[Tfug] Opine: Bricks, warts or...?

[Bexley Hall]

--- On Wed, 8/6/08, johngalt1 <johngalt1 at uswest.net> wrote:

> From: "Brian Murphy" <murphy+tfug>
> 
> > I'm going to hate myself for extending another ill-defined
> > no answer good enough Bexley thread but...
> 
> "ill-defined no answer good enough"
> LOL

Yeah, I found it amusing considering how many folks (Brian
included) *seemed* to understand the "ill defined" subject
of the thread!  :>

> Fscking engineers -- talking about things in voids between
> boundaries with respect to design philosophy. How dare they
> consider stuff that doesn't often result in a tidy
> conclusion, and in mixed company!

Note that I deliberately avoid asking technical questions in
this group (since it isn't a technical forum).  Those 
discussions happen elsewhere -- more appropriate venues.

Even there, it's often hard to explain subtleties of some
of the tradoffs involved to people without the same
skillsets.  :<  (Recent discussions re: psychoacoustics were
particularly tedious -- I've learned to just "trust those
who know" and nod politely  :> )  And, language barriers
often make things complicated (we aren't all English speakers...
*especially* the Brits! ;> ).

But, since we are each aware of -- and respectful of -- each
others' capabilities as "professionals", everything *always*
gets resolved -- by deferring to the "experts" in whichever
area is under discussion.  That's what makes engineering so
much more fun vs. more "routine" jobs...

OTOH, I'm also working with a different FOSS/H group designing
a VoIP phone around a Palm PDA (recall my earlier comments re:
PDAs  :> ) that tries to be more "democratic" about the decision
making process.  I'm convinced that group will splinter at least
once, maybe *twice* before coming to a "final design" (there are
just too many voices for too little work to be done).  And,
guestimate a 40% chance that they'll ever finish/release a 
working product!  :<  (for many of the reasons outlined below)
A shame as it is an amusing approach.  Probably it were only
two or three people it would stand a better chance of happening.
Hopefully I'll get something I can use out of it...

> Ahh, what if doing everything was like making stuff with
> Lego blocks. As long as you have enough of the right Legos,
> you could build anything.

ROTFL!  Sure, and if you always had "enough of the right
*money*", you coud *buy* anything!!  (therefore eiminating
the need to ever *build* anything!)

> And then if its cool enough, your creation could go on
> http://boingboing.net in between the comic books and artsy
> BS.  Nirvana!

The problem we FOS{S,H} developers face, is coming up with
the "right" approach that users can accept/tolerate.  This
is especially significant with FOSH designs!

With FOSS, you (as a consumer) *possibly* might be skilled
enough to modify the code to make it work the way you
want.  Since most FOSS is often "desktop based", you also
*probably* have a platform on which to run and develop those
changes.  You can come up with your own special version
without requiring anyone else's help/cooperation.  Barring
this, you may be able to influence someone else who *can*
do the work (for you).

This is a fair bit more difficult with appliances.  You're
working in a cross-development environment (this weeds out
many folks who are used to self-hosted designs) and, sometimes,
even stuck doing things like Canadian builds (with all the
headaches that involves for tracking changes (build/host/target
environments).  You probably *don't* have a real "target" to 
run on and are either working under a simulator, a (hardware)
emulator (if you have $$$) *or* writing VERY portable code :<
On top of that, you are working in a severely constrained
environment (you can't just use resources in your code as if
they were "unlimited" -- as they are in desktop environments).
But, it *is* "do-able"!

OTOH, with FOSH, the skills and *tool$* required are
usually out of most users/consumers reach.  You probably
can't read a Gerber photoplot -- let alone *modify* it.
And, having compatible PCB layout tools is probably not
on your list. (we've spent a fair bit of time trying to
figure out just what to be "compatible" with  :< )  As would
be manipulating the schematic "sources" (for the board as
well as any programmable logic employed on it).  So,
"hacking the sources" is already an impediment to your
"infuencing" the design.

Even if you *have* those tools/skills to modify the design,
prepare manufacturing documents, etc. the cost of making a
"one off" version of a design that fits *your* idea of what
the product should be usually doesn't make sense economically!
You have to find a board house, someone who can program the
FPGA's, assemble SMT/SOIC components, *test* the assembly,
etc.  The setup charges and overhead for each of these
tasks are just ridiculous!  (e.g., you can get three copies
of a prototype PCB for *exactly* the same price as one.
But, you can probably get 100 copies of that same PCB for
*twice* the price of that first "one").  You're talking
anywhere from a few hundred dollars to a thousand dollars
for the first prototype *bare* board(s).  That suddenly
makes rolling your own design pretty unappealing!

As a practical consequence, you end up having to create
your own "mini FOSH project" -- enlisting others who *might*
share your ideas of what changes need to be made -- so
you can take advantage of economies of scale.  Of course,
this *still* means you have to find a suitable manufacturing
facility for your "small lot", come up with enough *buyers*
for the boards, a cheap source of components in these
naturally reduced quantities, etc.

But, then you invariably end up with a bunch of NEW changes
as each of these people want his/her own "pet" change added
to the mix -- regardless of cost (effort/materials).  :<
This is usually enough to torpedo any "special design" you
might WANT to come up with.

This leaves you two choices:  live with the existing FOSH
design *or* pay through the nose to roll your own!  (or,
hope the "for profit" community comes up with something
more palatable to you -- at *whatever* price...)

As a result, FOSH developers have to ask "potential" and/or
"generic" consumers for ideas about how they perceive
certain aspects of *existing* products.  And, based on those
comments and the "logic"/reasoning behind them, decide which
of many competing approaches to adopt in implementing a
particular design.  Remember, unlike software where changes
only cost *time*, changes here cost *money*, too!  (you end up
scrapping your previous prototype and all of the components
soldered onto it!)

Of course, we don't *have* to do this -- since we can just
(selfishly) design the product that *we* want.  Usually, there
is enough of a "market" among ourselve (and friends, etc.)
to get the quantities up high enough to be affordable.

But, if we come up with a design that others will *also* want,
then it is easier to find someone to manufacture them in larger
quantities (so they get even cheaper FOR US *and* we have an
ongoing supply -- instead of having to rely on one large
"prototype batch")

I wish I could recall the correct attribution (as well as
the exact *quote*):  "Listen to your users (consumers), but
ignore what they say".  :>

(IIRC, Henry Ford is credited with uttering something to the
effect of "If I had listened to what my customers *said*
they wanted, I'd have given them a faster HORSE!"  Insightful
especially considering the first autos were considerably
SLOWER and LESS reliable than horses!)

<grin>