Re: OpenG VISA

[Jean-Pierre D. <jea...@tr...>]

In the light of recent licensing issues, I resurrect this old post from R=
olf.

At Avensys where I work I'm director of software development and in decis=
ion regarding sofware tools we use for R&D and commercial
applications. As much as I love LabVIEW I'm seriously thinking about drop=
ping it for application distribution, keeping it only for
internal sofware development needs.

We develop our own harware. Some are stand alone devices for which we hav=
e LabVIEW GUIs communicating with Serial/Modem/TCP. We also
have general distributed I/O boards (using BACnet protocol mainly used in=
 HVAC) in which LabVIEW applications acts as data
acquisition/logging  servers that serves data to network and Internet Lab=
VIEW clients. As such we're not a great source of revenues
to NI and eventually competitors if they ever get into the BACnet arena.

The prohibitive VISA licensing and the fact that NI can decide anyday tha=
t our applications are compition to their own products and
then against their licensing makes the use of LabIEW "complicated".

Rolf, what is this visa32.dll? It seems to be a NI product so I don't thi=
nk you meant OpenG VISA to be a replacement of NI-VISA. I
would like to have a serial interface to get rid of VISA installation whe=
n I want to use the serial port on target computers. Both
licensing fees and installation footprint make no sense for me for the me=
re use of the serial port.

Like VISA, I 'd like to see a generic communication API where the interfa=
ce would be the same whatever the actual communication
channel is (serial, TCP, etc). It would be OOP so when there is a need fo=
r another communication channel, just provide a specific
implementation that is simply integrated. How difficult would it be to bu=
ild an interface to Windows (or your favorite OS) serial
drivers? Gateways (like serial<->TCP to access serial port on remote comp=
uters) to connect any two different channels could also be
implemented.

Any comment to this?


Jean-Pierre






----- Message d'origine -----=20
De : "Rolf Kalbermatter" <rol...@ci...>
=C0 : <ope...@li...>
Envoy=E9 : 11 novembre, 2003 04:38
Objet : OpenG VISA


> Hello all,
>
> I have been thinking a lot over the weekend about this and I think ther=
e is
> a reasonable approach which might work.
>
> We might be able to create a modularized VISA (the current NI implement=
ation
> is similar in this) which intially might support serial only. The idea =
is to
> separate the implementation of the hardware related interface driver (I=
 think
> NI calls it in their architecture a passport) from the actual VISA32.DL=
L.
>
> LabVIEW and all other software always interfaces with VISA32.DLL only (=
well
> I guess T&M Explorer doesn't comply with this but I have really absolut=
ely
> no intention of supporting T&M in the beginning, both for the fact that=
 the
> interfaces it uses are entirely undocumented as well as possible licens=
e
> issues).
>
> OpenG Visa32.dll would enumerate the underlying passport drivers on fir=
st use
> and manage them accordingly. For the rest it would pass the VISA reques=
ts from
> application space to the according passport driver depending on the act=
ual VISA
> session. It would be a plugin architecture on shared library niveau, wh=
ich is
> also the drawback of this project. Although not inherently very complic=
ated, it
> is something not every programmer is comfortable to work in as it has a=
 serious
> amount of abstraction, so I'm at the moment afraid that there would be =
little
> if no active support from others for this and therefore it would silent=
ly die
> before it has reached any useful stage.
>
> OpenG passport drivers would intially and most probably always NOT be c=
ompatible
> with NI passport drivers as there is no documentation about that interf=
ace,
> although I have a fairly good idea how it works ;-), but the devil lies=
 in the
> detail and therefore I would not want to attempt to achieve compatibili=
ty on
> that level unless NI might at some stage want to help.
>
> As a side node, we might be able to support PORTIO as a passport driver=
 in this
> architecture as VISA has the according register access primitives altho=
ugh it is
> currently only used for PXI and VXI.
>
> Any comments to this?
>
> Rolf Kalbermatter
>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by: ApacheCon 2003,
> 16-19 November in Las Vegas. Learn firsthand the latest
> developments in Apache, PHP, Perl, XML, Java, MySQL,
> WebDAV, and more! http://www.apachecon.com/
> _______________________________________________
> OpenGToolkit-Developers mailing list
> Ope...@li...
> https://lists.sourceforge.net/lists/listinfo/opengtoolkit-developers