[Info-vax] Could XRDP be the next graphical interface for VMS?

Stephen Hoffman seaohveh at hoffmanlabs.invalid
Sat Mar 21 12:28:41 EDT 2015


On 2015-03-21 14:59:18 +0000, Dirk Munk said:

TL;DR: It'd be nice to have, but as an add-on and not as the sole path 
into an OpenVMS graphical display environment.  Whether RDP solves some 
particular problem here does depends on what problem you think RDP 
might address — and I don't yet understand which problem(s) you are 
seeking to address with RDP.

> Recently there was some discussion about X-Windows, Motif etc. in this group.

There was a discussion of the rather limited graphics driver support 
and the need to deal with graphics hardware — graphics hardware which 
tends to have the product shelf life of a fruit-fly, and where new 
device generations require reworked or wholly new device drivers, and 
which is work that can require potentially-restricted access to 
detailed device interface specifications if and as those are available 
— and also discussion around whether or not certain applications and 
certain applications did or did not require graphical displays — some 
folks here will emphatically never require a graphics display 
associated with OpenVMS — and there was the usual debate and 
disagreement over exactly what a desktop environment is, and what a 
desktop environment might provide folks.  The opinions and requirements 
and potential usages of what various participants refer to as a 
"desktop" do of course vary very widely, and — for clarity — the 
chances of OpenVMS ever hosting what most folks would consider useful 
home or office applications approaches zero.

> I'm not an expert on these matters,

Some of these discussions can require substantial effort to explain the 
associated technical details to an inexperienced or inexpert user, 
which can require a willingness to gather and present the associated 
background information and lints to associated resources and 
discussions.  Time and effort and research.   Put another way, it can 
be much easier to ignore some of these discussions.  The astute reader 
will note that more than various folks familiar with X and RDP 
technologies can and do ignore some of these threads, of course.

> but it doesn't stop me thinking about good solutions for these problems.

Which particular problems would you be proposing "good solutions" for, 
in this case?   Whether inexperienced or expert, some background 
information and a summary can help readers better understand the 
suggestion, or the proposal, or — if you were asking one — the question.

For this RDP discussion, this "good solution" is a reasonable approach 
in support of those Windows clients or other systems that cannot or do 
not have an X Windows server package installed, and that require 
viewing the OpenVMS graphical display from the physical or virtual 
frame buffer.  That's probably not a very big market, right now.

RDP can certainly be handy for some specific application cases and I do 
use it for various operations, but I wouldn't presume providing a 
somewhat better OpenVMS graphics experience for Windows users would be 
at the top of the work list over at VSI — there isn't all that much 
graphical application software for OpenVMS, and what is available can — 
in the absence of RDP — be displayed remotely via X.   Unlike the 
experience and the tools that most Windows or OS X users are presented, 
OpenVMS is still centrally managed from the command line.

OpenVMS on Integrity servers can also have RDP vKVM support via the 
management processor iLO, which means this solution is already 
available for those users — no RDP server needed.  There are also 
various vendors which offer hardware network KVMs 
<http://www.blackbox.com/Store/Results.aspx/KVM/n-4294964384/p-0> — no 
software required.  Not the cheapest way of doing this, but functional 
for those that really need it.

RDP is a reasonable solution that Microsoft uses, and Microsoft uses 
RDP because the Microsoft GUI does not have remote-display 
capabilities.  OS X has similar GUI display limitations, as do some 
other windowing systems.  The old VWS/UIS system was local, as well.   
Both Windows and OS X have functional, free, add-on X Window servers.

Beyond the effects of the network connection and ignoring the lack of a 
viable RDP server for OpenVMS servers lacking vKVM support, local 
experience finds RDP rather slow when working with virtual frame 
buffers.  But I digress.

Note that Windows and OS X still have graphics driver support, and 
extensive application frameworks for accessing and managing the display.

Running DECterm via X or via RDP is reminiscent of hunting fleas with 
anvils.  It works, but...

> But what is your opinion, could XRDP or a similar OpenVMS package be 
> the way for a new graphical interface?

RDP is a software-implemented network-remote KVM.   RDP is useful for 
some cases and a common protocol for an RDP client box that needs to 
access a remote RDP server box, but it is not what most folks would 
consider a "new graphical interface".   RDP is nothing of the sort.  
It's far too primitive to be considered a graphical interface.   Try 
doing some X programming on OpenVMS or Unix (whether Motif or CDE or 
Wayland or otherwise), or Xcode and the now-integrated Interface 
Builder on OS X, or the Visual tools over on Windows, or GTK+, or 
whatever.  These tools can control and manage displays, and 
applications can use these tools for displays.  RDP?  Not so much.  If 
VSI wants to go entirely virtual, then they probably would use RDP — 
but I suspect there'll be customers that want a local graphics display 
of some ilk, which will require a graphics device driver.

Read up on X and Motif and CDE and Wayland and particularly watch the 
Wayland video referenced at <http://labs.hoffmanlabs.com/node/1680> as 
a starting point, and read up on RDP 
<http://en.wikipedia.org/wiki/Remote_Desktop_Protocol>.






-- 
Pure Personal Opinion | HoffmanLabs LLC




More information about the Info-vax mailing list