[mythtv-users] Re: excellent review of mythtv vs. MCE

[Stephen Tait]

At 16:36 16/09/2004, you wrote:
>Paul Hyui <phyui at yahoo.com> wrote:
> > at anandtech
>
> > http://anandtech.com/linux/showdoc.aspx?i=2190
> > http://anandtech.com/linux/showdoc.aspx?i=2208
>
>I found that "review" to be quite flawed.  Choosing to use SuSE the first time
>around was boneheaded for a review.  He should have gone with something more
>commonly supported and documented (Knoppmyth, Fedora Core 1, Debian to name a
>few).
>
>And then choosing not to use PVR-250 in the second run makes it comparing
>apples and oranges anyways.
>
>Some of his remarks were so off-base it's not funny.  Blaming Myth for him
>not reading the docs well enough is how it came across to me.  And he should
>have included the user support part of it.  Myth has the mailing list, IRC
>channel, and numerous wikis and forums (official or not).
>
>He also didn't state (that I saw) which version of MythTV he was using.
>
>There was no *quantatative* comparison (i.e. CPU usage using the same 
>hardware,
>etc) that I could see.
>
>On the whole, I found it totally disappointing.
>
>Ciao
>Gavin

Seconded. Most of the review seemed to be gushing about how incredible and 
blue the MCE interface was (as evidenced by the pictues)... and then the 
conclusion said that Myth's interface was friendlier. Make your mind up! 
They also completely neglected to mention that there are a number of 
companies selling Myth-based PVR solutions.

And yes, comparing the HW card under MCE with software encoding under Myth 
makes any quality comparisons null and void.

I think the only valid point that was made during the whole review was that 
MCE would cost you an arm and a leg was was not expandable, whereas Myth 
would allow you to spend as much or as little as you liked *and* be expandable.

I've been having issue with a fair few of Anand's reviews recently. Did 
anyone else witness the utterly laughable comparison between the AMD64 and 
the brand new Nocona chip? Among the more comedic errors were benchmarks 
that used Intel-specific assembly code and GCC optimisations passed *during 
the linking stage*. Beer was expelled through nostrils at that point...!