[ale] MyLinux Personal Workstation initial production run (long)

[Joseph A. Knapka]

Hey folks,

A couple months ago I posted some material about the AZ Power
Personal Linux Workstation, a fairly hefty (performance-wise)
palmtop. AZ Power is going to sell these things for cost+$20,
which means that the more folks order them the cheaper they'll
be. This is a Hitachi SuperH-based beastie, base configuration
is 64MB RAM and 32MB flash for a filesystem, and I/O out the
wazoo. IOW, a machine that you could feasibly do actual work
with, not just PIM stuff. Here's the URL:

http://www.azpower.com/mylinux/

What follows is the announcement about the initial production
run. The target price for the initial run is $699/unit, with
an eventual target of ~$300/unit when the production runs
get big enough.

-- Joe

Rob Wehrli wrote:
> 
> I thought that I'd provide a quick update to the informal survey.
> 
> So far, we're about evenly split between the high end and the low end,
> with perhaps one or two more responses for the low end versus the high
> end...but a dead heat for all practical purposes.
> 
> Without a clear "winner," I'll need to default to a somewhat different
> mechanism to decide how we'll build the next round of 200 units.  How
> does the group feel about letting an "orders-based" approach work for
> us?  Our dilemma is that we need to do *a* production run, but which
> model?  The cost concern is that it will cost us less the further down
> the production path we go.  This first run will cost the most.
> Subsequent runs will cost a little less and a little less until we
> finally can do enough volume to make it cost quite a bit less...but that
> it at about the 2000 per production run point, so we're not going to see
> a big cost savings for quite a little while without a lot of orders.
> 
> Note that we're keeping strictly to the COST + $20 model!  With that in
> mind, I'd like to generally propose the following for your acceptance.
> Those wishing to place an order will go to our order page (being
> developed) and place their order for the configuration desired.  Credit
> card billing information is *taken* at that time, but no transaction
> occurs *unless* you state that you wish to make a deposit.  The idea
> here is that some folks might want to spread the cost of a unit over two
> credit card billing cycles, while we're building your unit(s).  We are
> currently seeing an approximately 4-6 week lead time on components,
> which puts an order placed for a unit at about 8 weeks delivery time,
> give or take a bit.
> 
> Now...since we're not charging an already known flat rate, but a rate
> that is capped at our list price less our actual cost (plus $20) plus
> shipping will mean that the final price will likely vary somewhat
> between the "target" price and what you actually have to pay to get your
> unit in your hands.  Here it is broken out in a bit more detail with
> numbers.  I'll use the example of the low-end units without prejudice.
> 
> Target Price: $699.00  (Final price can not exceed $699 plus shipping &
> handling)
> 
> Our Cost:     $612.00  (Bogus number, just used for illustration, actual
> cost is still unknown!)
> Our $20:        20.00  (Our "profit" ...only ~8,000 more units to go to break
> even!)
> Your Shipping:$ 18.00  (Guessitmate based on 2nd/3rd day to most US
> locations, Intl higher)
> 
> Total Due:    $650.00
> Less Deposit: $349.50  (50% of $699)
> Net Due:      $300.50  (Charged to your card when shipped)
> 
> Naturally, we're continuing to take every possible measure to constrain
> costs, but the first couple of thousand units are just going to be
> wildly expensive compared to someone with more money who can build a lot
> more than we can like Compaq or HP, for example.  Again, my personal
> goal is that we, like Linux and the GPL in general, only charge a
> relatively appropriate "distribution fee" rather than profits seen by
> big companies...while using the communitive buying power of our Linux
> Community to drive the costs down so that newcomers to Linux can also
> consider this product...but also so that we can eventually afford to do
> a complete product development cycle in less time and with equal or even
> better value-added than the Windows CE and Palm companies do today.
> Hey, I told you that this was a "grass roots" effort!  I feel my okole
> hit the ground just about every single step!
> 
> OK...so, let's get back to the process.  You decide that you want a
> unit.  You'll have the option of making the deposit the day you make
> your order, or waiting until we start building your unit.  Once we begin
> the process of ordering parts for your unit, your deposit is commited
> and charged against your card.  We'll initially be saying that deposits
> are NOT REFUNDABLE, even though none of us like to have to say it nor
> want to find situations where it causes problems for anyone.  The reason
> for this is that a few refunds could upset the price for everyone and we
> absolutely have to build at least 200 units at a time to stay within
> even these numbers.  I'm sure that some of you recall our goal of ~299
> per unit for the lower end.  The "unhappy" aspect of this is that that
> number is still possible, but not until we're able to build 10,000 at a
> time.  As I've said before, the "early adopters" will pay more so that
> the "generations" after them can *afford* to own a PLW."  The good thing
> is that we'll also be taking commercial orders from companies who want
> to buy them for their business needs, which will help to ease the burden
> on those of us who just want one for ourselves.
> 
