MARCH2006 Model Challenge

Dale,

One unexpected result is that several have submitted more than one
solution. I have my own solutions that are different (and not as good
as) some of you. Since some are having fun with two or three
solutions, why not allow more than one? Just take time for your family
and all that. The number of solutions so far are certainly adding to
the art and science of modeling.

Now as for the vary the edge part I had the thought that one property
of the truncated icosahedron is that all the edges are equal in length.
But they don't have to be. Granted it wouldn't be a proper Archimedian
solid, but it would be real world. One of the things you find in the
literature is a connectivity graph between vertices and edges. Imagine
that the restriction of having equal length edges were lifted, but not
the connectivity requirement. Let's say that next month's challenge
might be creating a solid of this type (though not necessarily this
one) with the abillity to vary individual edge lengths by say +/-10%.
I've had some thoughts on how to get there, but I'll keep them to
myself.

 

Sounds like I have to update my model. I think I can do it with one

I ended up with the same situation, but when I read this I decided to remove
the last circular pattern. So my model has now one sketch, revolve, cut and
circular pattern and no other features such as reference geometry.

-h-

 

Sounds like I have to update my model. I think I can do it with one

It seems, though, that my second contribution with 5 features is more than
seven times faster to rebuild than my last effort which has only three
features.

-h-

 

Heikki's got me beat. I think I see now what he did, and it's definitely

Well actually I just realized that a revolve feature has a temporary axis,
and the centerline for revolve feature doesn't need to be vertical...

To be honest, I would never have believed that this can be modeled with so
few features.

-h-

 

Interesting. That's not what I thought you did. I resubmitted a model

Sharing sketches is a cool way to reduce the total amount of features, but
as a result I had to use contour selecting. By using multiple sketches I was
able to cut the rebuild time from 3.52s to 0.70s. However, I was not able to
beat my second, more complex model, which rebuilds in 0.38s, so I did not
bother to send my last version.

Of course the rebuild times can be compared only on same computer.

-h-

 

I can't quite understand the factors of rebuild time. Earlier today I got
0.7s rebuild time but now I can't get under 0.9s with the very same model.
(Somehow this made me think that maybe I should get a life instead of
hunting for 0.2 seconds)

I have Dell Precision M60, 1,7GHz, 1GB RAM. Interesting detail is that
unplugging the power source drops the cpu to 560MHz, but it does not affect
the rebuild time.

-h-

 

The official timing machine is an XP 3000+ on an ASUS mobo with 1GB
ram.
NVidia 500XGL graphics card. This is slow enough that small differences
should be noted.

TSToolbox will be used for timing.

 

Just wanted to get this thread back to the top of the list.
TOP, have you gotten any more entries? Could you post them somewhere
so we can see how the others did it? I'm curious how some did it with
so few features.

Timelord

 

I'm getting ready to post them. I had the NISWUG meeting to get ready.
That is over with and now we can put the models up on their web site.
This is next on my list of things to do.

I must say there are just about as many solutions as there are
entrants.

 

Are these posted somewhere? I'd really like to see how everyone
modeled this thing.

 

Just waiting for stragglers.

 

Come on TOP, mid April already! At least post the models somewhere so
we can see how others did it, even if you cannot pick a winner.

Timelord

 

Yes mother.

 

I have gone through the first pass of evaluations. Quite interesting.
The most complex model won the rebuild speed award while the very
simple 2 sketch model was second. I have class tommorrow evening so
I'll try to get them up on the web site Friday. Sorry about the wait.


Could I have the respondents send me information on their timezones.
For example I am UTC - 5 hours. I think Ed was -6 and Heikki +2. Not so
sure about Dale and Kim.

 

Could I have the respondents send me information on their timezones.

+2 confirmed.

-h-

 

Who got done first. Heikki has a 7 hour penalty to be adjusted for. If
his model says he finished at 9am his time he really finished 2am your
time. It's part of the scoring.

 

I'm in Albuquerque, NM
Mountain Standard Time

Timelord (Kim)