Unix was designed by programmers, for programmers. You _are_ right that it's esoteric. (It's safe to assume that your comments on Bourne are, well, inaccurate. Have you ever seen the guy's biography?)
It takes time. You have to approach it little by little. I've been using it for 10 years, and it _still_ bears new things for me. This is an advantage: it allows you to grow. My recommendation is for you to never learn it; the ensuing love affair will make the PC-heads look like the 'pod people' in Invasion of the Body Snatchers!:-)
P.S. vi does suck. I think it was written to scare people away from unix.
The real problem is with software as a whole. There is no emphasis on doing lean designs, just quick implementations to beat time-to-market. We're all riding on the back of hardware -- either more &/or faster hw. (And hw is riding on the back of fab -- but that's another topic.)
The line between policy and mechanism _is_ what makes a lean, flexible system.
If the tangled mess that is an M$OS, is desirable to anyone, then you have to go back to university. If component 'programming' is desirable to you, then you've probably been raised in the 'e-business' world.
The absence of libraries is due to C not having a large standard library. The absence of large pre-packaged commercial libraries/tools, is due to UNIX's current lack of commercial success.
The supposed 're-use' which OOD or component programming allow, results in bloated apps which survive only thru fater h/w. While the initial creation is easier, the maintenance and performance are not. I'm much more for Niklaus Wirth's "A Plea for Lean Software" than for Miguel's yearning to develop for M$.
But the real issue is something else. All these developers who complain about unix, are fooling themselves if they think the criticism will lead to innovation. At a time when *nix is near dead (yup even linux), such irresponsible nit-picking will lead only to the homogenisation of platforms... and nothing like the revival of systems research. plan9 is not going to be your next OS; some M$ bloatOS will be.
The OS should protect the system from a stray app.
Also, I've been 'developing' on NT for several years now. It goes blue-screen fortnightly. If you don't do serious work on it, or if you reboot it every week, then you'll never see its mis-management of virtual memeory bring it down to a blue death.
I'm tired of these developers criticizing unix because their own development work is tough. We all know unix doesn't yet have what-comes-with-commercial-success.
The reason why M$ is easier to develop for is twofold:
1. there's a lot of money to be made, and so a lot of tools and libraries have been written for it;
2. the 'programming' has been reduced to monkey work. This is not an OS that's evolving in academia; its workforce are flavour-of-the-moment narrow-skill-set programmers.
"Most folks" is right. But they have no clue how to determine the root cause of speed/responsiveness. I don't mind; I let them spend the gobs of cash on the h/w & BW that they do.
But our job is different. Objective analysis is what I expected here not lashing out at this and that. Disciplined performance analysis is our job. People who judge the interface, the underlying s/w, the underlying OS, the underlying H/W, the CPU, the billions of transistors, the fab techniques, the nanosecond time-scales, the underlying network, the backbones, and the peers mirroring them... by casual observation, rely (whether they realize/admit it or not) on us to make the right choices.
Seen in this light, such critiques of X fail miserably.
Re:Uhh. . .pass the crack pipe this way. . .
on
X Windows Must Die!
·
· Score: 1
Much of this is a rant about various utills. I could not care less! X is customizable; people who argue otherwise are inept. X is insecure; correct. X's device in/dependence needs work, but is still better than anything that's viably out there. The UI discussions are a yawn.
Where are the discussions of tradeoffs when designing these things. An engineering trade-off is in every design.
As for the architecure, I've read his arguments for NeWS before. NeWS being a distributed windowing system, I have no fundamental arguments against it -- I was equally happy using NeWS as I am X. The tradeoffs (which are not discussed on this site) are not sufficient to warrant the huge shift that a move from X to NeWS represent. This is particularly so considering that the battle-lines (so to speak) are against the PC world. (See most other comments in this discussion.)
UI rant -- at best. Note the unix-haters/ directory. This falls under unix.
It's sad how often engineers/developers don't understand modular/hierarchical design. I once logged in as a different user (used a different wm) on a Sun ('andromeda'), and I had to spend 5 minutes reassuring this other engineer that we were *actually* on the same machine! He kept on saying, 'log into andromeda'.