> Once we get enough orders for the first 200 units, we'll fire up the
> production line and pop them out as quickly, but certainly with as high
> quality production results possible.  It would not do for us to send out
> units that do not work or break after a bit of use.  HOWEVER, we really
> need your help in working with us to correct any problems that we may
> have missed!  I can not stress this fact enough.  This is an entirely
> new product and I am quite certain that we'll have some quirks and plain
> old bugs that need to be fixed!  If we were a big company, maybe we
> could afford to build 10000s of units at a time and to have them perfect
> the first day you touch one.  I'm hoping that you'll have the patience
> and persistence to help us get over the hump and be a paying "wide area
> beta" tester with a smile :)  The first 200 production units will be an
> amazing accomplishment for the Linux Community as a whole and it will be
> a tremendously important advancement in ensuring that Linux is taken
> seriously by others.  Your order will put you into another special
> category as well...and that is in our list of credits!  Everyone who
> places an order for the first production run will have their "handle"
> pop up in the list of credits along with the core development team and
> contributors to the project.  Your handle can be anything you want that
> is clean and appropriate for public consumption; it can be your name,
> your email address, your nickname, your dog's name, however you'd like
> to be recognized as being a part of this community-oriented project.
> 
> Now: Let's assume that we don't get 200 orders...kinda worse-case
> scenario.  Your deposit will be credited to your card if charged and/or
> your order will be cancelled.  No matter what happens (barring
> destruction of the earth or some other such thing!), you'll get an email
> from us keeping you up to date with what's up.
> 
> Of course, if you can figure out a way to get everyone you can think of
> to order a unit (hint, hint), the price for all of us goes down...and,
> in the "best case" scenario, we can do a lot more than 200 units and
> perhaps even 1000, which would be a significant benefit to all of us in
> terms of final price we pay.
> 
> We are going to be very tight on this first run of 200 units without
> your help, your patience and your dedication to furthering the Linux
> movement.  With that in mind, those of you who can spare it, we're
> asking you to make your deposit at the time of order.  We will not spend
> your deposit until we actually commit to building the first 200 units,
> however, we will earn a bit of interest off of your money while we're
> waiting for a full 200 orders.  In the event that we do not proceed with
> production (not enough orders), we'll refund your deposit in full.  The
> amount of interest we make might be enough to help defray another
> expense or perhaps something that we have not foreseen yet.
> 
> Lastly, I want to be clear enough to advise everyone reading this that
> we're not a big company and we have a lot of financial risk associated
> with building 200 production units without your help...so much that we
> really can't do it without you.  For that reason, we must have the
> NON-REFUNDABLE deposit policy in place...at least for now.  If you find
> that you do not want to follow through with your order after production
> starts, we'll try to connect your order with a hopefully existing,
> uncommited order and free both your deposit and your obligation to
> follow through on the balance due of the order...HOWEVER, I can not
> guaranty any results of being able to "swap" orders with someone else at
> this time for the obvious reason that we do not know how many orders
> we'll get yet.  If, after production begins, but before shipment of your
> order you decide that for any reason you do not want to continue, you
> will forfeit your deposit in the event that we can not swap the order
> with another Linux advocate (or other kind of customer :)  We don't
> believe that you should have to pay the entire amount if you're in a
> pinch of just lost your job or whatever reason might compel you to pull
> out of the deal.  Our belief in our product says that we're willing to
> put up the remainder of the cost per unit to be able to offer it for
> sale to someone else.  Hopefully we'll do very well and have no major
> problems with orders, cancellations and such, but I figure worst case it
> bankrupts the company and we have a bunch of units (without customers
> for them) that some bank gets and auctions off at pennies on the dollar
> to Microsoft who converts them to run Windows CE and ...I'm going off on
> a tangent, huh?  Under no circumstances will we begin production (and
> thus commit your deposits) without being damn sure that we're going to
> complete the production run.
> 
> With that in mind, I'd like to make a short statement about production
> run and time.  With not being a large customer in the component
> industry, we do not carry enough weight to be able to make manufacturers
> ensure delivery of the components that we need in as timely a manner as
> larger companies.  We are somewhat at their delivery scheduling
> discretion.  We will endeavor to second source any component that
> doesn't seem like it will be delivered promptly enough to meet our
> production schedule requirements, but there are a number of components
> that are very much single source.  So far, I do not predict any
> unbalanced situation as a result of component availability, but "stuff"
> happens--is my point in all of this.  If it happens, after we've taken
> your deposit and commited resources, we're kinda stuck in a situation
