simnet@ssc-vax.UUCP (Mark R Poulson) (10/31/89)
I'm looking for ways to develop X-Windows application software and run it on an IBM PC or compatible (possibly under the Xenix OS). My final product should also need to be portable e.g. to a SUN. Any idea or comment will be greatly appreciated. A. Hsu
madd@world.std.com (jim frost) (11/12/89)
simnet@ssc-vax.UUCP (Mark R Poulson) writes: >I'm looking for ways to develop X-Windows application software and run it on >an IBM PC or compatible (possibly under the Xenix OS). My final product >should also need to be portable e.g. to a SUN. Any idea or comment will be >greatly appreciated. Well-written X applications are highly portable, so that shouldn't be a major problem. If you make use of UNIX-flavor-dependent functions, such as tty control, it's a little harder. This kind of thing is generally unnecessary under X. If you want to do this, I recommend using Interactive Systems Corp's 386/ix instead of Xenix. They have the best 386 X windows I've ever seen, period. ESIX's X had poor performance relative to Interactive even though they were running on exactly the same hardware. Xenix, last I heard, was still R2. Xenix also uses the flaky Microsoft compiler. Pcc isn't great but it's fairly reliable. If you want to write X applications on a 8086 or 80286 machine, good luck. There are fairly useful servers that run on those machines but no good systems to run servers and clients. jim frost software tool & die "The World" Public Access Unix for the '90s madd@std.com +1 617-739-WRLD 24hrs {3,12,24}00bps
chasm@attctc.Dallas.TX.US (Charles Marslett) (11/14/89)
In article <1989Nov11.175220.25696@world.std.com>, madd@world.std.com (jim frost) writes: > simnet@ssc-vax.UUCP (Mark R Poulson) writes: > >I'm looking for ways to develop X-Windows application software and run it on > >an IBM PC or compatible (possibly under the Xenix OS). My final product > >should also need to be portable e.g. to a SUN. Any idea or comment will be > >greatly appreciated. > > Well-written X applications are highly portable, so that shouldn't be > a major problem. If you make use of UNIX-flavor-dependent functions, > such as tty control, it's a little harder. This kind of thing is > generally unnecessary under X. > > If you want to do this, I recommend using Interactive Systems Corp's > 386/ix instead of Xenix. They have the best 386 X windows I've ever > seen, period. ESIX's X had poor performance relative to Interactive > even though they were running on exactly the same hardware. Xenix, > last I heard, was still R2. Xenix also uses the flaky Microsoft > compiler. Pcc isn't great but it's fairly reliable. As a corollary, I have been trying to get support from any of the Unix vendors for 256-color X-windows drivers. None offered to provide them, but Interactive has been working diligently, and has been testing my code for me against their applications, so I am very pleased with their support for X11. On the other hand, I have tried several times to discuss what I have to do to support X11 in 8-plane mode, and I have yet to talk to the same person twice at SCO (and I haven't really tried to talk to anyone at Everex yet). In general, I think Interactive is much more committed to Xwindows than the other vendors. [And our 256-color 800x600 and 1024x768 drivers will work there, too;^)] > jim frost > software tool & die "The World" Public Access Unix for the '90s > madd@std.com +1 617-739-WRLD 24hrs {3,12,24}00bps =============================================================================== "Those who would sacrifice ** Charles Marslett liberty for security, ** STB Systems, Inc. <-- apply all std. disclaimers deserve neither." ** Wordmark Systems <-- that's just me -- Benjamin Franklin ** chasm\@attctc.dallas.tx.us -------------------------------------------------------------------------------
madd@world.std.com (jim frost) (11/15/89)
