[comp.graphics] an idea: common question summary sheets

glassner@unc.cs.unc.edu (Andrew S. Glassner) (03/12/88)

This newsgroup seems to periodically raise and discuss the same questions, 
generating (usually) no new and useful information, but lots and lots of 
postings - some giving algorithms, some trashing other people's algorithms, 
some giving opinions, some trashing other people's opinions, and so on, all 
in response to some question that has gone through the whole thing before.

I believe this will continue to occur as long as new people join the net 
and try to solve the same common problems each of us has to answer.  It CAN 
be a good idea to post questions to the net if your own investigations lead
to no solution; there are some very sharp people on here.  But too often 
people ask questions that can be found in basic texts or common journals,
and that's where the problems with a flooded newsgroup begin.

I suggest we write a set of "Summary Sheets for Common Questions".  We could 
draft these things until they were rather complete. Probably lots of us 
would like to have local copies for our own use to give to colleagues who 
come for such information.  And when someone on the net asks one of these 
questions, someone will post the standard response.  IF there is a significant 
new point to be raised that is not properly addressed in the standard
response, be assured that someone will pipe up.  But if the standard answer
covers the best answers and debunks common poor ones, net traffic
and misinformation might go 'way down.

I don't propose that these sheets should be tutorials.  They should be
SHORT: mostly pointers, formulae, and literature references,, with terse 
and consise discussion where appropriate.  The question poser will probably
still have to do some independent work (probably in the local library);
these things will show the path.

Here are my proposals for starters, based on the last few weeks.  We can
always write more standard sheets or update these when it seems appropriate.

  Color conversions: RGB<->HSV, RGB->B/W, RGB->Lum
  Image conversions: 24 bits RGB->8 bits colormapped, B/W->dither
  Geometry: Point in polygon
  Rendering: intro Z-buffer references, intro ray tracing references

Before tackling these I'd like to know if there's some reason this
procedure won't be a Good Thing.  If you feel strongly about this one
way or the other, please let me know.  If you'd like to volunteer to
help draft one of these documents, let me know that, too.

Then we can get down to the Real Stuff: like how to insert aliasing
artifacts into real photographs so our stuff doesn't look so much worse.

-Andrew

   - -- ---- ------- ------------ -------------------- ---------------------
Andrew Glassner    UUCP:decvax!mcnc!unc!glassner    ARPA:glassner@cs.unc.edu

allan@calgary.UUCP (Jeff Allan) (03/16/88)

In article <1677@thorin.cs.unc.edu>, glassner@unc.cs.unc.edu (Andrew S. Glassner) writes:
> This newsgroup seems to periodically raise and discuss the same questions, 
> generating (usually) no new and useful information, but lots and lots of 
> postings - ...
> I suggest we write a set of "Summary Sheets for Common Questions".  We could 
> draft these things until they were rather complete.  ...
> Here are my proposals for starters, based on the last few weeks.  We can
> always write more standard sheets or update these when it seems appropriate.
> 
>   Color conversions: RGB<->HSV, RGB->B/W, RGB->Lum
>   Image conversions: 24 bits RGB->8 bits colormapped, B/W->dither
>   Geometry: Point in polygon
>   Rendering: intro Z-buffer references, intro ray tracing references


I think this is a wonderful idea. Is there some way that the net controllers
could make these summaries available or would that be up to some concerned
person like yourself? I'm not much of a participant in comp.graphics but
I like to read it and I do get tired of the same questions. I find myself
skipping over follow-ups but wondering if I am missing something good.

Another area where net traffic could be reduced is in requests and discussions
on particular computers (esp micros) and systems. Maybe this is what the 
group is intended for, I don't know, but it seems to me that this stuff would
be better posted under the groups set up for these specific machines.

I'd like to offer my assistance with the summaries but I am in the writing phase
of my M.Sc. right now. Besides my expertise would only cover the geometry 
question you mentioned and I think the answer to that one is pretty 
easy to write up.
-- 
        Jeff Allan,  University of Calgary
	..!{ubc-vision,ihnp4}!alberta!calgary!allan