Re: Mini-ITX PC's a the future of HA

["Robert Green" <ROBERT_GREEN1963@xxxxxxxxx>]

Dean Roddey and Robert Green discussed the future of HA:

<stuff snipped>

RG >> How often would you have to access the HA/S machine
>> compared to the entertainment server?  That differential in
>> access, as well as the difference in overall function seem
>> to cry out to me for separate boxes.  That's saying something,
>> at least for me, considering my goal is to *reduce* the number
>> of PC's that are running in my house.
>
> You are probably never going to 'do security' on the PC anyway. As a
> practical matter, the core security features are likely always going to be
> on some other piece of hardware in order to get it blessed by insurance
> companies so that you can get your discount on your home owner's insurance
> and all that. The only part of security that happens on the PC is the
> interfacing to the security system, and presentation of the security
status
> and such on touch screens, and the ability to arm/disarm.

Yes! YES!  I totally agree that the devil is in the details of how a
"blessed" alarm panel communicates with a HA system.  That's the part that I
contend needs to get a lot smarter to move into the next generation of
usability.  Right now, you're correct, all that's happening on the PC is
simple I/O.  But I want something smarter, and those smarts *aren't* going
to be found on a panel.

My HA/S system should be able to use video processing (that stays in my
house, incidentally, otherwise there's no SAF) to determine if *I* tripped
the alarm accidentally before the police are called.  The insurance and
alarm industry, along with all of the standards-making bodies that are
involved in the alarm installation process, act as subtle brakes on progress
towards a new way of doing things.

Current alarm systems are stupid.  They will false from thunderstorms and
all sorts of other causes that just a little bit of machine IQ could have
determined were not legitimate activations.  Set a variable called
"FalseFactor" and input TOD, weather conditions, previous conditions, etc.
I want a sensor on my garbage can that tells the system when move this can
the night before trash day, a false alarm is likely and to increase that
FalseFactor so that the system possibly waits for a second and third event
to determine if its a break-in.

I want a system that knows that the power's "blinked" three times in the
last 5 minutes, the rain is falling hard, the sky has darkened, there is
nearby and strong lightning and thunder and that glass break sensor that
just tripped may have only sensed vibrations from a nearby thunderstrike and
not a window being smashed with a crowbar wrapped in a towel.

Can Charmed Quark help me take out the trash and yet not wake up my wife?
She can sleep through the bag-dragging and door-opening but the alarm
warning tone gets her up every time, as it should.  Yes, I know I am an
idiot for not disarming and rearming the panel, so I need an idiot proof
HA/S system.  I haven't found one.

Worse, still, some of the commercial installers I've dealt with equate a
desire for "Fewer False Alarms" with a request for making the system less
sensitive and thus less secure.  What I want is a house that's at least as
smart as my dog.  She can distinguish between the UPS guy and a Jehovah's
witness.  Don't ask me how, but she waits for the UPS guy by the front LR
window but the JW's are waited for behind the front door with a barely
audible growl.  She won't start barking until they touch the door or make
some sort of loud noise.  Then it's Katy bar the door!

(Yes, I have thought of putting a GPS and a wireless microphone on the dog
but she generally eats harnesses and attachments. :-)  We had a Widgetcam
for a couple of days until we weren't watching.)

> But in order to meet the necessary requirements, there will
> probably always be a dedicated keypad of some sort for the
> security system that you can use as a fallback in the worst case
> scenario.

Yes - But for reasons that have little to do with technology, just custom.
I believe when the technological advantages of PC-based systems outweigh
those available to the traditional systems, there will be a sea change.  I
believe it's happening now because the form factor, reliability and
processing power have converged to a size and price point that makes
economic sense.  Certainly there were other options: expansion doohickeys,
extenders and other kludges.  But there really wasn't anything (other than
proprietary laptops) that could match the size, CPU horsepower, efficiency
and plain overall elegance of the mini-ITX.

> So I think that this is a bit of a strawman argument.

No, it's unfortunate that it might be cast that way but for me it's about a
decision fork I reached.  I've got an OMNI LT, but I can't really see going
to something bigger in that line or any other panel line.  What I can see is
using the LT as the core security panel.  I'll just use it to communicate
with a central station and dial a call list because that's what it's good
at.

But no Omni or Elk panel I've read about can provide the smarts I need to
take HA/S to the next level.  I can only get that from a form factor like
the mini-ITX.  If enough USB or Ethernet based control peripherals appear at
a reasonable cost, it will be even better and I am sure those devices are in
the pipeline.

Right now, I am using my Via with a Leadtek video card inside but I am not
happy with that option because the card's just a little too big for the
space and all the cables.  The Via, at least when mounted in a Casetronic's
case, uses a slot "bender" to mount the card horizontally.  It's already
worked loosed twice, confirming my overall suspicion of slot extenders and
expansion chassis.  I will probably switch to an external USB video unit
just to keep the case clear of extendo-trash but I would prefer integrated
solution from Via that offers as much as the Leadtek card does.

I don't think it will be long until Via incorporates on-board video
processing with a cable tuner.  Since they work closely with three large
chipmakers - Intel, IBM and AMD - they've seen how each company implements
its technology.  That interoperation with the "big three" gives them
valuable insights and viewpoints.

More importantly, because they worked with the big boys they are more immune
to the NIH syndrome.  In what I see as a replay of the American auto
industry, while the big three concentrated on speed, speed and more speed,
Via was working on cheap and efficient.  Sounds like Via's the Japanese auto
industry in that equation.  It's why I think that Via's the one to make the
big leap.  They're taking their swing last, after getting to watch all the
others.  That gives them an advantage.

<stuff snipped>

> But music is so light weight an operation for a modern machine (meaning
> either playback of a few streams locally on that machine via a
multi-channel
> card, or streaming of that data off the disc to other players around the
> house), that it's really not a practical problem. We are working with the
> Zoom Player guys to create a service based multi-player program so that
the
> playback all happens in the background without anyone logged on, and each
of
> the internal players in the app an be assigned to a different audio card
> output.

That sounds neat.  A great use for an older box fast enough to serve music
but perhaps not up to video speeds.  I assume that part of your reasoning
for doing it this way is to reduce the need for someone to be a network
engineer to deploy a client/server based program like Charmed Quark.

 >> When small touch screen LCD's become a commodity item,
>> I will be sorely tempted to move to a system like yours.
>
> You mean wireless or wired? For something like a home theater, you can
hard
> mount something like a 7" Lilliput touch screen between the chair arms (or
> on a table beside the couch) on a swivel mount for a couple hundred bucks.
> If you mean wired, then the new UMPC boxes are probably the best bet
(though
> currently still untested, by us at least), and they should start getting
> pretty reasonable by the end of the year and into next year. Even now, at
> $1000, as long as they work well, they will be quite competitive with
> proprietary touch screens from folks like Crestron.

Dude!  You I and clearly have differing opinions on what "commodity" items
are.  You're used to higher priced components than I am.  I'll mount
touchcreen controllers on all my walls when the price hits $50 per unit.  It
will be sooner than you think as the great worldwide gear-up to produce LCD
panels hits its stride and the glut comes.  I *hope* it will be soon,
anyway!  :-)

--
Bobby G.