jw@pan.UUCP (Jamie Watson) (10/22/87)
The following is a request for advice on a decision that I have been wrestling
with for quite some time now. I suspect that it will be of relatively wide
interest; in fact, I'm really quite surprised that it has not been brought up
before.
I am one of a small group of consultants working on an application for a
customer here in Switzerland. The people working on this project have quite a
bit of experience in Unix internals, system software, and applications; we can
do what is necessary to complete a project, given enough time and money. At
the moment, however, we have to make a decision about using X windows.
We, and our clients, believe that X windows, and the X tookit(s) are the best
choice for developing application interfaces now and in the near future. The
questions is, what version of X, what toolkit, and when to actually start the
work.
We have not made a final decision on hardware yet. The contenders are:
. Vaxstation (II,GPX,2000)
. HP 9000/300
. Opus
. Sun
The choices we have before us now are:
. Using X.V10R4 as distributed by Project Athena, and taking it upon ourselves
to bring it up on the particular hardware we choose.
+ We have the source code; lots of well known benefits to this.
+ V10R4 has been around for quite a while, so it is relatively bug free, and
is pretty well optimized.
- V11R1 is out (although we haven't received our tape yet). This implies
that we would have to make a conversion effort in the next 6 to 12 months.
Not my idea of a good time.
- Our clients would really prefer that we spend most of our time working on
their application, not on the windowing system.
. Using the X windows release supplied by the hardware vendor.
+ We would save considerable time by not having to bring up X ourselves.
+ In most cases, the vendor has 'optimized' for their particular system.
+ We have someone else to yell at about problems, instead of having to fix
everything ourselves.
- None of these manufacturers are willing to give or sell us the source code
of their own implementation of X. Yelling at them about bugs is fun.
Waiting for them to fix the bugs is not.
- The manufacturers are all going very slowly in bringing out V11. We would
have to start with 10.4 and convert later.
- The manufacturers are currently very inconsistent in toolkits. DEC has no
toolkits at all in 32w 1.0, and I hear there is only the DEC toolkit in
1.1. HP has only Xray in the current release. Opus has a preliminary
version of the V11 library, adapted (somehow) to 10.4.
- This option effectively rules out Sun right now, since they will not be
shipping X until next year sometime.
Other pertinent data:
. There is no shortage of bugs in the V10 toolkits we have on our systems
right now (DEC and Opus).
. Information on the net indicates clearly that the V11 tookits are not a
stable as the V10 toolkits yet. This does not make us anxious to use V11
right now.
. Interest and activity are clearly focusing on V11; if we start with V10 we
are afraid of being isolated from others doing similar things. Judging by
the net traffic, we would have little opportunity of exchanging ideas, bug
fixes, etc. with others.
Missing information:
We don't have our V11 tape yet, so we don't really know
- How much work it would be to bring it up.
- How broken it *really* is in the areas we need.
- How different the V11 tookit is from any of the V10 toolkits. (i.e. could we
minimize conversion by starting with a particular 10.4 toolkit.)
So, what is everyone else doing? Surely others must be faced with this kind of
a decision. How are you proceeding? A fair amount of the decision will be
based on input from other X users (direct, as a result of this message, and
indirect, from what we see on the net).
jwwohler@milk1.istc.sri.com..istc.sri.com (Bill Wohler) (10/27/87)
Jamie Watson writes: [X10, or X11, that is the question...] >So, what is everyone else doing? Surely others must be faced with this kind of >a decision. How are you proceeding? A fair amount of the decision will be >based on input from other X users (direct, as a result of this message, and >indirect, from what we see on the net). we had to make the same decision. we have decided to go with X11, write our software, and let the window system "catch up" with us. we will be writing software that will be running on suns and hp 330's. thus, i would like to get in contact with hp engineers who are working on the hp x11 server and toolkit port. please reply to wohler@spam.istc.sri.com if you have any pertinent information. --bw