The KAI contribution is only relevant to C++. For straight C it doesn't help. For C++, as you said, it was very good, particularly at optimizing template code and generally getting rid of abstraction layer penalties.
This is the reputation that Gcc has, but it's not quite true anymore.
Gcc improves all the time. They have been using the SPEC benchmarks to monitor performance at every build for quite some time now, and it shows. There are commercial players behing Gcc (and Apple is one of them), for whom performance is really an issue.
The numbers for recent versions of GCC aren't that bad, really. In an objective comparison you will find that Gcc sometime yields faster code than the native compiler of choice. This has been true for me on Alphas, PPCs, Sparcs and x86. Sometimes this is not true and the native compiler produces significantly better code, but this only points out where things can be improved in Gcc. There is nothing set in concrete there.
I don't think that there are too many cooks. Each of them can focus on a particular back end.
Also I've consistently found native compiler buggier. As recently as yesterday the Intel Compiler (latest version: 7.1) could not compile tth, on Linux RedHat 7.3. It does after allocating all the available memory in the system (tth is a very long single-file program). Try it yourself if you don't believe me.
Finally Gcc is the cornerstone of the Free Software movement. Without it no Linux, no *BSD, nothing. It's probably the most tested and solid compiler on the face of the Earth.
I can't believe this got moderated insightful. Without 64bit you are stuck with a 2GB (possibly 3GB) limit on memory for every process. Swapping has nothing to do with it.
Where I work people are running processes that require well above 3GB or RAM. Those don't run at all on 32-bit machines.
Next year or the year after that *you* will be crying for 4GB of RAM, and you will need it. 32-bit computing is as good as dead, it's only a matter of time (and a short amount at that). The only thing that will maintain it is Intel's insistence that you don't need it, like they did with the 80x86 line with 32-bit vs. 16-bit.
Well the killing statistic is the Google access rank. Only 1% of requests to Google come from Linux, see Zeitgeist.
From that one statistic it would appear that the number of Linux desktops is in fact overestimated, which is very depressing. Unless you have a good explanation of why people would install Linux on their desktop but then choose to browse from a Windows box.
From that same site the number of Apple-based request is 3 times that of Linux. Linux is not likely to pass over Apple as a whole anytime soon.
Well that's the whole point that I hoped the lawsuit against the author of DeCSS would bring out. What have we bought when we buy a DVD?
Clearly the DVD content providers want you to buy the disk, *plus* an approved player with attached royalties, *and* a copy for each region you want to play the disk in.
Movie studios don't want you to do your own imports. They want to control the distribution of films around the world. They want to be free to delay the distribution of a film, say in Australia, for several years. At the same time they don't want to impede the sales of DVDs of the same film, say in the USA. They don't want people in Australia to be able to view the film in their home theatre before it's distributed there.
They do not want you to view your paid-for films from your hard disk. They don't want you to make backups of their content, they want you to pay again if your disk gets damaged somehow.
In the future they want to be able to introduce DVD media that become unplayable after a few days or a number of runs.
All the above is made impossible by DeCSS. Movie studios have only themselves to blame. They tried to be too greedy and it backfired on them.
However don't be too optimistic about movie studios learning a lesson from this. The next digital format to come along will have strong encryption.
I'm not the only one having false memories, then. Bill Gates actually said that about Windows NT 3.x.
Do a Google on 'Microsoft Windows "better Unix than Unix"', I got 7 pages of people remembering the exact same quote, from an add; here is an exerpt.
> One of the things that upsets us about Microsoft > is that they depend a lot on what seems to be > boastful marketing and sometimes, in their race > to catch up with Unix, lets quality slip. A > colleague of mine wrote: "It was in 1993 I > believe when I saw a full page ad for Windows NT > 3.1 with the title "A Better Unix Than Unix." I > sure wish I'd kept it." >
Actually, yes I have survived a project like that but it wasn't pretty. I was 22 at the time and I had 1h commute each way on top of the 12h days. I was absolutely 100% a wreck after 3 weeks of that, and the code I produced was a complete mess and had to be re-implemented after. It was all x86 assembly and you know how easy this is easy to read after the fact (it was more than 10 years ago, and a standard when writing device drivers then).
We met the deadline and delivered the product, which was promptly shelved by the client, never to be used until 6 months later. It turned out the deadline was completely artificial and had been used in an internal warfare in the client company. One of the executive was touting this project as infeasible and another decided to `prove him wrong' using our services rather than their own internal resources.
I resigned shortly after hearing of this, both the client and my ex company where history about 18 months later. Actually the client company was quite big and politically important at the time, and the collapse resulted in a spate of lawsuits. The CEO of that company was in real trouble (at some point he was looking at a prison term), and my own boss lost his company and was actually forbidden to run one for quite a long time.
