[roseman at dimension4.math.uiowa.edu: Re: alpha and rib files on the NeXT.]

[munzner]

> Alpha information is ignored if transparency is not enabled. By
> default, transparency is off. You want a command like this:

  (merge-ap g1 appearance { +transparent material {alpha .2}})


Thanks. 


Unfortunately it still doesn't seem to work exactly.  


I tried the above command. 


In my geom I have two transparent objects one opaque.
In the rib file, the opaque one comes out as expected:

.............

Polygon
"P" [0.015025 -0.189848 -0.317004   0.000618 -0.182992 -0.317753    
-0.016667 -0.180484 -0.288611] 

"Cs" [0.685996 1 0   0.682745 1 0   0.508653 1 0] 

"N" [0.056514 -0.273463 0.960221   0.056514 -0.273463 0.960221    
0.056514 -0.273463 0.960221] 

Polygon
"P" [0.134974 0.606116 -0.162452   0.219549 0.57289 -0.184977    
0.187373 0.593787 -0.124963] 

"Cs" [0.115866 0.992916 0.1   0.115307 0.993165 0.1   0.112673  
0.994341 0.1] 

"N" [-0.282604 0.027059 -0.958855   -0.282604 0.027059 -0.958855    
-0.282604 0.027059 -0.958855] 


ETC ETC ETC

and the others look like:

........

Polygon
"P" [0.62261 -0.154888 -0.149333   0.6459 -0.137943 -0.048951    
0.702518 -0.120302 -0.16982] 

"Cs" [0.860628 0.660371 0.1   0.944241 0.591616 0.110648   0.914655  
0.545838 0.130638] 

"Os" [1 1 1   1 1 1   1 1 1] 

"N" [-0.176927 0.539333 0.823296   -0.176927 0.539333 0.823296    
-0.176927 0.539333 0.823296] 

Polygon
"P" [0.746979 -0.098987 -0.078997   0.702518 -0.120302 -0.16982    
0.6459 -0.137943 -0.048951] 

"Cs" [0.866754 0.471721 0.163004   0.914655 0.545838 0.130638    
0.944241 0.591616 0.110648] 

"Os" [1 1 1   1 1 1   1 1 1] 


ETC ETC ETC

So the "good news" is that we have transparency, the "bad news is  
that it is and the completely opaque kind. 


.."close but no cigar"

&&&&&&&&&&&&&&&&&

But at least, now, I can simply alter this with a simple  
search&replace since all my semi-transparent ploygons-to-be are  
triangles.

That's the "good news".

The "bad news" is that, for some reason, when I did this, it came out  
wrong.  This is the files trans13.rib and trans13.tiff in /u/roseman.   
Some of the polyons ended up with the wrong alpha.

All I did is to use the commands as suggested on the two geoms, save  
the file to .rib format and do a find/replace (using Edit), 

replacing 

"Os" [1 1 1   1 1 1   1 1 1] 

with 

"Os" [.2 .2 .2   .2 .2 .2   .2 .2 .2] .


For further information, (to get an idea what it should look like) I  
put the original stuff in /u/roseman/allThree13s.4off.

This has NO alpha set here--by idea is to make  
onetwfig8m013.4off.slicey13 and onetwfig8m013.4off.slicey13b semi  
transparent and the onetwfig8m013.4off.slicey13b left opaque.  This  
should be viewed with start-up file /u/roseman/CopyOf.geomview

In the .tiff, I had turned the camera around to get a better view.

Note: there is another problem I have mentioned to Stewart about  
normal vectors in ndviewer--I don't know if this is somehow connected  
with this problem.

I am trying to get together stuff to finish a video next week.  In  
the grand scheme of things, I guess I can live without this example  
presented in this way.  Without being able to see it well, I'm not  
even sure how effective it would be.  My visual guess is that in  
order to look right one would need to make a rotating animation of  
it--but with something that Stewart wrote, I would only need one  
working .rib file.

&&&&&&&&&&&&&&&&&

I might point out, in the spirit of a (possibly) helpful suggestion,  
that (as far as I can tell) "+transparent" is not documented in the  
manual.  In a way I see why--the SGI pannel has a button for this  
which was left off of the NeXT pannel.  


It is confusing, at least to me, as you see, since the information is  
NOT ignored in geomview files but IS when .rib files are made.  I can  
see a reason for this--make the .rib files a more reasonable size.

My suggestion is that the documentation (somehow) clarify this.