Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ucbvax!DIAMOND.BBN.COM!dm
From: dm@DIAMOND.BBN.COM ("Dave Mankins")
Newsgroups: comp.windows.news
Subject: Re: Traffic on the X list
Message-ID: <8807061628.AA22039@quartz.BBN.COM>
Date: 7 Jul 88 23:51:57 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 65


> [talk about how the $100 fee for News binaries is a tiny
impediment.]

A couple of years ago (post News, although at the time it was
still called Sundew) I worked for a company that builds a
large-scale parallel processor which does not run SUN binaries,
and at the time didn't even run UNIX.  For the price of a few
hours spent running FTP, and a couple of days spent debugging
my Berkeley socket emulation library, we had X up and running
on this machine.  We were able to show our management that X on
this machine was a reasonably good thing, and they agreed, and
a little while later it was a part of our product line (it
helps that we liked to distribute sources with our products).

(It was this easy because our machine is a computer, not a
workstation, so I didn't have to port the display portion,
which is more hardware-dependent.)

What would it take to talk my boss-beings into springing for
the $40K source license, signing a semi-restrictive agreement
with what was, more or less, a competitor, just to do an
EVALUATION? Especially when that $40K could immediately be
applied to buying workstations to make our programmers more
productive?  I'm not flaming SUN, I think that if you do good
work, you're entitled to ask money for it, if you so choose. 
However, I think that in order for something to become a
standard, you have to be pretty loose with the sources, so people
can port it to their machine without spending more time talking
to the company's lawyers than they spend doing the port.

I hate talking to lawyers.  Even believing that Sundew was a
superior paradigm, X was preferable, because we could spend the
News source license fee putting Sun 3/50s on 10 of the
department's desks immediately instead of a year later.  It would
be a LONG time before the value-added of News over X came any
where close to the value-added of those 10 workstation-years.

News ain't ever likely to run on that machine, because:

 o   they've got a ``pretty good'' window system now, one that
     interoperates with all our other workstation and CPU vendors
     --- we can even put PCs on the net running X

 o   since they've got X, they no longer pay much attention to
     yet another window system from SUN, so they don't know
     what they're missing 

 o   the above two reasons have to be overcome to justify the
     expense of a source license (which they need to port the
     software to their multiprocessors and their vaxen)

I'm sure that they're not the only hardware vendor in a similar
fix.  The News BINARY license is comparable to the expense of
getting the X SOURCE tape, if you're on the internet, X is FREE. 
The value-added of News over X isn't immediately obvious, so
why spend all that money on a gamble (at the time Sundew first
came out, it was Suns THIRD incompatible window system.  Who
could predict that there wouldn't be a fourth)?  The first
they're likely to see of News will be after the X/News combined
server comes out, so they can run their existing software on it.

X is like a weed, it can germinate and spread to your machine
without much attention on your part.  With News you have to spend
$40,000 in garden tools and fertilizer before you even get the
seeds.