km@mathcs.emory.edu (Ken Mandelberg) (05/03/90)
In theory, X provides a framework to mix clients distributed from many diverse hardware/software platforms on a single desktop. In practice, almost all the clients I know are programs that run on Unix or were ported from Unix clients. For an X user on a potent Unix workstation, the advantages of using X over a framebuffer based window system with necessarily local clients, are fairly subtle. X lets you use the CPU power of even more potent Unix like hosts, use several CPUS at once, and run software not ported or licensed to your particular workstation. On the other hand X consumes more resources and has generally poorer performance on local clients than those running on a framebuffer based window system. I think it would be easier to sell people on X, if there were actually significant X clients, which are not ports or cousins of Unix clients. For example, I see no reason why a Mac or DOS program which was written in its host assembler with dependencies on the host OS, could not be converted into an X client (although the graphics host dependencies may in fact be exactly the hardest part to change). I suppose since DecWindows is a strategic VMS windowing product, there may be VMS X-clients that are not directly related to Unix clients. I'm not familiar enough with VMS offerings to know. So now to the question. Is there a signficant set of X-clients that run somewhere other than a Unix platform, that offer applications not otherwise available on Unix? -- Ken Mandelberg | km@mathcs.emory.edu PREFERRED Emory University | {decvax,gatech}!emory!km UUCP Dept of Math and CS | km@emory.bitnet NON-DOMAIN BITNET Atlanta, GA 30322 | Phone: (404) 727-7963
JONESD@kcgl1.eng.ohio-state.edu (David Jones) (05/03/90)
I don't see what your point is. It seems like you want evidence to support or refute premise 3 of the argument: since: 1. The advantages of distributed applications are not easily perceived. 2. X applications run much faster if the X is replaced by a window system optimized for access to local hardware. 3. All X applications can run on my workstation using a different window system because the application was written under unix and my workstation runs unix. therefore: Its hard to justify writing X-based applications because X overhead degrades the quality of the application while apparently gaining little. The problem with premise 3 is that it assumes that the framebuffer window system on all hardware made by different manufacturers will be the same just because they are running unix. You need to define a standard that people can conform all their applications to. I think premise 1 will become less valid over time as people's work habits change. As for X-applications with a non-unix origin, VMS has a few pretty slick applications. The bookreader application has no widely known analogue under unix. There are also DECwindows front ends for the symbolic debugger and the performance coverage analyzer, which are VMS specific. David L. Jones | Phone: (614) 292-6929 Ohio State Unviversity | Internet: 1971 Neil Ave. Rm. 406 | jonesd@kcgl1.eng.ohio-state.edu Columbus, OH 43210 | jones-d@eng.ohio-state.edu Disclaimer: A repudiation of a claim.
mouse@LARRY.MCRCIM.MCGILL.EDU (der Mouse) (05/05/90)
> In theory, X provides a framework to mix clients distributed from > many diverse hardware/software platforms on a single desktop. In > practice, almost all the clients I know are programs that run on Unix > or were ported from Unix clients. Me too - but that doesn't prove anything; the same is true (in my case, at least) if we don't restrict the discussion to X programs. > For an X user on a potent Unix workstation, the advantages of using X > over a framebuffer based window system with necessarily local > clients, are fairly subtle. Not for me. I use X instead of Suntools on a Sun, for example, simply because I can get the look&feel[%] I want from X but not from Suntools. (From what little I know of sunwindows, this is primarily a documentation problem, but it's no less real for that.) Network transparency is (just) a nice side benefit. [%] I don't like the term, but it's annoyingly accurate for what I want to say. > On the other hand X consumes more resources and has generally poorer > performance on local clients than those running on a framebuffer > based window system. But if comparable clients are not available? I had to write my own clients to get the environment I wanted under X. I would have done the same with sunwindows, except that there is no documentation I could find for the low levels necessary, and X came along before I got fed up enough to go grubbing around and figuring things out on my own. > So now to the question. Is there a signficant set of X-clients that > run somewhere other than a Unix platform, that offer applications not > otherwise available on Unix? Either "yes but it doesn't mean anything" or "no", because if a client shows up for some other windowing system that's worth having, someone will soon port it to UNIX (or clone it, if its supplier is being too pigheaded about source to make a port practical). (The reason for the first answer I mentioned is that I can reinterpret the wording of your question to get a slightly different question.) But that's all my opinion; what do I know anyway... der Mouse old: mcgill-vision!mouse new: mouse@larry.mcrcim.mcgill.edu