Re: graphics device driver

From: Fred Kleinsorge (my-last-name_at_stardotzko.dec.com)
Date: 09/19/03 

Date: Fri, 19 Sep 2003 15:17:45 GMT

"Chip Coldwell" <coldwell@frank.harvard.invalid> wrote in message
news:bkf170$rb1$1@news.fas.harvard.edu...
> I found this quote in an article posted to this newsgroup in 1998 by
> Fred Kleinsorge:
>
> You *must* write a device driver, and can't use the Unix one... but
> the latest VMS graphics drivers are written in C, and there is a
> common skeleton that allows the programmer to just write relatively
> O/S independent action routines - and not need to know about how
> all the graphics drivers hook together. So, using the Unix driver
> as an example, hacking up a VMS driver for the card is not too
> difficult.
>
>
http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&selm=35A0A2EE.44
3ED402%40star.enet.dec_nospam.com
>
> Is the "common skeleton" referred to above publicly available, and if
> so where?
>

The driver is the simple part.

Let me respond as I have to most of these types of queries...

OpenVMS does not provide a general toolkit for doing X11 DDX/Driver support.
There is not enough general demand for it that would justify the packaging,
documentation, and ongoing support for it.

If a 3rd party approaches us who wants to develop support for a special
card, we will work with them to provide a development toolkit on a
case-by-case basis. We have done this twice in the past. The first one
developed a fully functional driver/ddx, but in the end decided to use a
supported graphics option. The second is in the process of finishing their
support for a highly specialized large format (2k x 2k +) LCD monitor and
proprietary graphics engine (verrrry cool stuff).

I have not yet been approached by an individual (as opposed to a company)
who wants to make a seious committment to developing support for a graphics
card. I generally don't want to entertain people who are just toying with
the idea, interested in the idea, or thinking about it. It's just a
logistical pain-in-the-*** to carve out a toolkit that does not disclose
any IP that was not strictly derived in-house in VMS and remove any licensed
code (like OpenGL), as well as stripping out the licensed tools (like the C
compiler) - and then packaging it up (it's about a full CD worth of stuff).
Plus trying to explain how to go about doing the development, testing, and
packaging.

In the cases I described above, the first employed a developer with VMS and
VMS graphics knowledge (who needed nothing but the toolkit). The 2nd was a
UNIX shop. and I spent several days helping them get a skeleton driver/ddx
to come up on their card, and a number of hours since then helping and
consulting with them. Both were companies with bonafide interest and
experience.

So to sum it up. All you have to do is convince me that it is worth the
effort. It's not that hard to do -- if you are _really_ serious.