This was my very first real job as a software engineer. I went back to graduate school after that.
My experience is telling me loud and clear that such projects are the hallmark of terrible, in fact destructive management and should be avoided at all costs.
> Home Use Desktop: Microsoft OSes are fantastic as > home desktop OSes, they beat Linux in 95% of home > use situations. The only exceptions in favor of > Linux is non-game, surf and email purposes and > people who code as a hobby.
You do pay for that home use software, don't you? Myself I don't have deep enough pockets and most of the free equivalent stuff does not run on Windows at all.
The only things I used to use Windows for was for (properly paid for) games but with NWN I'm more than satisfied now.
There is a major difference. In IE's case it only works on one platform (initially 2) and cannot be made to evolve by a third party. With all open-source code even if the initial developer is no longer interested in the project, others can take it on and make sure it's not dead technology.
In IE vs. Netscape the objective was to cut Netscape's "air supply" (the Microsoft memos and e-mails said so). In the OSS case the objective is to share the technology around, not to destroy competition.
The objective of OSS is to drive innovation and invention.
This is interesting, in fact. With a proper and efficient DB as a backend rather than a traditional FS, you could not only make fast searches, you could reorganize the way your data is presented to you.
Now you are entering data in directories and files, but what if you want to find out what you worked on a particular series of days? You could reorganize your data on the fly so that your directory structure would now be presented in a timeline fashion. E.g. you would be looking in for files in/2003/June/1[5-8]/*
A nice database combined with version control could potentially find for you the changes to a particular file you made that day. All pretty much instantly.
You could decide to organize all the data any which way that would make sense to you at the time. You would not be stuck with a fixed directory structure anymore.
If this is what MS has in mind this could be a real advance.
> All depends on how you define "best" I guess. I'm > only after a single language and platform, and I'm > judging based only on execution speed of the code > produced. gcc is not in the running.
The word you are looking for is `uniformly'. Gcc is not uniformly better than the sum of all other C/C++ compilers. Some compiler smaller or faster code, some are more ISO compliant, some come with a GUI.
However as a single piece of software Ccc is arguably the best compiler because it compiles very reasonable code on just about any platform known to Man, from your lowly Palm to IBM mainframes and various supercomputers via your current platform of choice. You would not use it for extreme performance but you can use it as your benchmark. If a compiler is not better than Gcc in some useful way it is not worth much.
Have you read the article? No Sunday, no Saturday, no overtime, no compensation, nothing. Just 6*84=504h of straight work with not even enough time to sleep.
This is not a job, this is a Death March. Have you done one of them yourself? After two weeks of 12/7 regimen they will not recognize their own family members. After that there will be actual sickness (depression, heart attack) and possibly suicides if the Death March is managed efficiently.
In Australia even the plans for businesses are laughable. This is actually a huge problem, how can internet services get even started without bandwidth?
The only entities with good access are universities and the CSIRO. They layed down AARNET in Australia 25 years ago, and now they are enjoying 100mb+ accesses. On AARNET the typical 600MB CD-ROM gets downloaded in 10-20 minutes...
I don't want to make it a religious issue, but my interpretation is that if you want to use Isp, it should be in (s), and you can use both metric and Imperial measurements. In Imperial you must use pounds of thrust and pounds of mass, in metric your formula is slightly different because you must use Newtows and kilograms, but there is a (g) correcting factor so that eventually everything evaluates to the *same* values on both sides of the Atlantic.
Do you agree with that?
Re:I'll care when native compilers become the norm
on
Preview of Java 1.5
·
· Score: 1
Sorry I should have made it clearer, that wasn't my point. Even if you run your path-optimized executable on the *test* data, which is basically the most advantageous figure case, you don't get that much of a speedup over a normally optimized executable, if the test case is sufficiently complex to represent reality (if it's a toy test case, then sure you do get some speedup, but who cares?).
I haven't seen the documentary that you speak of, but I have seen one called `Citizen Welles', which tells a similar story.
My reading of the `Citizen Kane' story is different from yours. Welles was a very bright young man when he started on that film. He had negociated and obtained unheard-of conditions to make his film, and used them to the fullest. He made his film, finished it pretty much to his satisfaction and then had a hell of a time to get it distributed and recognized thanks to Hearts' efforts.
However it did not finish him, by a long shot. According to IMDB He directed another 36 films after CK, produced another 12 and acted in another 110 films! This is just the sort of guy he was. A cocky, overconfident little brat who had no qualms about making a film on the super-rich media magnate of his time. This achievement brought him an immense deal of respect *outside* of Hollywood at first, and eventually within again.
