[AG-TECH] Development questions for vic on Windows

[Robert Olson]

>1)       Is the code on sourceforge really the latest?

It is the latest version of the UCL-based AG vic. I haven't incorporated 
the locking patch (would like to hear from more sites on its use, to see if 
it causes any problems).

>2)       Why does the vic I compiled work (well, crash really :-) 
>differently than the released AG vic.

The latest binary release of vic may be lagging the latest in the 
sourceforge CVS archive.


>3)       Should the Windows code (tcl, tk, vic) be compiled with the 
>single-thread or multi-thread run-time libraries (see Project Settings -> 
>C/C++ Tab -> Category="Code Generation"). Default seems to be single-threaded.

I've always built single-threaded; perhaps if VfW capture requires 
multithread support it should be multithreaded.

>4)       What is the status and direction of vic development with respect to:
>
>a.       Direct draw, display and capture.

There are some optimizations that could be made to the display side of 
things - support for hardware overlays for h261 (there is support in there 
now for jpeg); support for YUV visuals if they are available in non-overlay 
mode (RGB->YUV conversions take up a remarkable amount of CPU time); etc.

I'd love to see someone build a DirectShow-based capture module for vic.

>b.       AG vic vs. OpenMash vic vs. ucl vic vs. nlsde vic (am I missing 
>anyone?).

The OpenMash folks have put quite a bit of effort into making mash vic 
compatible with the Access Grid. Since they are actually doing maintenance 
& research work with vic, I'm leaning toward integrating mashvic as a 
supported tool. Merging the Admire fixes into UCL or Mash would be great as 
well.

>One of my main tasks is to improve the reliability of vic on SMP Windows 
>machines. I am looking forward to helping the AG effort, but I want to 
>make sure I'm starting with the right base and have the necessary 
>background info.

If you are going to be doing capture on Windows, I think moving to 
DirectShow for capture would likely lend some stability.

--bob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/ag-tech/attachments/20020529/a2585163/attachment-0001.htm>