there's a good chance the update will
get shot down in the Parts Tracker. I'm not saying that any of this is
good or bad -- my point is that making the entire library BFC-compliant is
not going to happen soon. Especially since we still get a large number of
new parts and files that are not BFC-compliant.
To get back to your question: I can see two cases for not culling surfaces
that BFC tells us should be dropped. Basically, the deal is that
transparent surfaces/materials reveal polygons that BFC says are hidden.
If your favorite rendering engine renders transparents so that multiples
layers of transparent surfaces look different than single layers of
transparents, then it should *not* cull backward-facing transparent
polygons. However, this test needs to be performed in the renderer, *not*
by using BFC CLIP/NOCLIP in the code. This is because the part author
can't know the color of the surfaces she's modeling, and because she has to
allow for variances between rendering programs.
Keeping transparency in mind, NOCLIP would be useful whenever we're coding
surfaces that should be visible when viewed from behind -- that is, when we
can see the surface on the back side of a transparent solid. The only
example of that I can think of is printed decorations.
IMO, *all* patterned parts should wrap NOCLIP / CLIP around patterned
surfaces, so if someone chooses to model a patterned part with a
transparent color (Star Wars canopies are a real-world example), the
patterns can be seen through the backside of the part.
> Is there a line in the header already that declares that file is
> a 'Part' or a 'Primitive' ??
Currently, it's the 0 LDRAW_ORG meta-statement that appears in the header.
Besides the release/version number, this line also indicates whether the
file is in the parts library or not (and what type of file it is -- part,
shortcut, alias, primitive, subpart). Previously, we had the 0 Official
line, which provided most of the same information.
Files which are *not* in the official parts library often have a
similar 0 Unofficial meta-statement.
Steve
|
|
Message has 1 Reply:
 | | Re: Latest BFC Spec?
|
| In lugnet.cad.dev, Steve Bliss wrote: [snippity-do-dah] Sorry, I found an old post still waiting to be authorized. I hit "Post" instead of "Delete". Steve (19 years ago, 20-Apr-05, to lugnet.cad.dev)
|
Message is in Reply To:
| | Re: Latest BFC Spec?
|
| I saved this thread for later reading, and I've got a few questions now. Sorry if I'm resurrecting something everyone thought was dead :D (...) Once the parts library is fully BFC certified, will there ever be a need to use CLIP or NOCLIP? I'm not (...) (21 years ago, 20-Mar-03, to lugnet.cad.dev)
|
38 Messages in This Thread:
 
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
