dingler@hc.DSPO.GOV (Robert Dingler) (03/23/88)
My group needs to make a choice between X and NeWS. Does anyone care to express their opinions about each window system to help us make the right decision? Thanks in advance, John Sutton Los Alamos Nat'l Lab Direct replies can be made to: sutton@olin.lanl.gov
bob@allosaur.cis.ohio-state.edu (Bob Sutterfield) (03/23/88)
In article <14208@hc.DSPO.GOV> dingler@hc.DSPO.GOV (Robert Dingler) writes: >My group needs to make a choice between X and NeWS. >Does anyone care to express their opinions about each >window system to help us make the right decision? > ... > John Sutton (sutton@olin.lanl.gov) > Los Alamos Nat'l Lab On my desk is a paper entitled "Using the NeWS Window System in a Cray Environment" by R. L. Phillips and D. W. Forslund at LANL. It seems to be more of a sponsor-sales sort of thing ("look what we did with your money!") rather than presenting the results of any new research. Still, the paper indicates to me that NeWS is pretty well accepted in the LANL environment. You might consider NeWS as the window system of choice if only for interoperability with the other work that's going on there. -=- Bob Sutterfield, Department of Computer and Information Science The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 bob@cis.ohio-state.edu or ...!cbosgd!osu-cis!bob
bob@kazoo.cis.ohio-state.edu (Bob Sutterfield) (03/25/88)
Mr. Phillips, My sincere apologies for any offense caused by my note on Xpert and NeWS-makers regarding the paper you wrote for your Cray User Group meeting. Nothing derogatory was intended by my remarks. I have written similar papers for both "sponsor sales" and user groups, and recognize that the level of detail chosen must be (as ever) appropriate for the purpose and audience. I didn't expect that my remarks would constitute a review of your work; rather more to protect you from a barrage of requests from people expecting a different level of presentation. For the purpose and at the level intended, the paper was very informative. Frankly, the largest intent behind my posting was to persuade the author of the posting to which I responded (and authors of other similar postings) to check around with the local folks, which it seemed he hadn't done yet. For a user asking the questions he was asking, the level of your paper would be appropriate as well. -=- Bob Sutterfield, Department of Computer and Information Science The Ohio State University; 2036 Neil Ave. Columbus OH USA 43210-1277 bob@cis.ohio-state.edu or ...!cbosgd!osu-cis!bob
dingler@hc.DSPO.GOV (Robert Dingler) (04/16/88)
About a month ago I posted a request for opinions about which window system
to use, X or NeWS. My group uses both UNIX (SUNs) and VMS for development and
needs a common window system that is available on both these systems. Both
X and NeWS will be available on these systems in the near future. (Someone,
and they have requested that I not say who just yet, is porting NeWS to VMS.)
My request was just the beginning of a process to evaluate and select a window
system.
Below you will find a summary of the responses I got from my March 22 posting
of "NeWS or X - which to choose?". I thank all of those who took the time to
respond to my request. The informations has helped us get started on our
evaluation process. Clearly, to make a correct choice one must evaluate both
systems and judge which is best based on their own unique requirements.
This we are doing now. If anyone would like to see the complete text of all
the responses please mail your requests to my address below.
John Sutton (sutton@olin.lanl.gov)
Los Alamos Nat'l Lab (505) 667-3623
P.O. Box 1663
MS H820
Los Alamos, NM 87545
"Symbolic representation of quantitative entities is doomed
to its rightful place of minor importance
in a world where flowers and beautiful women abound."
----------------------------------------------------------------------
This document contains a summary of information obtained from USENET articles
about the NeWS and X window systems and responses to my request to USENET
readers for opinions about X and NeWS. The comments summarized here are
condensation of the opinions of the authors of the USENET articles. I make no
claim as to the correctness of their opinions.
The following is simply a tally of the preferences stated in the USENET
articles comparing X and NeWS. This is raw data and does not take into account
how the case was stated.
