--- Stefan Steiniger <[EMAIL PROTECTED]> wrote:
> - Does it actually matter to have a 2D versus a 3D
> Triangulation? Or do
> you intend to triangulate in 2D and use 'z' as an
> attribute? (The
> simplex insert algortitm from Paul Chew should be
> 2D?, I am not sure
> about Pirol's Triangulation algorithm)
>
For the surfaces were doing in OpenJUMP, the DT
algorithms will be 2D. Three dimensional DT algorithms
produce connected tetrahedrons rather than triangles
and are primarily used for solid object modeling
rather than surface modeling.
> - When refering to H. Samet and Tin implementation
> you are talking about
> a table. Do you really mean a table or rather a
> list/collection (i.e an
> object oriented realization vs. a database
> realization)? (I don't have
> the book here)
>
In this case, a table is equivalent to an array (O(1)
lookup of elements given an index). In my
implementation, they will be represented as arrays.
> - on the intermediate TIN file format based on ITF:
> We need to get sure
> not to miss other standards. But all the extensions
> you mentioned make
> sense to me. Most important would be to have a very
> good documentation..
> so that others implement support for the format too.
> [btw.. it may be recommendable to ask Frank
> Warmerdam/the OSGEO list if
> he knows about something similar and/or send an
> email with your proposal
> to the java-collab list and ask for comments?]
>
Ill send those messages off today.
I only found vterrain and their binary TIN format
recently (due to the java-collab list actually). Im
using it more as inspiration rather than calling my
file format an extension or new version; I wouldnt
want to step on any toes. I think it is a slick format
that is simple, compact, and expressive, but not quite
what we need.
I intend to heavily document the file format both on
the OpenJUMP wiki as well as in the svn codebase.
> - is Java OpenGL for the visualization platform
> independent (i.e. would
> it run on MacOSX and Linux as well?)
>
It is the most platform independent Java,
hardware-accelerated, 3D library out there. Jogl is
just a wrapper around the OpenGL C interface which is
supported by not only Mac, Windows, Linux, and Solaris
operating systems, but also by pretty much all 3D
graphics cards made in the last decade.
The real question you should be asking is whether
hardware accelerated jogl painting to a JPanel is
present on all platforms. Up until recently, SWING
rendering wasnt hardware accelerated. Only with Java6
and the hardware accelerated 2D did this aspect of the
project become possible. This will currently work on
Windows and Linux/Solaris, but Im not sure about Mac.
I just recently downloaded the non-developer release
of Java6 for Intel OSX and havent had time to test
it, but unlike the pre-release, there is no caveat in
the readme about non-implementation of hardware
accelerated 2D. I imagine that by the time I get
around to doing that portion of the project everything
will be ironed out on all platforms. And, since Im
developing on the lagger of the bunch (OSX), by the
time I can make things work on my machine, everyone
else should be a ok.
As a side note, when doing the 3D visualization, I
intend to have a 2D/3D toggle available to the user so
that those without hardware acceleration or a older
java version can still run the code using the original
2D implementation. Im specifically thinking about
palm tops and future mobile versions of OpenJUMP.
> - on the Task-List: I agree with you that we need to
> accomplish Taks 1
> and 2, and Tasks 3 and 4 as far as possible. All
> other stuff 5-8 is from
> my perspective optional, i.e. nice to have. Note,
> task 1.2 (Write a USGS
> "bil" file reader) is for me optional too.
>
I agree that the USGS reader is functionally optional,
but I am interested in including it so that there is
data out there that this library can utilize. The SRTM
data covers the entire earth at 1 degree arc with
smaller datasets at 1/3 degree and 1/9 degree. And
its free over the internet. I think that this would
be an incredible functionality increase for a modest
expenditure of time. Im targeting the .bil format
because it is the only non-propriety data format that
this information is available in.
> and my request:
> - Can you send me the original word document?
> I would then add a few more comments to this doc,
> and also make some
> suggestions with respect to the layout (so I we can
> use it for future
> GSofC reports)
>
The original document is written in OpenOffice and
thus is in a .odt format. Will this be useful to you
or should I convert it to .rtf or .doc?
Thank you for your feedback,
--Christopher
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Jump-pilot-devel mailing list
Jump-pilot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel
