X-windows

Tim Cutts tjrc1 at cus.cam.ac.uk
Wed Nov 23 05:25:01 EST 1994


nash at nrcbsa.bio.nrc.ca (John Nash) writes:

>In article <1994Nov22.142022.2232 at nomina.lu.se>,
>Tomas Johansson <tomas.johansson at biokem.su.se> wrote:
>>Is there anyone how knows if there is a free software to emulate X-windows
>>on a PC running Windows for Workgroups and Winsock. We just received the
>>GCG Software pack v8 running in UNIX which I would like to use with all its 
>>new features.
>>
>>Sincerely, Tomas

>G'day,

>If you're running Winsock, you may want to take a look at Starnet's
>Xwin package.  It is very reasonably priced (around Can$100) and very
>stable.  A free demo version, which is fully functional is available
>by anonymous ftp from ftp.cica.indiana.edu /pc/win3/demo as
>xwindemo.zip.  The demo has a restriction that only one copy per
>subnet can be run at a time.

>If you like it, dick at starnet.com would be pleased to help you with
>purchasing details.  They have a nice DOS-based X-server also
>(micro-X-enlite).

>Disclaimer: I have no association with Starnet, I just like their
>program and their style... they answered my questions about their
>demo, before I bought a copy.

I have just downloaded this X server to give it a workout, and compare it with 
the Linux X11R6 server I use most of the time.

So just to start with, here's the hardware and software setup in question:

Hardware:

486DX-33
20Mb RAM
S3801 graphics, with 2Mb RAM.

Software (for MicroX):

OS/2 Warp, or Windows 3.1, Trumpet WinSock

(for X11R6):

Linux 1.1.62, XFree86 3.1

Speed:

The Linux server is interactively much much faster than MicroX, but this is to 
be expected; it is a 32-bit program running under a 32-bit operating system.

Colour support:

MicroX should in theory support any Windows driver.  XFree86 supports most 
common chipsets, in 256 colours, and a subset in 65,535 or 16,000,000 colours.

My experiences:  Using a 65,000 colour driver under OS/2, MicroX displayed 
most things correctly, although NCSA Mosaic running on a Sun could not display 
any inline images properly at all.  Using a 256 colour driver and seamless 
Windows, everything displayed as black (not very useful).  Under full screen 
Windows, this was not a problem.  I found the same results running under plain 
DOS/Windows 3.1, so this is not an OS/2 problem.

Linux' 256 colour support is flawless.  65535 colour support works for the most 
part, but Mosaic displays images with a green tint (since the RGB values are 
coded as 5 bits Red, 6 bits Green, 5 bits Blue).  It's better than MicroX' no 
image at all, but is probably Mosaic's problem anyway.

Other visuals:

MicroX displayed a lot of video corruption of the root window on my machine, 
especially when transferring focus from the X window to another application.  
This may not be MicroX' fault.  I am tempted to blame the screen driver (some 
other apps have problems occasionally).

Of course, the ability to cut and paste between X and Windows apps is a major 
point in MicroX' favour, but Linux can write to MS-DOS disks, so you're not 
completely sunk for data transfer, although it's hardly as good as using the 
clipboard.

One major problem is that any application making use of the Xpm library does 
not display properly on the 256 or 65535 colour drivers, so there is some 
problem in MicroX there.

Robustness:

MicroX has not crashed on me yet (though I've only been using it a couple of 
hours).  Unfortunately (and this I think is quite serious), it doesn't run all 
X apps properly.  For example, emacs 19 will not run at all, and quits with the 
following error:

/a/bootes-intracus/home_3/tjrc1$ emacs &
[1] 25141
/a/bootes-intracus/home_3/tjrc1$ X protocol error: BadAtom (invalid Atom 
parameter) on protocol request 17

The other problems I could live with.  This one is very serious, and really 
would prevent me from purchasing MicroX.  If its implementation of Xlib is this 
flakey, I wouldn't touch it.  I know its competitors, such as eXceed, do not 
have problems in this regard.

eXceed for Windows is a rather more polished product, especially in the 
configuration department, and if you want a Windows X server then I'd recommend 
that one.  If you're willing to try Linux and XFree86, then I recommend that; 
it is genuine MIT X (XFree86 is part of the X Consortium), and is far faster 
and more robust than either Windows server that I have tried.

That's my summary...

Take your pick!  Oh, and feel free to e-mail me with your comments.

Regards,

Tim.
-- 
-------------------------------------------------------------------------------
194, Vinery Rd, Cambridge, CB1 3DS, UK (+44) 1223 572622
http://cyclin.zoo.cam.ac.uk/
-------------------------------------------------------------------------------




More information about the Bio-soft mailing list