[mythtv-users] Why cant we make a STB?

[Mark Knecht]

On Wed, Feb 4, 2009 at 3:57 PM, Robert McNamara
<robert.mcnamara at gmail.com> wrote:
>>
>> Am i the only one who really wants this?
>>
>>
>> --
>> Mitchell
>
>
>
> Of course you're not the only one who *wants* it, but there are
> realities to consider-- Travis touched on a part of this.  *Someone*
> would need to foot the bill for the design and manufacture of a fixed
> number of these to begin.  I'll give the idea the benefit of the doubt
> and say we could find the EE and ME from within the Myth community.
> But then you'd need to get specs for the chips used in these devices,
> and consent to open source the interface with them.  And *that* is
> likely where this idea would fall apart, huge startup cost aside.
> Sage can do this easily because they are closed source and a paid
> product.  Myth is neither.
>
> Robert

I'm an EE. I worked in Silicon Valley for about 25 years before I
bailed 5 years ago. I did mostly chip design and managed chip design
groups but for about 2 years managed a group that also did board
design for the chips we developed.

My #1 input on this sort of thing is that most folks probably don't
have a clue about either the amount of time it takes or the costs of
doing hardware design - even board level design. It seems simpe and
straight forward until you really get into it. Chip specs are one
thing, but there's the cost of tools, the cost of manufacturing
boards, the cost of assembling boards, the costs of building systems
around those boards, the costs of testing, and then the costs of
revisions because a bunch of folks like us on the list aren't going to
do it perfectly the first time. It's expensive, takes a LONG time and
a lot of money, and when you get done and it's built and working the
market has moved on and our design is underpowered and overpriced.
Your market for selling these things is always smaller than you think
due to features in the design and compitition.

The best thing someone can do from a price POV is to use standard
parts to build the hardware and buy in quantity. Buy 100 motherboards,
100 video cards. Show some sales guy at NewEgg that you're serious
about making a product that he can sell you parts for once it's in
production. Configure Myth around that hardware and make it turn key
for the user. Even if the price is a bit higher support from a
community like this may make the difference in terms of actually
selling it. Sell it to everyone subscribed to the list for $300 or
$400. Get a reputation that it actually works. Bring in some money and
then, ONLY THEN, worry about the cheapest cost or doing something
custom.

Just my 2 cents. I've been there multiple times.

Cheers,
Mark