Pierce,
I think that you are assuming the following:
1. You assume everyone who purchases WO will build a one-time custom
Internet app, like a store or something. By assuming this, it is very easy
to justify the cost of WO as it drowns in the development costs. But this
assumption is not true. Some of us build almost shrink-wrapped
applications on WO and sell them with WO. In our case, the WO cost becomes
a significant cost of the total solution for the client.
2. As I mentioned before, you also assume WO will be used only for
Internet use, where the cost of communications lines are high. Our apps
are for intranet use where bandwidth is practically free.
When you make your cost justificaiton charts, I hope you will address these
issues.
-Afshin Behnia
-----Original Message-----
From: Pierce T. Wetter III [SMTP:[EMAIL PROTECTED]]
Sent: Wednesday, March 10, 1999 8:13 AM
To: Multiple recipients of list
Subject: RE: Unfair WO4 pricing - the attitude of apple
At 2:32 AM -0800 3/10/99, Jeroen Clarysse wrote:
> I think that apple isn't interested in small developers. WO is simply
> targeted at big projects : $100.000 or more. Then, indeed, the software
> cost is acceptable. But WO pricing is not 'splittable' over multiple
small
> projects since tpm is usuallly the same for all sizes.
Well, I'm trying to put together a scaling chart to show everyone
exactly how you get from a small project up to a large one, but in
the meantime, I've been trying to point out that WO doesn't really
cost $25K.
But no one seems to be hearing me, which means that I'm probably
don't understand what the others are saying any more then they
understand me.
Would someone explain to me what it is I'm not getting about this
point of view? I would think that I do understand it, because I've
used WO for small projects before. But I guess I don't.
Here's my point of view:
1. Profit > WO costs long before you reach the TPM limit.
$1000 or $5000 (with kick-ass hardware) gets you a WO deployment
for 50 tpm running on MOSXS, including OpenBase Lite as your database.
If you have to look at 10 pages to do an order, 50 tpm is 5 orders
per minute. If your average order is $30, with profit of $3, then
your profit per hour is: $900/hr. Profit per 8-hour day is $7200.
So one day's _profit_ would pay for the next cost upgrade to WO.
2. Network costs > WO
I argued this before, but if your web pages are 5K each (including
images) then 50 tpm is 250K/minute, or 34kbs, which means an ISDN
line, which means 200/month minimum in Internet costs.
At 100 tpm, you need a T1, which is $1000/month.
So the WO cost is on the order of half the network cost of running
your server for a year.
3. TPM is not users.
In my experience, I see 3-10 users for each TPM. That is, users
take 3-10 minutes on average per page, even in intranet apps. Sure,
while they're using it, they spend a minute or less per page, but
they don't use it constantly. They talk on the phone, they go to the
bathroom, they drink coffee, they talk to their coworkers, and some
of them even spend time thinking! No matter how cool my WO
applications are, they aren't the most important thing in my users
lives...
So 50 tpm is 150-500 users, so $1000 seems pretty cheap.
4. Hardware costs > WO
This is the weakest point in my argument because I don't have hard data,
but:
At 50 tpm, WO at $1000 is 20% of the system cost ($5000).
At 100 tpm, you're going to need to get a better database.
OpenBase Lite is ok, but I don't think it can handle 100 tpm (but I
don't know for sure, but it seems to be choking on my little
application at 1 tpm). So you're going to need to upgrade your
database technology. From what I've seen of ASP, SQL Server, and all
the other "cheap" technologies, they can't handle 100 tpm either.
That means a "real" Database like Oracle or some such.
So expect WO to eventually become a small fraction of the total
system cost, with a small bump at the 100 tpm level.
Here's what I think people get stuck on:
"unlimited" transactions is a fallacy. "Unlimited" means "as many
as your hardware can support", not infinity. If you want to support
more TPM, you have to start buying more RAM, and more processor, and
so on.
As a preview of the scaling chart I'm building, Standard & Poors spent
$1,000,000 just on RAM for their servers. The $150,000 they spent on
WebObjects was nothing compared to that.
Pierce
----------------------------------------------------------------
Pierce T. Wetter III, Director, Twin Forces, Inc.
e-mail: [EMAIL PROTECTED] Phone:520-779-4227
<http://www.twinforces.com/>
U.S. Mail: 1300 South Milton Rd, Suite 206, Flagstaff, AZ 86001
