btb@hogpc.UUCP (B.BURGER) (06/03/84)
> The obvious reason for vendors spliting up UNIX when they sell it is > it's a good marketing trick to increase how much bucks they get. It's > the same type of thing that's done in sell cars or anything else, > make the entry price low and get them on the "options". From a business > standpoint it clearly makes sense and from a consumers point of view > it sucks. Ah yes, consumers would be better off if they *had* to take a car with every possible option, whether they wanted it or not. As a matter of fact, they'd be better off if all the auto manufacturers stopped making so many different cars and just made fully loaded Cadillacs. THE ONLY PRICE THAT COUNTS IS THE PRICE OF THE CONFIGURATION *YOU* WANT! It shouldn't worry you that there may be other configurations at other prices. Partitioning increases flexibility, which minimizes the price for each configuration. (The only caveat is that giving people choice creates a cost for keeping track of who ordered what and making what they ordered.) Prices can, of course, be reasonable or unreasonable with or without partitioning. But partitioning *in itself* reduces the price for a most configurations. This applies to cars, computers, and everything else. --Bruce Burger AT&T-Information Systems Lincroft, NJ {...ihnp4!}hogpc!btb
boylan@dicomed.UUCP (Chris Boylan) (06/03/84)
The point is that most people WANT a loaded [object goes here].
Partitioning makes the APPARENT price lower while in reality
people don't buy the base [object goes here]. An example or two:
The price of the recommended car configuration from consumer
reports for a particular brand is about 20% over the "price"
of the car as advertised.
UNIX for (say) the Apple Lisa is 495, BUT if you want to write
a C program on it or relink the system it costs 1495.
As I said before, the really great thing about partitioning is
it's usefulness as a marketing tool. It's so easy to get sucked
in by it ...
--
Chris Boylan
{mgnetp | ihnp4 | uwvax}!dicomed!boylanguy@rlgvax.UUCP (Guy Harris) (06/04/84)
> The point is that most people WANT a loaded [object goes here].
Is that truly the case for "object" == "UNIX system"? A car with power
steering, power brakes, stereo, air conditioner, etc., etc. doesn't require
extra skills to use the options. However, a shop buying a small UNIX box
(remember, the 3B2 isn't a big 40-100 user development machine, it's a
small box) may not *have* any C programmers, or indeed any programmers at
all; it may be a small business that wants to run its accounting and word
processing on a small computer and possibly provide financial analysis
programs for the owner. If these people were told, "Well, if you want to
give up some disk space you could otherwise use for accounting files and
documents and spreadsheets, you can get a program that allows you to write
programs in the C language and run them on this machine," they'd probably
say thanks, but no thanks.
I agree it can be used as a marketing tool, but I don't know that that's
how it actually is used. Given that the computers are being sold by AT&T,
the argument that "well, the manufacturer pays AT&T the same price for
the binary sublicense, regardless of whether the system offers a C compiler
or not" doesn't apply. If AT&T unbundles UNIX in the same fashion, the
same would hold true for other systems.
If partitioning is used as an excuse to raise prices, I agree, it's annoying.
It's annoying that the machine-readable form of the UNIX documentation, other
than the User's, Administrator's, and Programmer's Manuals, is no longer
provided as part of the basic System V package; on the other hand, if
unbundling that documentation makes the purchasers that actually use it
pay for the actual cost of providing it, I can't really fault AT&T for it.
Let the market decide. If, indeed, most purchasers of small UNIX boxes
including the 3B2 want the full system, and shun the 3B2 because the full
system costs extra, AT&T will come around or end up with warehouses full
of 3B2's. If most purchasers of small UNIX boxes couldn't care less whether
such a box comes with all of System V, the 3B2 won't be hurt by the partitioning
and most customers will get what they want.
Guy Harris
{seismo,ihnp4,allegra}!rlgvax!guyok@edai.UUCP (Richard O'Keefe) (06/08/84)
I'm emotionally opposed to unbundling (why can't we be a little
more honest and call it "dismembering") because I'd like to get my hands
on *everything*. But I'm not a computer buyer.
The claim has been made that dismembering is good for the customer
because it means his 10Mbyte disc isn't cluttered up with things he
doesn't want. Nice argument. But it depends on two assumptions:
A) the only place you have to store software you are not
using right at the moment is on the 10Mbyte disc.
As we've been told that when you buy a new package, it comes on
floppies, this is evidently false. The storage cost of getting
programs you're not using at the moment is a drawerful of floppies.
B) you know exactly what tools you will need when you buy
the system, and either your requirements never change, or
when they do change you can either do without anyway or
be sure of getting the money for your upgrade out of your
bosses and then afford to wait for the floppies.
This is unlikely to be true either. There are lots of environments
where it is easier to get another $500 at the time of the initial
purchase than another $200 subsequently. (E.g. they come out of
different pockets.)
Given that the text formatters have been split off into a separate
package, I foresee a boom in "roff" lookalikes powerful enough to format
new manual pages. Standardisation was great while it lasted. (Have
AT&T noticed that TEX82 is public domain?...)