They are _all_ replacable (even xinit), and not integral to X.
All of these are from the "unix-haters handbook" site. Great choice! How telling!
Re:Uhh. . .pass the crack pipe this way. . .
on
X Windows Must Die!
·
· Score: 1
I am commenting on OSs, not X -- notwithstanding the misnomer (methinks) of 'NOS' being applied to it. As to your comments:
Many developers still need to work with multiple boxes. Remote display (implemented in X, NeWS, etc) caters to this. I see PC developrs _walk_ from box to box whereas I control many boxes/apps from one spot. And I frequently work 'in' those labs *from home*!! PC folks simply cannot do this with the same ease & power. That "the vast majority" of X users do not utilize its power, is a side-effect of the fact that an increasing number of developers have known only PCs.
Display API is only one part of the X stack, and a replacable one at that (to wit: Motif, Athena). The decision to take these into h/w, is a general one and does not reflect on the particular API involved. To wit: openGL's implementation in h/w.
OSs do need to be done clean-sheet every once in a long while. I totally agree, and was accordingly concerned even _before_ I read Rob Pike's comments. But my point is, once that is done, it will take years for that implementation to reach the stability and robustness that was designed into it.
The comments I've read today are mostly of a general nature. Contrary to people's expressed emotions, they are not specific to, or often even _applicable_ to the architecture of, X.
Re:We Have X Because Sun Wanted to Keep Da Goodies
on
X Windows Must Die!
·
· Score: 1
I read something along these lines by one of the original designers of NeWS (forgot his name). Many companies have been releasing their less-than-successful projects under an open src license. When Plan9 was released this way, I wondered if Sun would do so for NeWS. That cannot be a bad thing for Sun/NeWS.
Re:Uhh. . .pass the crack pipe this way. . .
on
X Windows Must Die!
·
· Score: 1
In the world of OSs, new is not necessarily equal to good. Most engineers do not understand this: an OS is not like an application. It's a product of research, and it's of such sheer complexity that you cannot just develop-it-and-put-it-out. An OS *matures*, and can do so like a good wine (if the foundation's design was good). It takes years and years for it to emerge into the intended goal. Applications do not mature.
Re:Uhh. . .pass the crack pipe this way. . .
on
X Windows Must Die!
·
· Score: 1
Hear, hear! I'm yet to hear *one*/. opinion (or on any other website) that dissects X, demonstrates understanding of what's involved, and shows what needs to get done, and what the engineering tradeoffs are.
Every once in a while/. puts on articles which were clearly meant to agitate. These often come from PC sites which just want to increase their hit counts by agitating folks. This is no technical critique of a network-transparent windowing system, it's akin to the incoherent rantings of a petulant child, masked only by the sprinkling of half-baked techy tastes & opinions. Unfortunately, an increasing number of developers have emerged knowing only PCs. They use UNIX networks like they would their PCs, not utilizing much of the power that X provides. I see these people every day manually doing what UNIX/X can handle for them beautifully. The Digital Libraries of IEEE and ACM provide proper analyses of these topics. But reading them would require patience & thinking; it's so much easier to just lash out incoherently.
The Net is a _medium_, just like books is a medium and movies is a medium and radio is a medium. A child raised soley on books, will be quite different from one raised on television alone. The Net, particularly at its current state of human usage, allows for _specific_ expressions of humanity. Even at a (later) point when equal access is available for all to the Net (thus forming a more representative reflection of humanity), the AI's development will have this further factor contributing to it. That is, it will not develop as it would if it had access to eg all printed media.
Interpreted languages are _justly_ becoming more prevalent. But the higher you go, the less flexibility you have. The same data structures or mechanisms may not be available. This is one cause of the bloat we see in much software nowadays, and is a down-side of what is, otherwise, a positive trend.
I think this is just something M$ would like to fan the flames of... merely to help them in their negotiations with the US government. PS It wouldn't be bad if M$ got kicked out of Canada altogether. True North, Strong and Linux-friendly.