Pro X Pro NeWS Undecided Info Only
5 11 3 3
This section shows the tally of preferences from those who seemed or were known
to have experience with both window systems and were providing more than just a
"gut feeling" about which window system they preferred.
Pro X Pro NeWS Undecided
2 4 1
What did they say?
This section is a summary of what people are saying about
the two window systems. In a few places I have inserted my own comments, these
will be found in the square brackets []. Note that the number of statement
roughly reflects the preferences listed above.
Pro X
* X has the lead in the market place
* Toolkit lives with client
* X is in the public domain
* Application will be more portable to other machines
* Wider range of platforms to run on
* X is superior in performance
* X will prevail in the short term (1-3 years)
* Fits with existing technology, easy to port,
could get very high performance
Pro NeWS
* Easier to program with NeWS, including shorter codes, lower levels of
frustration, and interactive/object oriented development environment
* NeWS can do non-rectangular windows
* Better use of communication bandwidth
* Toolkit lives with server [a changed widget - changes for all clients]
* NeWS is more hardware independent (arbitrary spatial and color resolution)
* NeWS is technologically superior
* NeWS will soon be part of AT&T's System V for all the world to have
(hence in the public domain)
* NeWS is more powerful
* Can dynamically download application into server
(but this is only occasionally useful)
* NeWS is more widely use at LANL
* NeWS is a better technical solution
* NeWS requires only one font file for each font regardless of size
and orientation
* NeWS is very easy to port
* NeWS has a very clean "virtual event" mechanism
* NeWS will supersede X in the long term due to its more abstract device
with higher functionality and arbitrarily high spatial and color
resolution, its "future-proof"
* NeWS allows you to lay down site rules for scroll bars, windows, icons,
cursors, etc. without having to bow to the design decisions taken about
the server
* NeWS is not slow
General
* The best thing to do is to pick a high level toolkit and ignore what
window system is used, counting on it being ported to new window
systems as they come along
* X is C based and NeWS is Postscript based
[note: NeWS has a C toolkit layer]
* Each has a few unique characteristics outside the vast majority of
qualities they have in common
* Performance is not an issue ... performance can be distributed to the
best place
* X is pixel based, NeWS is stencil/paint based
* X provides a procedural interface, NeWS provides a programming interface
It is my general impression that the major points that were being made in the
articles summarized above are:
* X is currently more widely accepted than NeWS and will probably be the
standard at least in the short term
* NeWS is technically superior to X but current implementations display
poor performance, however, NeWS is expected to become more popular in
the long term.
* The major differences between the two window systems are: (1) where the
division between client and server is made (client does most window
management in X where as NeWS relegates most windowing tasks to the server)
(2) the degree of device independence (NeWS using a more abstract device)
and (3) X provides a protocol interface to the server and NeWS provides
an object oriented programming language interface to the server allowing
dynamic loading of server functions.davidh@dent.Berkeley.EDU (David S. Harrison) (04/16/88)
Last time I saw a message similar to this I posed a few questions
to the comp.windows.news group about operations I feel are essential
for engineering graphics that don't seem to be possible in NeWS.
I have received no responses or feedback. Once again, I would
1. Raster operations other than copy or xor.
2. Color map modifications to directly specify a color.
3. Plane management (masks) for independent control
of multiple planes.
4. Stipple fill of rectangular and non-rectangular regions
on *color* devices (this is NOT the same as tiling with
an or operation). The stipple must be interlocked (relative
to a known point regardless of the location of the
stippled object).
I have written an editor for Computer-Aided Design of Integrated
Circuits under X10 which uses *ALL* of these features. Furthermore,
most CAD of IC graphics editors I know of require most if not
all of these features. As far as I can tell, these editors
cannot run on top of NeWS.
David Harrison
UC Berkeley Electronics Research Lab
(davidh@ic.Berkeley.EDU, ...!ucbvax!ucbcad!davidh)