[cad-linux-dev] Re: spec
- From: Eric Wilhelm <ewilhelm@xxxxxxxxxxxxx>
- To: cad-linux-dev@xxxxxxxxxxxxx
- Date: Wed, 25 Aug 2004 16:05:09 -0500
# The following was supposedly scribed by
# Chris Somerlot
# on Wednesday 25 August 2004 02:18 pm:
>Yes, actually this brings up a whole other entity that I'm missing:
>coordinate systems.
Right, some sort of header entity?
>In acad circles are always drawn in the plane of a
>particular UCS.
Actually, I'm pretty sure that everything is stored according to OCS and that
there's no relationship to UCS saved in the file.
I'll rant for just a second about Object Coordinate Systems below.
Object Coordinate System as used by acad is absolutely stupid. In OCS, a
circle centered at the global point [5,3,7] will have a completely different
value for that point in OCS (which is a plane that shares the global origin.)
I can't see why they have done it this way, other than that it is some kind of
display optimization. If it makes sense to anyone, please educate me.
>I'm wondering if I should just give a vector attribute
>to circles and arcs,
Yes. Other entities will also need this (text, polylines), but note that this
is just a "normal vector".
>or give every geometric entity a coordsys
>attribute
NO!
>have the coordinate system defined as a separate entity,
>like layers.
yes, the UCS should be saved, but I don't think they should be associated with
entities in rhizopod (e.g. all entities live in one global coordinate space.)
--
"These crispix get soggy so quickly."
-- Tina Connolly
Other related posts:
