Re: Drivers, Upgrades and SW design (was Re: Dedicated Z-wave sites?)

["Dean Roddey" <droddey@xxxxxxxxxxxxxxxx>]

"Robert Green" <ROBERT_GREEN1963@xxxxxxxxx> wrote in message
news:CbidnXiEvYM6yAfYnZ2dnUVZ_sSmnZ2d@xxxxxxxxxx
> Why am I "stuck" on this issue of drivers?  We alluded to HomeSeer, if you
> recall, and their decision to begin offering "plug ins" for an extra fee.
> There was a lot of discussion some time ago about how that decision might
> have affected the overall quality of the product.  While it's clearly only
> a
> guess, one reason they might have chosen to use third party plug-ins so
> was
> the sheer magnitude of the problem. There are so many devices needing
> "plug
> ins" or "drivers" (or whatever label one chooses for the SW that
> communicates among the application, the OS and the HW) that it's easy to
> see
> that the man doing the core programming might have trouble doing it all.
>

I'm sure I've already explained this, but just for the fun of it... HS
allowed people to write plugins in any language they want, and to pretty
much do anything that the system allowed. CQC doesn't allow this. We have a
strictly defined driver architcture and they can only write drivers in our
PDL or CML languages, which means that we control how drivers are structured
and what they have access to. We brought forward all our existing drivers to
our 2.0 release without trivial effort. HS took years and still hasn't
gotten them all moved forward.

> Part of my dubious attitude comes from starting out learning linear coding
> techniques and morphing through OOP.  I'm still not certain that OOP has
> produced all the benefits
> that it promised although I'll grudgingly admit it has brought some.  But
> who's to say what happened to Fortran won't happen to object oriented
> platforms five or ten years from now?  There's a revolution coming in
> multiple CPU motherboards that's going to require an OS very different
> from
> anything the current Windows platform can deliver.  A media-savvy home
> automation application program would seem to me something that would very
> much benefit from having multiple CPUs to handle decoding multiple video
> and
> audio streams.
>

OOP is the only way that our system could exist. It wouldn't be remotely
practial to write this system any other way really. It would be too fragile
and require far more resources than would be reasonable. OOP exists at two
levels. One is the way you often here it waved around in the press, which is
that you'll magically plug in software for various vendors and it'll all
work. It can be done, but it's very difficult. We use it in a completely
different way, which is to create a very powerful single OO framework that
is all built as a system to work together. In that way, OOP is every bit as
powerful as it is made out to be.


> Dennis talked about that but he also talked about the profound design
> changes in Vista that are going to seriously alter the way drivers are
> written and how they work. Vista represents such a change in a number of
> areas that even MS, the *authors* of the OS, couldn't get the new
> data-based
> file system working correctly in time for the initial release.
>

You making contradictory arguments. You argue that drivers are a problem
because of errors that destablize the system, but you then argue that Vista
is horrible because it changed the driver model in order to fix this
problem. Yes, it will be painful in the sort term, but the long term
advantage will be less problems with drivers because they'll be less able to
destablize the system.