I did think about this as well. Though not an ideal situation (US has plenty of bilateral agreements with Kanucks, not to mention errr 'influence'), the Charter of Rights and Freedoms may prove more powerful. Copyright protection will still hold, but free speech may have better protection.
Slashdot Mirror
It takes time. You have to approach it little by little. I've been using it for 10 years, and it _still_ bears new things for me. This is an advantage: it allows you to grow. My recommendation is for you to never learn it; the ensuing love affair will make the PC-heads look like the 'pod people' in Invasion of the Body Snatchers! :-)
P.S. vi does suck. I think it was written to scare people away from unix.
The real problem is with software as a whole. There is no emphasis on doing lean designs, just quick implementations to beat time-to-market. We're all riding on the back of hardware -- either more &/or faster hw. (And hw is riding on the back of fab -- but that's another topic.)
If the tangled mess that is an M$OS, is desirable to anyone, then you have to go back to university. If component 'programming' is desirable to you, then you've probably been raised in the 'e-business' world.
The absence of libraries is due to C not having a large standard library. The absence of large pre-packaged commercial libraries/tools, is due to UNIX's current lack of commercial success.
The supposed 're-use' which OOD or component programming allow, results in bloated apps which survive only thru fater h/w. While the initial creation is easier, the maintenance and performance are not. I'm much more for Niklaus Wirth's "A Plea for Lean Software" than for Miguel's yearning to develop for M$.
But the real issue is something else. All these developers who complain about unix, are fooling themselves if they think the criticism will lead to innovation. At a time when *nix is near dead (yup even linux), such irresponsible nit-picking will lead only to the homogenisation of platforms ... and nothing like the revival of systems research. plan9 is not going to be your next OS; some M$ bloatOS will be.
Also, I've been 'developing' on NT for several years now. It goes blue-screen fortnightly. If you don't do serious work on it, or if you reboot it every week, then you'll never see its mis-management of virtual memeory bring it down to a blue death.
I'm tired of these developers criticizing unix because their own development work is tough. We all know unix doesn't yet have what-comes-with-commercial-success.
1. there's a lot of money to be made, and so a lot of tools and libraries have been written for it;
2. the 'programming' has been reduced to monkey work. This is not an OS that's evolving in academia; its workforce are flavour-of-the-moment narrow-skill-set programmers.
But our job is different. Objective analysis is what I expected here not lashing out at this and that. Disciplined performance analysis is our job. People who judge the interface, the underlying s/w, the underlying OS, the underlying H/W, the CPU, the billions of transistors, the fab techniques, the nanosecond time-scales, the underlying network, the backbones, and the peers mirroring them ... by casual observation, rely (whether they realize/admit it or not) on us to make the right choices.
Seen in this light, such critiques of X fail miserably.
The X-Windows Disaster
Much of this is a rant about various utills. I could not care less! X is customizable; people who argue otherwise are inept. X is insecure; correct. X's device in/dependence needs work, but is still better than anything that's viably out there. The UI discussions are a yawn.
Where are the discussions of tradeoffs when designing these things. An engineering trade-off is in every design.
As for the architecure, I've read his arguments for NeWS before. NeWS being a distributed windowing system, I have no fundamental arguments against it -- I was equally happy using NeWS as I am X. The tradeoffs (which are not discussed on this site) are not sufficient to warrant the huge shift that a move from X to NeWS represent. This is particularly so considering that the battle-lines (so to speak) are against the PC world. (See most other comments in this discussion.)
What happens to XCalc when you resize it too many times?
Trivial. Most X apps handle resizes very well. I'd talk to the xcalc developers if you often need to resize calculators!
Official Notice, Post Immediately!
venting!
The ICCCM Sucks
I sometimes come across people who criticize X window managers. It's always fun to watch them at their own desk ... moving at a snail's pace.
As for X, if you don't like your wm, change it! It lets you! ;-)
jojo on UI
UI rant -- at best. Note the unix-haters/ directory. This falls under unix.
It's sad how often engineers/developers don't understand modular/hierarchical design. I once logged in as a different user (used a different wm) on a Sun ('andromeda'), and I had to spend 5 minutes reassuring this other engineer that we were *actually* on the same machine! He kept on saying, 'log into andromeda'.
