[nycbug-talk] Re: [nylug-talk] open source HVAC control software

alex at pilosoft.com alex
Fri Jun 3 14:35:14 EDT 2005

On Fri, 3 Jun 2005, Michael Shalayeff wrote:

> Making, drinking tea and reading an opus magnum from alex at pilosoft.com:
> > On Thu, 2 Jun 2005, George wrote:
> > 
> > > Actually, do some heavy googling because FBSD's PHK is apparently doing
> > > many things in that arena with Soekris boxes. . .
> > > 
> > > But you'd have to agree with Alex that it seems frivilous to go this
> > > route.  It might make sense to figure out getting data over RS485 to a
> > > box, then what would allow you to tail the data over that route. . .
> > Bottom line is, when your open source-picked thing fails and sets the
> > building on fire by setting your HVAC to heat instead of cool, what will
> > you tell the board members?
> this is typical american boolshit. what does it matter who to blame? or
> do you mean "commercial" software does not fail? every single commercial
> piece of software is a pile of poo. and guess what? there is nothing you
> can do about it. you live w/ it and that gives you false sense of
> security and makes you think it will just work because there is no other
> choice and if it does not you will find your way around to make it
> _kinda_ do what you want. and you live w/ it blaming the vendor for your
> own inability to make things work...
No, not really. There are commercial products that are excellent. There 
are open source products that are stacked piles of poo. Poo-pooing 
(literally) commercial software and saying open source is always better is 
just as silly as saying "open source sucks". The reality is, there are 
cases where open source fits, and there are cases where it doesn't.

> > RS-485 interface is fairly low-level and you can control specific
> > parts of the unit via that.
> > 
> > This isn't linux on desktop, kids. This is real stuff. Code fails,
> > things catch fire.
> yes and real stuff is not about this stupid bigotry it's about YOU
> making sure it WORKS before YOU deploy it into USE irregardless of what
> it's made of.
Don't get me wrong, I love open source. However, unlike the zealots, I
recognize that there are cases where open source is simply not appropriate
for the business requirement.

Let's compare two cases, one is when Joshua bites the bullet and purchases
the big bad proprietary software, and second one is when he finds hackers
on the internet to write code to control his HVAC unit, and see what
really makes sense in *this specific case*.

Case 1:

a) Software is developed based on manufacturer published docs and/or 
reverse engineered manufaacturer's software.

b) Software is tested on Joshua's single unit, while being developed by
hackers around the world. Nobody knows if the unit will catch fire (or
just die) doing development and testing if you sent a wrong command.

c) If the software malfunctions, it will be Joshua's head on the line.

Case 2:

a) software is developed by manufacturer's software engineers in close
contact with engineers who developed said hardware.

b) software is tested on manufacturer's units, which are designed to be 
'test units', if unit dies, it isn't a big deal

c) If software malfunctions, Joshua's legal counsel will be firing off 
negligence lawsuits.

Don't get me wrong - I would love if all manufacturers released full 
documentation and open sourced basic code libraries necessary to talk to 
their hardware. But, that's not the world we live in. In the real world, 
there are cases where open source just isn't appropriate

More information about the talk mailing list