home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.os2.misc
- Path: sparky!uunet!spool.mu.edu!umn.edu!csus.edu!netcom.com!samiam
- From: samiam@netcom.com (Scott Moore)
- Subject: Re: Why not standard VESA
- Message-ID: <1992Dec19.070020.4266@netcom.com>
- Organization: Netcom - Online Communication Services (408 241-9760 guest)
- References: <29437.1088.uupcb@satalink.com>
- Date: Sat, 19 Dec 1992 07:00:20 GMT
- Lines: 120
-
- bert.tyler@satalink.com (Bert Tyler) writes:
-
- >Forgive me for responding to a message here that really has little to
- >do with OS/2, but as a person who occasionally writes VESA TSRs for
- >IBM, I just couldn't let this one go by without a response...
-
- Teed you off eh ? I'll have to point by point this one as well !
-
- >Useless? You have the "right" to make any claim you want, but this
- >one is way off the mark. Virtually every SuperVGA adapter sold today
- >comes with VESA either on the BIOS or as a TSR. IBM ships PS/2s with
- >SuperVGA adapters(*) with MS-DOS SuperVGA modes accessible *only* via a
- >VESA TSR (the ValuePoints ship with video adapters accessible via both
- >VESA and chipset-specific methods.) Many SuperVGA MS-DOS apps
- >(Links386 Pro is an example, Microsoft C 7.0's graphics libraries are
- >another) support SuperVGA *only* for VESA-compliant adapters. Others
- >check for VESA-compliance first, and chipset-specific implementations
- >afterwards (the problem with chipset-specific implementations is that
- >they tend to work with this week's chipsets, but not with the ones that
- >get released next week). Hardly a textbook case of the term "useless".
-
- Yes, adapter support of the VESA standard is basically universal at this point.
- Just as univeral is the software writer's avoidance of it. Many cards also
- come with a TSR that supports the "8514" standard, which if you remeber was the
- all software interface to the 8514 board. Anyone out there using that thing ?
- The important programs have virtually bypassed VESA, such as AUTOCAD and
- windows itself.
-
- >(*) The PS/2 model 25SX, 35, 40, 56, ThinkPad 700C and others, using
- >the PS/2 25SX VESA driver (available in Compuserve's IBMPRO forum in
- >Library 12 (VESA), filename IBMVES.ZIP). There's also a freeware VESA
- >TSR for the XGA adapter (same Compuserve library, filename XGAVES.ZIP).
- >I have seen nothing indicating that IBM's commitment to the VESA standard
- >is dying off.
-
- Since IBM claims to be commited to OS/2, and os/2 has nothing to do with VESA,
- I have to wonder about that. But I don't remeber saying anything about IBM.
-
- >Maybe you meant to say "Useless in terms of GUIS like OS/2 when those
- >GUIs aren't running DOS sessions". That I would agree with at the moment.
-
- >SM> [discussing bank-switching:]
- >SM> Now, I could stop right there, and any programmer worth his salt
- >SM>will tell you that a high speed driver does NOT stop to make an OS call
- >SM>whenever a change of vga memory location is required. I believe that
- >SM>this fact alone has kept the VESA standard from spreading even in DOS mode
-
- >See above. Under MS-DOS, VESA bank-switching is just as fast as
- >chipset-specific bank-switching - in fact, the DOS App just makes a
- >far call to what is, in effect, the VESA driver's implementation of
- >a chipset-specific bank-switching routine. (There's also a VESA BIOS
- >call supporting bank-switching, but any DOS programmer concerned about
- >speed and "worth his salt" avoids that overhead and uses the direct call.)
-
- And there is the core of the matter. That direct call is useless with anything
- but a true blue DOS app. In my 32 bit adapter/extender days (before os/2),
- I was able to call all the VESA functions but that one. That, plus the
- requirement to pass the driver an address to a parameter buffer (which any
- junior os textbook will tell you is trouble) made creating even a SLOW driver
- for VESA impossible. I talked at length with VESA about this problem,
- proposing several solutions (such as providing a 32 bit version of the
- direct call), and was basically told that 32 bit programs were not
- representative of the mainstream. VESA is supposed to plan for the future, not
- intrench the past.
-
- >SM>(for instance, why is there no windows driver for VESA ? windows is
- >SM>perfectly capable of using VESA BIOS calls).
-
- >There is no incentive for a chipset manufacturer to write a Windows
- >(or OS/2) VESA driver, as it would work on his competitors' chipsets
- >as well. Not the cleverest marketing move in the world. A Windows
- >(or OS/2) VESA driver *is* technically possible. Realistically, though,
- >the only vendor with any incentive to build one is the vendor selling
- >the GUI itself (Microsoft for Windows, IBM for OS/2). Also, any
- >protected-mode GUI using the current real-mode VESA standard would
- >suffer performance penalties due to the need to switch to real mode
- >to perform any bank-switching (there is a protected-mode VESA standard
- >in the works, but we're all talking about the real-mode version that's
- >been around for awhile).
-
- Man, you really doged that one. If VESA was so great, microsoft could have
- simply output windows with a VESA driver built in, and that would have
- been the end of it. They did not because it would have been a DOG.
- Do you really think that microsoft would not have avoided the video driver
- wars if they could have ? remember that win 3.0 drivers were hard to get
- as well.
-
- >OS/2 (and Windows NT as well) has the additional problem that most VESA
- >drivers ship today in the form of DOS TSRs, making them unusable for OS/2
- >video drivers. If IBM provided a VESA-based OS/2 video driver and
- >Microsoft provided a VESA-based Windows 3.1 driver, IBM would be in the
- >embarrassing position of having to explain why their VESA driver didn't
- >work on systems which worked with Microsoft's VESA driver. They would
- >have an excellent technical explanation, but it would still not be a good
- >image situation - and image is important.
-
- Your ending wrapup ^^^^^^
-
- Mine:
-
- VESA could have avoided all this nonsense by simply standardizing a single
- bank selection port, that, duuuuhhhhh, selects which video page is active.
- They could have issued that along with the software standard. Yea, the chip
- makers would have screamed. But if enough programs come out to that standard,
- they would have gotten in line rapidly.
- The unfortunate result has been that SVGA goes virtually unsupported in the
- DOS enviornment. Please don't send me a list of VESA programs. Instead,go
- down to your local software store and look at side panels. Then tell me how
- many off the shelf applications support SVGA. Further, of those that do,
- how many list each adapter they support individually ?
- Compare this to the virtual market sweep of VGA and I think I am justified in
- calling the situation sad (but probably not VESA's fault).
-
- [sam]
- --
- Scott A. Moore [SAM] | "Cash is more
- samiam@netcom.com | important than your
- Santa Cruz, CA USA | mother"
- 408-423-1624 | Allan Shugart - CEO Segate Corp.
- ------------------------------------------------------------------------
-