Thanks for your quick response.  As you know, we are new to this
community, and it's often not easy to find your way around right away.

> Some cameras also have bugs that can't be easily worked around. For instance,
> I've received reports of webcams that had a bad auto-exposure algorithm that
> got stuck in very low-exposure mode, making the image extremely dark. There's
> not much the host software can do about that.
>
> Fixing application bugs and improving multimedia integration doesn't usually
> require access to every webcam model. What is needed there is a few developers
> with free time and access to a few webcams. Providing hardware samples can
> definitely help.

You articulated well a number of the benefits we'll get from
additional outside testers:  Workarounds, troubleshooting, feedback.
It's a very inexpensive way to get feedback IMHO.  I'm always
surprised that more manufacturers don't do it.

>
> Obviously, bugs in proprietary applications (such as Skype) can not be fixed
> by open-source developers. The end-user might feel that a particular webcam
> model is "crap because it doesn't work with Skype", but little can be done
> about that if Skype is at fault.

Sure, I'm not expecting miracles, and I would naturally expect to work
with the skype folks and have some samples in their hands as well.  Of
course, ultimately we, as the manfuacturers have the responsibility to
make sure our devices work well, but having outside help and feedback
can only help.

>
> How many different webcam models do you plan to sell ? Do you plan to perform
> some basic testing yourself and provide hardware samples to developers when
> problems are detected, or would you prefer to donate hardware samples for all
> models to someone who would test them all ?

I would guess that we'll see about 3 models.  My personal view is that
getting samples to developers early is always better.  My main job is
at neuros (neurostechnology.com) and we have a history of releasing
early "gamma" or developer kit samples to the market even before full
production.  We might even try to do some of that here.  The biggest
issue for us is finding the legitimately involved users, who will at
least poke around with the device.  It's certainly not the standard
that a free unit obliges a developer to fix bugs (a free unit is
hardly fair compensation for doing bug fixing), but its always sad to
send out free units and get no feedback whatsoever.  To me, if you get
a developer to use the unit and provide feedback then the free unit
was well invested.

To be honest, the best way I have found to "self select" the right
devs/users is to sell the units at discounted rates rather than give
them away.  But in either case, if you can point us to the right
folks, we can certainly seed the community with a few samples.

Joe
_______________________________________________
Linux-uvc-devel mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/linux-uvc-devel

Reply via email to