[bionet.software] X on Internet

SBAIRD%UOTTAWA@ACADVM1.UOTTAWA.CA (Stephen Baird) (02/26/91)

I see more and more an interest in molecular biology software on Unix
using X-windows.  The reasons for heading that way have been discussed
on thisboard often.  Molecular biologists have a lot of work to do and
do not want to learn how computer systems work.  They want something
easy to use.  I am going to set up a SPARCstation 2 to be shared by
a few labs around me over an ethernet  network.  Has anyone ever tried
to log in over Internet from an X-terminal or PC/X-terminal and use an
X-window application?  What about a high-speed modem?  Would it work
like a bogged down networkor is there no way to login this way?


      Sincerely,


      Stephen Baird
      Molecular Genetics
      Children's Hospital of Eastern Ontario
      Ottawa, Ont., Canada
      Sbaird@acadvm1.uottawa.ca

JRAMON@ccuam1.uam.es ("J. R. Valverde") (02/26/91)

Hi,
        this may be of not much use to you. And I must admit we still haven't

tried it yet.

        Here, we have purchased the PCSA software from DEC for communication
in our LAN. This software supports connections through modem, LocalTalk and
Ethernet, and provides X-windows client support for micros (PCs and Macs). It
seems that X-client works fine with Ethernet, but we haven't set up yet the
modem support for remote connections, though I bet that trying to run X with
a modem line will be too slow to be useful (maybe at 64000 bauds? can't
believe).

        The bad news is that PCSA runs only under VMS and Ultrix, not on
Sun or other Unix boxes. But maybe, you could use the X-client software
with different underlying comm protocols. Since all we have here as servers
came from DEC I can't say.

        Anyway, theoretically, you should be able to remotely run X programs
with an X-client and an internet comm. package, even through slow speed lines.
However using low speed lines should be of less use.

        Hope this helps.

                J. R. Valverde

------------------------------------------------------------
Do you like Toloose Lautrec?
No Sir, I don't like to loose anything.

spam@iastate.edu (Begley Michael L) (02/27/91)

In article <9102261032.AA22393@genbank.bio.net> JRAMON@ccuam1.uam.es ("J. R. Valverde") writes:

>modem support for remote connections, though I bet that trying to run X with
>a modem line will be too slow to be useful (maybe at 64000 bauds? can't
>believe).

Actually, this isn't completely true...X will be painful but nearly tolerable
at 9600, and it actually passable at 19.2K.  A couple of trailblazer modems
running data compression will actually work reasonably well (assuming good
telephone lines, of course).

-mike begley
--
************************************************************************
* mike begley       "I will not waste network bandwidth"               *
* spam@iastate.edu  "I will not waste network bandwidth"               *
* hzl01@iastate.edu "I will not waste network bandwidth"               *

lear@turbo.bio.net (Eliot) (02/27/91)

spam@iastate.edu (Begley Michael L) writes:

>[...]...X will be painful but nearly tolerable
>at 9600, and it actually passable at 19.2K.  A couple of trailblazer modems
>running data compression will actually work reasonably well (assuming good
>telephone lines, of course).

Actually there is a new tool available called Xremote that distributes
the task of processing X protocol requests between a local UNIX system
and a remote X server (``remote'' being on the other end of some serial
connection; presumably anything acting like a terminal would do).
This is not unlike what Graphon and NCD do to improve serial line
performance.

I'll check into the availability of this software and get back to this
group.
-- 
Eliot Lear
[lear@turbo.bio.net]

lear@turbo.bio.net (Eliot) (02/27/91)

I write:

>Actually there is a new tool available called Xremote that distributes
>the task of processing X protocol requests between a local UNIX system
>and a remote X server (``remote'' being on the other end of some serial
>connection; presumably anything acting like a terminal would do).
>This is not unlike what Graphon and NCD do to improve serial line
>performance.

Because the program documentation I had read was for the client side
of the NCD serial communications protocol.  Thus it requires an NCD on
the other side, so that should answer the question of availability.

Oops..

-- 
Eliot Lear
[lear@turbo.bio.net]