In 1942, Orson Welles himself was nominated for best picture, best actor and best director for CK. He WON the best original screenplay Oscar. In 1943 he directed `The magnificent Anderson' and that picture won the Oscar for best film.
in 1971 Hollywood awarded him a `lifetime achievement' oscar.
Among his other film, there are `touch of evil', `the 3rd man', an adaptation of Kafka's `the trial', `the lady of Shanghai', classic renditions of Macbeth and Othello, which all took awards and are considered great films.
On the other hand sure Hearst managed to blockroad CK as much as he can. He spent all of his influence making sure the film was never going to be a success, and he FAILED. CK is regularly voted among the 10 best films of ALL TIMES. People everywhere have seen this film and critics never fail to mention that Kane's character is based on W.R. Hearst and how he tried to have this film banned. Now Hearst is seens as one of greatest bastards of all times, and this film effectively ruined his business and his life.
Sure CK cost a lot to Welles, but you don't get that sort of success (because CK was eventually a huge success) without pain. With 20/20 hindsight and the wisdom of an older man, Welles admits that it might have been an error to have made that film, or that he should have waited until he had an unassailable position in Hollywood to make it, but he had this film in him, and it made him, and it unmade him at the same time.
But what a story! How many people have one like that to tell? Would have Welles been happier if he had been more calculating and more patient? maybe, but who can tell for sure?
Yes sir, I wasn't very clear. With due respect Microsoft was not convicted (confirmed on appeal) of obtaining a monopoly, but of illegally leveraging the monopoly that they have to effectively kill other competing companies. The case in point was Netscape but it was in no way the only one.
We can all agree that the `punishment' was extremely weak, although being convicted of illegal business practices is never nice and did hurt Microsoft in the PR department.
Re:I'll care when native compilers become the norm
on
Preview of Java 1.5
·
· Score: 1
Ah, "the JIT knows better" argument.
In fact, in a C/C++ environment you can optimize a particular program knowing exactly which paths and variables it is going to use, this is called `profile-based optimization'. You can try this with the Intel C/C++ compiler or with a recent gcc.
Now you need a test sample of the data your application is going to work on, but then the optimizer will make use that data exactly as a JIT compiler would do.
If you try this with a realistic application (not a toy 10 liner) you will find that with few exceptions that super-duper optimization strategy does not do significantly better than normal optimization (it does get a few percent faster). What I mean is if you try this method on an already somewhat big, mature and tested application you will get few benefits.
This is because in a realistic application, things get optimized by the interaction between programmers and testers already. Often inefficient code will have been removed already, and what is left is there for a reason so there is not much to work on. Try it for yourself sometimes.
> If that statement ever shows up in court he is > royally fucked.
Not necessarily. A lot of patents really are ridiculous, you only have to show that. I'm sure that the community would help footing his legal bill. We might end up with a saner patent system as a result.
> No matter what you say, improving weapons is a > noble goal
Yeah right, until you are at the receiving end of such weapons.
So the new weapons make war a possibility again, right? and that is noble?
Which one is best: huge weapons that kill lots of people that are so bad that they can't be used (strategic nukes) or little weapons that kill the people you want only -- mostly, so they can be used (so called smart bombs)?
with the former no one gets killed, with the latter lots do get killed, including some innocents. I find your reasoning higly dubious.
The US weapons program is not being developed with humanistic goals in mind. It is developed so that the US can continue to wage war against less powerful opponents with minimal bad press. The question of whether such wars are justified or not is never addressed.
Now to fight against such weapons the ennemy has no choice but to fight dirty. Remember this next time a tower crumbles because maybe you will be a victim of the collateral damage.
Slashdot Mirror
The KAI contribution is only relevant to C++. For straight C it doesn't help. For C++, as you said, it was very good, particularly at optimizing template code and generally getting rid of abstraction layer penalties.
This is the reputation that Gcc has, but it's not quite true anymore.
Gcc improves all the time. They have been using the SPEC benchmarks to monitor performance at every build for quite some time now, and it shows. There are commercial players behing Gcc (and Apple is one of them), for whom performance is really an issue.
The numbers for recent versions of GCC aren't that bad, really. In an objective comparison you will find that Gcc sometime yields faster code than the native compiler of choice. This has been true for me on Alphas, PPCs, Sparcs and x86. Sometimes this is not true and the native compiler produces significantly better code, but this only points out where things can be improved in Gcc. There is nothing set in concrete there.
I don't think that there are too many cooks. Each of them can focus on a particular back end.
Also I've consistently found native compiler buggier. As recently as yesterday the Intel Compiler (latest version: 7.1) could not compile tth, on Linux RedHat 7.3. It does after allocating all the available memory in the system (tth is a very long single-file program). Try it yourself if you don't believe me.
