Todo / Ideas / Concerns Status
Load slices, collaborative viewing, custom pointers Done
Dynamic insertion of slices via the network Done- currently demosend client sends slice#, frame# and image filename to cibrserver which distributes message to cibrview. cibrview downloads image from cibrserver. In future we need a separate download server and we need to bundle its address and port inside the frame,slice message.
Export of script file for caching Done
Generation of sample image Done
Startup busy wait icon Done
1 shared cutting plane Done 11/10/99
N shared cutting planes  
Add 1 higher level switch to allow for multiple view points  
Real Corridor1 data set for LBNL Cosmological strings 11/11/99
Generation of stereo versions of each image  
Shared manipulation of viz- rotate as a trackball  
Menu interface  
Audio streaming Keep as external program
Currently slices are loaded in as script is loaded. This may not be viable for lots of slices, frames and viewpoints. May need to load as you go.  
Posibility of inserting a sub section volume rendered with SGI hardware that user can move around within the slices- kind of like a microscope  
What is the advantage of ibr vs just using SGI rendering hardware?  
Are there operations on the slices themselves that need to be done? - various texture mapping filtering functions?

Also is there an interactive component that is necessary such as selection of thresholds to map colors, transparencies etc?
Use of SGI O2 compression hardware to compress images to reduce bandwidth requirements O2 hardware can compress arbitrarily large RGBA images into JPEG with varying compression ratios.

We may not gain anything if we have to ship the image to an O2 to compress it. Ideally we need the compression hardware on the machine generating the rendered images. But if the rendering is slower than the data distribution then it doesn't really matter too much except that you need to distribute the same image to multiple clients. Bandwidth can go way up if there are also animation frames involved.

May need to pick a good software streaming compression algorithm (RLE?)
Think about how this folds into TIDE architecture Notion of TI server that mediates the rendering server on behalf of the TI clients is still present in CIBR. Notion of client interface and server handler is still present as the client needs to submit new viewing angles or color map/thresholding parameters etc.

Contact the CIBR View Group