Guy Harris says let the market decide. But there is no reason at
all to expect the market to converge on a rational approach. There are
quite a few places over here planning to switch from 4.1 to System V.2
when/if V.2 sees the light of day; not because they know it to be
technically superior (or indeed because they've been able to find out
anything much about it; I was at a meeting where a manufacturer's
representative told us that we didn't want 4.?bsd because V.2 had
everything we wanted, and then answered most specific questions by
saying he hadn't the faintest idea and wouldn't have until AT&T let his
company have the documentation) but simply because of the magic of the
AT&T name. The market will go in whatever direction minimises present
fear, not future regret.
YES the authors of Unix are entitled to recompense, YES people
shouldn't have to pay large sums of money for programs they don't want,
but what have AT&T done to yacc/lex since V7? If, for example, yacc
were made to generate Ratfor (or EFL) again, or perhaps Pascal... I'm
sure AT&T have put a lot of effort into the kernel, but that's the bit
everybody buys.
Yours in waiting for GNU.guy@rlgvax.UUCP (Guy Harris) (06/09/84)
> Given that the text formatters have been split off into a separate > package, I foresee a boom in "roff" lookalikes powerful enough to format > new manual pages. Standardisation was great while it lasted. (Have > AT&T noticed that TEX82 is public domain?...) Depends on the market. I suspect the typical UNIX user five years from now (who will probably, *noli voli*, not look like us) may never see "UNIX" *per se*. They'll probably talk to the computer through a tightly- sealed application interface, or possibly through a "desktop" shell *a la* Xerox PARC's systems, Lisa, Mac, etc.. They may never even *see* the UNIX manual. Look at it this way; yes, you can buy shop manuals for your car, and all the specialized tools you need to work on it. Most people don't. And if it's a company car, almost nobody does. > Guy Harris says let the market decide. But there is no reason at > all to expect the market to converge on a rational approach. There are > quite a few places over here planning to switch from 4.1 to System V.2 > when/if V.2 sees the light of day; (It already has, at least on this side of the big pond.) > not because they know it to be technically superior (or indeed because > they've been able to find out anything much about it; I was at a meeting > where a manufacturer's representative told us that we didn't want 4.?bsd > because V.2 had everything we wanted, and then answered most specific > questions by saying he hadn't the faintest idea and wouldn't have until AT&T > let his company have the documentation) but simply because of the magic of > the AT&T name. The market will go in whatever direction minimises present > fear, not future regret. Which isn't necessarily a bad thing. Just ask the current owners of DECSystem-10s and DECSystem-20s (not the universities, the commercial customers) how they feel about the flyer they took, now that DEC's canned the top-of-the-20-line Jupiter project. Technological superiority isn't everything. Iron isn't the only thing that costs money. A case *can* be made for picking up S5R2 instead of 4.1BSD; if S5R2 had paging, it'd be a very strong case. Most "straightforward" software can be made to run under any "V7 derivative" (meaning just about anything except 4.2BSD, and even there the differences aren't *too* bad as long as you know they're there), but not all packages *will* be made to run under all such flavors of UNIX. Furthermore, a lot of customers will buy S5R2 for the support, although that may change now that DEC's offering Ultrix-32. > Yours in waiting for GNU. I'll believe in GNU when I see a copy, and I'll believe it succeeds when the number of GNU sites greatly exceeds the number of UNIX sites. Not before. Lauren Weinstein discussed rather cogently the risks of a system with the "freedom to hack" proposed for GNU, and the likelihood of the proliferation of incompatible versions of the beast, a while ago. I say "let the marketplace decide" because I think a lot of us are arguing from the point of view of a UNIX hacker with an emotional attachment to the system we grew up with. Well, it's not that system any more. Yes, it may not be sold under the same terms as we've gotten used to. As I said before, "that's capitalism". No moral judgement intended. AT&T isn't selling UNIX for the goodwill anymore; they're in it for the long green. If you want to engage in behavior modification, you'll have to understand their reward and punishment system, and reward them by buing their product or punish them by not buying it. Moral suasion won't do it unless you can back it up with something to make their bean-counters sit up and take notice. Guy Harris {seismo,ihnp4,allegra}!rlgvax!guy
josh@joshua.UUCP (Joshua Gordon) (06/13/84)
I'm running on a "partitioned" system right now: an altos-586 running
XENIX with, yes, a 10-meg disk. Indeed, there is not a heck of a lot of
room on here (I have the development package); only a few man pages (mostly
for stuff I got on the net, and a few pieces of local stuff); it is
somewhat of a nuisance, but I can certainly understand a company partitioning
the system. Even without the development package, ten megs is not much.
and I've stripped the thing myself even more (got rid of spell and its
froofraw; chucked out useless phototypesetting utilities; that sort of
thing) so I can keep at least a few days news online at a time.
--
from the blithering idioms of josh gordon
{ihnp4,ucbvax,cbosgd,decwrl,amd70,fortune,zehntel}!dual!joshua!josh