PAYNE@latlog.UUCP (10/20/87)
There seems to be a slight ambigouity with one of the `fixes' sent out. It regards the fix to server/ddx/mfb/mfbimage.c: The synopsis says that the ZPixmap format was equivalent to XYBitmap but it should be equivalent to XYPixmap but the fix changes XYPixmap for XYBitmap ????
RWS@ZERMATT.LCS.MIT.EDU (Robert Scheifler) (10/20/87)
Date: Thu, 15 Oct 87 10:45:32 -0100
From: mcvax!latlog!PAYNE@uunet.UU.NET
It regards the fix to server/ddx/mfb/mfbimage.c:
The synopsis says that the ZPixmap format was equivalent to XYBitmap but
it should be equivalent to XYPixmap but the fix changes XYPixmap for
XYBitmap ????
An == also changed to !=. There are three formats involved: XYBitmap, XYPixmap,
and ZPixmap. Think about it.spaf@PURDUE.EDU (Gene Spafford) (03/15/88)
I just got mail that indicated I ought to further clarify some things:
I am *NOT* trying to put anyone down about the form & content of the
X11 distribution. Please do not take it personally I do *not* mean to
sound like a chronic, irrational complainer, or worse yet, an ingrate.
I realize that putting the thing together was a massive task, and it is
a product provided basically "as is" to the public.
However, at the same time, I wanted to convey back to you some
observations about trying to install and use the X distribution. It
seems to me that if you want people to use the system on a day-to-day
basis, and if you want people to feel good about using X, it has to be
documented reasonably so that anyone can install it on their system and
run it. If it doesn't install, or they want to use features, they
should be able to either find the answer in the documentation or know
who to contact. Or have I misinterpreted the V11.R2 release and it is
not supposed to be widely installed and used by anyone other than
programmers intending to be X hackers?
I have many, many years of Unix experience. I have considerable
experience with installing and maintaining large software packages on
networks of machines. I work in software testing and configuration
management issues. I've written two operating systems, from boot code
and device drivers on up. All in all, I believe I have some advantage
over a random site admin trying to install all of this. So here I am,
in the entertaining position of putting X11R2 up on the Suns in our
Center and I'm having all these difficulties. (Why am I trying to
install X11 you ask? Well, I'm the lab manager and systems expert for
our NSF Industrial/Academic Software Engineering Center. All of our
industrial affiliates and researchers were interested in how X11 worked
and if it was usable for our projects; I am the "guinea pig" for this, and
submit my interim report next week.)
I am *not* very familiar with X, and I do not intend to learn how to
write widgets or make X library calls -- I doubt if I am the only person
with an interest in installing X but not developing software in it
(yet). And here I run into all these problems on a simple Sun workstation
(and I won't even mention the disaster trying to build on an IBM RT)
which I thought was supposed to be one of the base machines for X. I'm
not even sure how to tell if some of my problems are bugs or if they're
stupid mistakes, or misconceptions because of lack of documentation,
There really isn't a mechanism in place for me to find out --
there's too much of the wrong kind of documentation for me to find the
answers by myself, and there is no "official" support where I can find
the answers (or fixes) externally in a reliable, timely manner. Are
users like me supposed to back off and wait for Release 3? Wait until
the vendors offer it? Just give up?
Outside my evaluation duties, I'd be interested in helping, but what
can I do? I made some suggestions about documentation; I'd offer to do
it myself and contribute it back to you, but if I knew enough to write
the docs, I wouldn't be asking these questions!
I'd offer to beta test to help you find these problems before a
release, but I'm not an X-pert, so you probably never would have
considered someone like me for a Beta test anyhow (which is one thing
wrong with *most* Beta tests, by the way -- lack of trial by naive
users). And anyhow, Release 2 is in the field.
I'd help debug, but the learning curve for some of the code and so on
is too steep for me to be of much use. Plus, I don't have the time or
motivation to learn all of that stuff right now -- academic faculty
around here keep too busy to be able to play with code outside their
own projects.
So what can I do? I can contribute some observations and suggestions
based on professional experience, and hope it is taken under advisement
by all the bright people who put the system together without their
taking it too personally. I can also shut up and go away. I tried
the first approach, and it seems to have generated some resentment, so
I'll use the second approach from now on. My sincerest apologies if
I offended anyone with my mail -- it was unintentional.
Gene Spafford
Dept. of Computer Sciences, Purdue University, W. Lafayette IN 47907-2004
Internet: spaf@cs.purdue.edu uucp: ...!{decwrl,gatech,ucbvax}!purdue!spafRWS@zermatt.lcs.mit.EDU (Robert Scheifler) (03/15/88)
You haven't offended me, and I appreciate the comments. Given your situation, the best you can probably do is to take the time to document the problems you find as well as you can and let us know about them (for example, you had problems building, send us the relevant part of your make log, explain what you did and what failed, etc). I still believe the best place to ask questions is a public mailing list like xpert/comp.windows.x. In doing so, you reach hundreds of people who can probably answer your question (and thousands more who can't but might :-), and hopefully one of them will. Querying just MIT has simple limitations: when thousands of people send mail to just a few people, it will take those few people a *long* time to respond in a reasonable way, and it is almost guaranteed not to be timely. You send in a bug which you think is absolutely critical, but unknown to you a dozen other people have sent in what they think are critical bugs, and meanwhile the MIT staff is currently looking at R&D issues that they view at even higher priority than your bug reports. In the end, the reasonable (but not timely) response is the next MIT release, but of course people want answers and quick fixes sooner than that. We'll do the best we can, but there should be enough Xperts out there now to help spread the load. Different people will view MIT releases in different ways, depending on how much time, money, etc. they have. One view is that they are developer's releases; if you want "creature comforts", you should buy X from a vendor who puts effort into high-quality end-user documentation, Software Support, etc. Others view the releases as inexpensive alternatives to paying Big Bucks for software, and rely on their local expertise and documentation to see them through. The MIT releases are basically provided "as is", with minimal support from MIT, and you have to make up your own mind whether it serves your purposes.