[mythtv-users] HD-PVR2 support (please vote)

[Devin Heitmueller]

> I see it for sale on Amazon - clearly the hardware isn't in beta.

The references John made to "beta" refer to Hauppauge's Linux driver
for the HDPVR2, not the HDPVR 2 itself.

> I take it they don't just publish the specs on how it works?  I guess
> reverse-engineering would be a pain - I'm not sure how obfuscated
> existing driver software is.

Considerably more complicated than the original HDPVR was.  There's a
good reason that you haven't seen another HD encoder supported under
Linux in the last six years.

> I'm not really super-excited about the idea of having binary blobs in
> mythtv.  At the very least it should be easy to strip them out when
> packaging it in a free-software-only config.  They should also be
> issued under a redistributable license unless they're separated from
> the rest of the sources so that we don't have to stop mirroring
> mythtv.

The support would not ship by default with MythTV.  While all the
MythTV specific parts (i.e. improvements to recording frameworks and
other bug fixes in MythTV itself) would be part of the distribution
and GPL, the driver itself would not ship with MythTV.  It would be
downloaded/installed from Kernel Labs's website.

> Right now I have MythTV listed as licensed under GPL-2.  That seems
> pretty simple to me.  Why complicate things?
>
> None of this is a slight on Kernel Labs.  I'm sure they do good work.
> It just sounds like it wouldn't be FOSS.

Let's be clear:  the support for the HDPVR 2 would *not* be FOSS.
While I'm sure everybody would love to have an FOSS solution where
they would get all the source code and wouldn't have to pay any actual
money, it's pretty clear that nobody is willing to create such a
driver (based on the many years where nobody has done so yet and we've
all been stuck with the HDPVR1).

> Would they be willing to accept payment to put the code in the
> mainline kernel instead?  I'm sure they'd want more money for the
> source, and that seems fair enough assuming people are willing to chip
> in.

While I'm happy to privately discuss other options (and I'm not
looking to rule out any possibilities), the costs would likely be
*much* higher, and it's not clear that we even have enough interest to
warrant making the driver available in binary-only form at a cost
reasonable to individual MythTV users.

-- 
Devin J. Heitmueller - Kernel Labs
http://www.kernellabs.com