[mythtv-users] Sell mythtv "set-top" boxes

[Ray Olszewski]

At 07:48 AM 6/3/2003 -0500, Pete Hartman wrote:
[...]
>my main point was not that the existing database is bad, but that it is 
>not the same thing as a particular "recommended" configuration that is 
>documented in detail.  If that configuration is documented in its entirety 
>(mb, memory config, cpu, etc.--all the usual system variables, as well as 
>some outline of setup steps that is more detailed than the main docs, 
>including the steps to set up things that myth relies on but rightly 
>aren't really part of myth) it will be a good guideline for those who are 
>less interested in the science project aspects of setting this up and are 
>looking for something that they can just put in place to be an OSS 
>PVR.  People could then choose how much they wanted to deviate from that 
>config based on how much they wanted to wander off in their own directions.

How specific do you really want to be? Your starting list is "mb, memory 
config, cpu, etc.". These strike me as the least useful things to specify 
specifically ... especially the motherboard. (I'm reminded of old Consumer 
Reports product-review articles, in which the top-rated choice had a very 
exact model number, a model that in practice proved to be unfindable ... 
leaving the reader unsure how similar other models from the same 
manufacturer were. Mobos come and go way too fast for a recommended 
configuration to suggest a particular one.) Much better, I think, would be 
to specify minimum CPU and memory requirements for a standard 1-tuner 
combined system ... but I think the HowTo already does that.

Specifying minimums (rather then specifically "recommended" hardware) also 
helps because price/performance tradeoffs are central to decisions. I could 
imagine the recommended configuration being finalized by somebody with a 
"price is no object attitude", so calling for parts that priced Myth out of 
most users' range.

In contrast, where people really need very specific hardware advice is in 
these areas:

1. Tuner-card selection. This is an ongoing effort, not something someone 
can do once and put on the shelf. The quick example of why is the recent, 
slipstreamed change in the Hauppauge WinTV Go vidcap card. I'd recommend 
this card with enthusiasm, but today's users would be sandbagged by the 
fact that the card that is on the shelves today is not the one I use (And I 
don't even know how to tell the two variants apart without opening the box 
and examining the chipset on the card itself). Another example is these 
ivtv cards ... they are bleeding edge right now, but I'd guess that in 3 
months (6 at the outside) they will be very conventional.

2. Video card selection, especially video for TV-out. The only cards I 
would personally recommend unequivocally are the nVidias that are supported 
by the proprietary nvidia X driver. I haven't been able to make any other 
card work to my satisfaction (though there are a couple of candidates I 
have not tried).

Whether we could come to a consensus even on recommended hardware for these 
2 uses is unclear to me. A couple of months ago, I drafted some 
minimum-spec guidelines, and the response they got suggests to me that we'd 
have trouble agreeing about anything involving hardware. (That's the 
problem a database approach avoids.)

Marc's database does offer some help, and the suggestions Ben and others 
are making would, if impplemented, improve it. But I think it will always 
have two limitations:

1. It is not specific to MythTV. Marc says this is by design, and I respect 
that, but it still limits the usefulness of the datbase to prospective 
MythTV users.

2. It is structured. A structure encourages respondents to answer the 
questions that are asked, while omitting information that is not 
specifically requested. (Yes, I know there is a comments field; I'm 
predicting that it will not collect the unstructured information beginners 
need to supplement the specific questions.)

 From time to time, we see on this list postings of particular 
configurations, along with discussions of what the problems were and what 
special considerations were involved. That, I suspect, is what beginners 
really need ... well-written descriptions of particular approaches that 
succeeded. Several of these would, in the end, help newcomers more than a 
generic "recommended" configuration ... which in practice will be 
recommended by nobody in particular and runs the risk of being a 
portmanteau configuration incorporating components that no one has actually 
used in combination.