Motif Angst Page
I'm getting tired of this. IT'S A GRAPHICS LIBRARY, STUPID!
XBugTool Horror Stories
It's an app!
Here are more apps/utils: xinit, xterm, fvwm, ....
They are _all_ replacable (even xinit), and not integral to X.
All of these are from the "unix-haters handbook" site. Great choice! How telling!
Many developers still need to work with multiple boxes. Remote display (implemented in X, NeWS, etc) caters to this. I see PC developrs _walk_ from box to box whereas I control many boxes/apps from one spot. And I frequently work 'in' those labs *from home*!! PC folks simply cannot do this with the same ease & power. That "the vast majority" of X users do not utilize its power, is a side-effect of the fact that an increasing number of developers have known only PCs.
Display API is only one part of the X stack, and a replacable one at that (to wit: Motif, Athena). The decision to take these into h/w, is a general one and does not reflect on the particular API involved. To wit: openGL's implementation in h/w.
OSs do need to be done clean-sheet every once in a long while. I totally agree, and was accordingly concerned even _before_ I read Rob Pike's comments. But my point is, once that is done, it will take years for that implementation to reach the stability and robustness that was designed into it.
The comments I've read today are mostly of a general nature. Contrary to people's expressed emotions, they are not specific to, or often even _applicable_ to the architecture of, X.
I read something along these lines by one of the original designers of NeWS (forgot his name). Many companies have been releasing their less-than-successful projects under an open src license. When Plan9 was released this way, I wondered if Sun would do so for NeWS. That cannot be a bad thing for Sun/NeWS.
In the world of OSs, new is not necessarily equal to good. Most engineers do not understand this: an OS is not like an application. It's a product of research, and it's of such sheer complexity that you cannot just develop-it-and-put-it-out. An OS *matures*, and can do so like a good wine (if the foundation's design was good). It takes years and years for it to emerge into the intended goal. Applications do not mature.
Hear, hear! I'm yet to hear *one* /. opinion (or on any other website) that dissects X, demonstrates understanding of what's involved, and shows what needs to get done, and what the engineering tradeoffs are.
Every once in a while /. puts on articles which were clearly meant to agitate. These often come from PC sites which just want to increase their hit counts by agitating folks. This is no technical critique of a network-transparent windowing system, it's akin to the incoherent rantings of a petulant child, masked only by the sprinkling of half-baked techy tastes & opinions. Unfortunately, an increasing number of developers have emerged knowing only PCs. They use UNIX networks like they would their PCs, not utilizing much of the power that X provides. I see these people every day manually doing what UNIX/X can handle for them beautifully. The Digital Libraries of IEEE and ACM provide proper analyses of these topics. But reading them would require patience & thinking; it's so much easier to just lash out incoherently.
The Net is a _medium_, just like books is a medium and movies is a medium and radio is a medium. A child raised soley on books, will be quite different from one raised on television alone. The Net, particularly at its current state of human usage, allows for _specific_ expressions of humanity. Even at a (later) point when equal access is available for all to the Net (thus forming a more representative reflection of humanity), the AI's development will have this further factor contributing to it. That is, it will not develop as it would if it had access to eg all printed media.
Interpreted languages are _justly_ becoming more prevalent. But the higher you go, the less flexibility you have. The same data structures or mechanisms may not be available. This is one cause of the bloat we see in much software nowadays, and is a down-side of what is, otherwise, a positive trend.
Where was this obtained from? It just seems too dubious a presentation! Neither email, nor a proper document ready for publishing!
I think this is just something M$ would like to fan the flames of ... merely to help them in their negotiations with the US government. PS It wouldn't be bad if M$ got kicked out of Canada altogether. True North, Strong and Linux-friendly.
I did think about this as well. Though not an ideal situation (US has plenty of bilateral agreements with Kanucks, not to mention errr 'influence'), the Charter of Rights and Freedoms may prove more powerful. Copyright protection will still hold, but free speech may have better protection.