[comp.windows.x] Summary of responses

levene@aplpy.jhuapl.edu (Robert A. Levene) (04/08/91)

 The following briefly summarizes my responses to the SLC versus
X Terminal inquiry to this newsgroup.

 I also have on file a 1400-line document of selected responses, which I've
compiled according to respondent and stripped out the header boilerplate and
re-quoting of the inquiry.  Please e-mail me at levene@aplcomm.jhuapl.edu if
you'd like a copy, and if my mbox is flooded, I'll just post it.

My original inquiry was:

>  I'm currently considering the relative merits of purchasing a color
> X-terminal xor a comparably-priced Sun SPARCstation SLC to accompany
> a color SPARCstation.
>
>  If you've had to make this choice, have strong opinions either way, or
> have relative performance benchmarks, please send me e-mail, and if
> there's sufficient interest, I'll post a summary.
>
>  Comments from Sun, NCD, Tek etc. are welcome, provided that you make
> your bias explicit.
>
>  My application is both Graphics- and CPU-intensive, so the big
> question is "Does the supposed performance increase of the SLC
> outweigh its lack of color/greyscale graphics?  Is there a major
> loss when attempting to display color images on it?"


  As it turned out, my colleagues and I put in a request for a Sun IPC for
us, and a Sun SLC for our subcontractor's system. (We'll see if the finance
types permit this...)  We've also hung a dumb ASCII terminal from a serial
port for shell-type operations, as well as a PC for file transfer, since we
can't connect our SPARCstation to any external networks (due to
administrative constraints).  This will permit primary users on the console
and the IPC, who can simultaneously xfer files via the PC or start up a
spare shell on the other terminal.

  Here's how we came to our choice.  While it's not too late to back out,
we think we're on solid ground.  Note that we were impressed by many
of the products on the market, and may well have chosen differently if
the constraints were ever so slightly different.  In fact, a major lesson
is that in hardware decisions such as this, there's a different solution
for every set of narrow constraints.


Regarding Color vs. Grayscale vs. Monochrome
--------------------------------------------

  Several respondents suggested that even though monochrome can technically
work, color for primary use is far superior.  We also expect our application
to require correct use of color, almost to the angry-fruit-salad level.  So
we therefore figured that an IPC would have both speed and color, and our
subcontractor could still use color on their SPARC, as well as experiment
with monochrome implementations on their SLC.


Capability to Handle our Environment
------------------------------------

  We will be running Sun's OpenWindows and foresee problems with using some
X terminals with and without built-in window managers.  If we were using a
more generic X environment, then an X terminal with its own window manager
such would probably be a good solution.  (One respondent chose to "avoid
nasty comments about Sun software" perhaps for this reason :-) )


Resource Utilization
--------------------

  If we were to run a large network, then this would be a major concern to
us, however we expect to be configured as two local machines essentially
sharing one or two large disks.  If this sharing significantly reduces
performance, we'll be looking into a small local disk for swap space and
/tmp, etc.  We expect to be running a larger program on the SPARC and a
smaller program on the IPC.  We had a strong recommendation (outside NCD)
for using NCD X terminals as part of larger network.  There was
high praise for Tek's and Visual's terminals as well.

  Several people noted that there is a threshold load on a diskless
workstation where the benefit of local processing is lost and the
machine becomes a weak X terminal.  We expect to be below that threshold for
the most part, in order to gain the benefit of freeing up CPU cycles for
other tasks.

  Thanks to everyone who responded, and keep pushing the state of the art.

     Robert A. Levene             \ I am the sole legitimate representative 
       "/  //  / /"                \ and policy maker for my race, culture, 
 Bitnet:   RXL1@APLVM.BITNET        \ country, religion, political party    
 Internet: levene@aplcomm.jhuapl.edu \ employer, and extended family.