Hey guys,

In the event that 3 of our 15 summer of code slots don't work out, we 
will need a mentor for the Geotools #4 choice:

3-Dimensional Rendering Pipeline Component for GIS Servers

This is a duplication as we will already have accepted a 3D proposal 
(unless it is one of the 3 rejected), but the other projects are out of 
good applications and two competing 3D implementations wouldn't be so 
bad.  Apparently the second 3D proposal requires little supervision, but 
a official mentor is still required (keep in mind that the mentor can 
delegate to others -- the official mentor is just the point man).

We need to resolve this TODAY.  If anyone is willing to mentor, please 
chime in; Jody is the current mentor, but we should stick to having one 
student per mentor.

As a mentor, you would be responsible for staying in contact with the 
student a couple times during the week, keeping him on track, getting 
him the help he needs, and finally evaluating his performance at the end.

Thanks,
Cory.

----

Full text:

3-Dimensional Rendering Pipeline Component for GIS Servers

Synopsis
   The proposed code is to implement an interactive 3-D viewer for the
massive terrain-based data sets provided by a GIS server. This will be
substantially more interesting than a standard real-time 3-D viewer
(e.g. as found in a video game), because the data needs to be streamed
from a remote source, and because the data is typically massive.

Benefits
   Data provided by a GIS server is most naturally viewed in three
dimensions. Existing 2-D viewers can only present a limited amount of
information to the user, and can be awkward to use. A 3-D viewer
component, written in Java (with JOGL) as a client to a GIS server,
will allow full access to dense geographical data, allowing a number
of new possibilities: from simple walkthrough- and flythrough-type
applications, to more specific visualizations such as for route
planning, statistical analysis, or network modeling.

Deliverables
   The code is to be implemented as a re-usable component written in
Java, accessing the native environment's OpenGL driver using JOGL, and
likely embeddable into a web application (with Java Web Start), or
into uDig directly. The component should be as generic as possible, so
as to allow a high degree of reuse (e.g. across different types of
layers provided by the GIS server, or different specific viewing
applications desired by the user), while still taking care to do all
of the necessary client/server communication and visibility
computation. This will be a typical split betwen policy and mechanism;
the latter will be implemented, and the reusability means that the
former can be easily changed. But a complete, simple application
(implementing a typical flythrough application, for simple exploration
of a 3-D environment), will also be implemented.

Project Details
   The client viewer will need to be very careful about how it fetches
data from a GIS server. Advanced visibility algorithms will need to be
applied; the author's experience indicates that this will be a 
split-server/client implementation, in which the client requests the 
minimum amount of geometry to render a view from its local position, and 
the server is able to provide at least a very good estimate of the 
correct amount on demand (perhaps with some additional culling by the 
client). Because the author does not have experience with GIS 
specifically, some background research will be necessary. For example, 
augmenting the server's data set with a visibility graph may work, in 
which case a fair amount of pre-computation will be necessary; but if 
GIS primitives alone suffice, then likely only a client needs to be 
written. These details will be worked out in the initial stages of the 
project, by the author.

Project Schedule
   Background research will take place between now until April 11
(when publication of application acceptance occurs). After this time,
the author expects to communicate more directly with the project
community and possible mentor(s). Coding is to follow the standard
GSoC timeline, beginning on May 28 and ending around August 20, with
regular progress reports and evaluations.

Bio
   Though the author has not yet used GIS, he has experience with a
very similar client-server application, as part of the Building Model
Generation project sponsored by MIT's Computer Graphics Group. In this
project, the author built programs to analyze floorplans of MIT's
campus buildings and extract detailed 2-D and 3-D geometry and
topology information. These data were then stored in a central
(non-GIS) database, accessed by a client viewer (also written by the
author as a Java-based web application) to intelligently fetch
geometry on demand in response to the user's movement commands. The 
geometry was stored in a visibility graph to make this a reasonably easy 
task for the client; the author expects to need to do some research 
regarding how to
implement a similar technique with a GIS server.
   More information about the author's professional background can be
read at http://web.mit.edu/mic/Public/resume.pdf.

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Geotools-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to