> that truly sucks all the way around.  My general thinking on this is as
> follows: Please feel free to comment however you'd like.
> 
> In the event that we're merrily ordering components for 200 units and
> suddenly a supplier says, man, you're not getting those for 20
> weeks...what am I going to do?  My first check will be to evaluate a
> second source.  If one exists, then we're going forward again, no harm,
> no foul.  If one exists, but they're are a bunch more expensive...then
> I'll see if we've got enough to cover it in our $699 target price
> maximum.  If we do, well, the cost just went up for all of us by a
> little bit, but is at least at or under or max price per unit.  If it
> would go over the $699 but be contained within the $20 "profit"--then at
> least you guys and gals don't feel any impact other than the maximum
> price per unit.  The worst case on this is what would happen if it went
> over the $699 max AND the $20 in the COST + $20.  Well, we're a bit
> screwed...is the short answer.  The long answer is that we take a look
> and see where we are.  If we haven't ordered too many components at that
> time and can cancel the production run and refund everyone's deposit,
> that's what we'll do...(This *is* taking worst-case out a ways, huh?)
> If we've ordered a BUNCH of components, then we have to do something
> else.  My choices legally include bankrupting the company and everyone
> loses their deposits and the banks sort the mess out...or refunding
> whatever portion of deposits to people is possible and declaring
> ourselves a center of a "natural" disaster (maybe state and federal aid?
> :) or, sucking it up in some way that includes postponing some deposit
> holders for a production run cycle, building forward what units we can
> still complete on "budget" and making it back up in subsequent
> production run(s).  The last thing that I want to see is anyone take a
> loss, including me and all of you.  Our entire goal for this project
> from the start was to have an open source Linux hand held that was based
> on a low-cost model that made it easier for us to buy one.  Well, we're
> now at the edge of that little cliff and are flapping our virgin wings.
> Will we fly?  Will we crash and burn?  (Stay tuned for the next episode
> of...)
> 
> Seriously, I think that we're in a very good situation with the best
> possible odds (short of having tons of cash) to make a successful go at
> this.  My plan (no plan is perfect, huh?) extends the commitment time of
> resources until we have 200 orders waiting to go.  At that time, we'll
> get updated commitments from our suppliers before commiting your
> deposits.  We'll keep you in the loop by keeping the messages, when
> relevant, flowing to this mailing list.  I'm a "very much in favor of
> voting" kind of guy.  If there is additional risk at that point, I'll
> push it out to the body of this group and let you, as a community,
> decide what you want to do...whether we go forward or pull out.  I can
> not stress how much work we've put into this project to get to this
> point, but I am not about to gamble on bad odds with your money...even
> if a whole lot of my money is in it too.  I am able to accept this loss
> if that is what happens, but I do not feel that it is my right to put
> you into a position where you're faced with a possible loss.  Therefore,
> I'll do everything in my power to make this project successful and with
> our combined efforts, I think we'll have the coolest Linux product in
> the world.  (We've got a new, high-tech plastics design center
> (http:www.rapidpro.com) working on the "professional" version of the
> production plastics.  I'll get some photos up online when I get
> something back from them on it.)
> 
> Anyway...I better go, before this becomes a book.  I like keeping in
> touch with you and making sure that you have the benefit of what I know
> at the time decisions get made.  We're in this together.  I can't build
> them without you.  You can buy an Agenda for a lot less money.  You can
> buy a Compaq and work with getting a port up and running on it, or you
> can have the real deal, from the ground up, Linux device from start to
> finish, built and supported by the Linux Community for the Linux
> Community...even has Linux in the name, MyLinux!  I'll wait about
> another week to two weeks before starting to take orders (or deciding
> against it) based on your input over the next several days.  Now is the
> time to start communicating anything you've got to say.  We're fairly
> busy trying to get the production line started, but we want to answer
> your questions, too.  Feel free to send me mail privately if you have
> questions you don't want going out over the list.  Talk to your company
> and see if they'd buy you one...or if there are others you know who want
> one.
> 
> BTW...Hitachi is doing a write-up of the entire MyLinux Project in their
> newsletter magazine that should be fairly exciting.  It should come out
> just about as we go into production, give or take a bit, and might
> really assist our order volume and put us up to another level of buying
> power...which translates into another level lower cost.  Oh well, gotta
> go.
> 
> Take Care.
> 
> Rob!

-- Joe Knapka
"You know how many remote castles there are along the gorges? You
 can't MOVE for remote castles!" -- Lu Tze re. Uberwald
// Linux MM Documentation in progress:
// http://home.earthlink.net/~jknapka/linux-mm/vmoutline.html
* Evolution is an "unproven theory" in the same sense that gravity is. *
--
To unsubscribe: mail majordomo at ale.org with "unsubscribe ale" in message body.