chasm@attctc.Dallas.TX.US (Charles Marslett) writes: >In article <1989Nov11.175220.25696@world.std.com>, madd@world.std.com (jim frost) writes: >As a corollary, I have been trying to get support from any of the Unix >vendors for 256-color X-windows drivers. None offered to provide them, >but Interactive has been working diligently, The last time I spoke with the ISC technical people, they stated that they didn't support 8-bit color because it was too slow. Having used a Sun386i with 8-bit color, let me tell you that they're probably right. A lot of optimization needs to be done to make it acceptable. ISC seems to have done fairly well so far, though. >On the other hand, I have tried several times to discuss what I have to do >to support X11 in 8-plane mode, You shouldn't have to do anything special; you'll just have a pallette of 256 colors to work with instead of 16 unless you're doing something like colormap-based double-buffering. If you need fairly well-written color code, I have a bunch of it laying around; one such program is xloadimage, an image display utility, which was recently posted on comp.sources.x if you want to grab it. It has a LOT of color support for n-bit displays. Happy hacking, jim frost software tool & die "The World" Public Access Unix for the '90s madd@std.com +1 617-739-WRLD 24hrs {3,12,24}00bps
chasm@attctc.Dallas.TX.US (Charles Marslett) (11/21/89)
In article <1989Nov15.003835.4889@world.std.com>, madd@world.std.com (jim frost) writes: > chasm@attctc.Dallas.TX.US (Charles Marslett) writes: > >In article <1989Nov11.175220.25696@world.std.com>, madd@world.std.com (jim frost) writes: > >As a corollary, I have been trying to get support from any of the Unix > >vendors for 256-color X-windows drivers. None offered to provide them, > >but Interactive has been working diligently, > > The last time I spoke with the ISC technical people, they stated that > they didn't support 8-bit color because it was too slow. Having used > a Sun386i with 8-bit color, let me tell you that they're probably > right. A lot of optimization needs to be done to make it acceptable. > ISC seems to have done fairly well so far, though. The 256 color server I am currently working on is only a week from being 100% C code. It runs considerably faster than the original 16-color server for the VGA shipped with 386/ix in September. [It also runs a bit slower than the current "soon to be released" 16-color server that it should be compared against.] The screen-to-screen copy (bitblt) runs almost twice as fast as the new 16-bit server though (character output is about a third as fast, to balance things out). I think on average the VGA 256-color X11 server should run "almost" as fast as the 16-color server. And the individuals at ISC who have disagreed from the first simply do not "know" the VGA and its architecture [I DO!]. Though the 16-color mode often has twice the bandwidth of 256-color modes, and the 256-color modes have to shove twice the bits around, they involve lots of I/O (and in protected mode 286, and especially 386, code that slows everything down). I still have to stand by my statement that they are helping me a lot more than the competition (unnamed ;^). Even if they, generally, don't believe it can be done! > >On the other hand, I have tried several times to discuss what I have to do > >to support X11 in 8-plane mode, > > You shouldn't have to do anything special; you'll just have a pallette > of 256 colors to work with instead of 16 unless you're doing something > like colormap-based double-buffering. No . . ., try looking at the way the video memory is organized in the VGA 16- color modes, and how it is organized in the 256-color modes. An efficient server for one set of modes is not likely to be well organized for the other. Specifically, one is a heavy user of I/O instructions, the other cannot usefully take advantage of the VGA's internal latches. ETC. > Happy hacking, > > jim frost > software tool & die "The World" Public Access Unix for the '90s > madd@std.com +1 617-739-WRLD 24hrs {3,12,24}00bps Charles Marslett [a bad boy, at interactive now, since they told me to shut up, and <since I don't work for them, and since I like to shoot off my mouth> I ain't] chasm@attctc.Dallas.TX.US (or some other reasonable facsimile)
madd@world.std.com (jim frost) (11/22/89)
chasm@attctc.Dallas.TX.US (Charles Marslett) writes: >> >On the other hand, I have tried several times to discuss what I have to do >> >to support X11 in 8-plane mode, >> >> You shouldn't have to do anything special; you'll just have a pallette >> of 256 colors to work with instead of 16 unless you're doing something >> like colormap-based double-buffering. >No . . ., try looking at the way the video memory is organized in the VGA 16- >color modes, and how it is organized in the 256-color modes. I thought he was asking about client-side, not server side. If this is not the case, I apologize. jim frost madd@std.com