Finally Gcc is the cornerstone of the Free Software movement. Without it no Linux, no *BSD, nothing. It's probably the most tested and solid compiler on the face of the Earth.
I can't believe this got moderated insightful. Without 64bit you are stuck with a 2GB (possibly 3GB) limit on memory for every process. Swapping has nothing to do with it.
Where I work people are running processes that require well above 3GB or RAM. Those don't run at all on 32-bit machines.
Next year or the year after that *you* will be crying for 4GB of RAM, and you will need it. 32-bit computing is as good as dead, it's only a matter of time (and a short amount at that). The only thing that will maintain it is Intel's insistence that you don't need it, like they did with the 80x86 line with 32-bit vs. 16-bit.
Well the killing statistic is the Google access rank. Only 1% of requests to Google come from Linux, see Zeitgeist.
From that one statistic it would appear that the number of Linux desktops is in fact overestimated, which is very depressing. Unless you have a good explanation of why people would install Linux on their desktop but then choose to browse from a Windows box.
From that same site the number of Apple-based request is 3 times that of Linux. Linux is not likely to pass over Apple as a whole anytime soon.
Hi, thanks for that.
Well that's the whole point that I hoped the lawsuit against the author of DeCSS would bring out. What have we bought when we buy a DVD?
Clearly the DVD content providers want you to buy the disk, *plus* an approved player with attached royalties, *and* a copy for each region you want to play the disk in.
Movie studios don't want you to do your own imports. They want to control the distribution of films around the world. They want to be free to delay the distribution of a film, say in Australia, for several years. At the same time they don't want to impede the sales of DVDs of the same film, say in the USA. They don't want people in Australia to be able to view the film in their home theatre before it's distributed there.
They do not want you to view your paid-for films from your hard disk. They don't want you to make backups of their content, they want you to pay again if your disk gets damaged somehow.
In the future they want to be able to introduce DVD media that become unplayable after a few days or a number of runs.
All the above is made impossible by DeCSS. Movie studios have only themselves to blame. They tried to be too greedy and it backfired on them.
However don't be too optimistic about movie studios learning a lesson from this. The next digital format to come along will have strong encryption.
Hi,
I'm not the only one having false memories, then. Bill Gates actually said that about Windows NT 3.x.
Do a Google on 'Microsoft Windows "better Unix than Unix"', I got 7 pages of people remembering the exact same quote, from an add; here is an exerpt.
> One of the things that upsets us about Microsoft
> is that they depend a lot on what seems to be
> boastful marketing and sometimes, in their race
> to catch up with Unix, lets quality slip. A
> colleague of mine wrote: "It was in 1993 I
> believe when I saw a full page ad for Windows NT
> 3.1 with the title "A Better Unix Than Unix." I
> sure wish I'd kept it."
>
Sorry for late reply,
Actually, yes I have survived a project like that but it wasn't pretty. I was 22 at the time and I had 1h commute each way on top of the 12h days. I was absolutely 100% a wreck after 3 weeks of that, and the code I produced was a complete mess and had to be re-implemented after. It was all x86 assembly and you know how easy this is easy to read after the fact (it was more than 10 years ago, and a standard when writing device drivers then).
We met the deadline and delivered the product, which was promptly shelved by the client, never to be used until 6 months later. It turned out the deadline was completely artificial and had been used in an internal warfare in the client company. One of the executive was touting this project as infeasible and another decided to `prove him wrong' using our services rather than their own internal resources.
I resigned shortly after hearing of this, both the client and my ex company where history about 18 months later. Actually the client company was quite big and politically important at the time, and the collapse resulted in a spate of lawsuits. The CEO of that company was in real trouble (at some point he was looking at a prison term), and my own boss lost his company and was actually forbidden to run one for quite a long time.
This was my very first real job as a software engineer. I went back to graduate school after that.
My experience is telling me loud and clear that such projects are the hallmark of terrible, in fact destructive management and should be avoided at all costs.
> Home Use Desktop: Microsoft OSes are fantastic as
> home desktop OSes, they beat Linux in 95% of home
> use situations. The only exceptions in favor of
> Linux is non-game, surf and email purposes and
> people who code as a hobby.
You do pay for that home use software, don't you? Myself I don't have deep enough pockets and most of the free equivalent stuff does not run on Windows at all.
The only things I used to use Windows for was for (properly paid for) games but with NWN I'm more than satisfied now.
There is a major difference. In IE's case it only works on one platform (initially 2) and cannot be made to evolve by a third party. With all open-source code even if the initial developer is no longer interested in the project, others can take it on and make sure it's not dead technology.
In IE vs. Netscape the objective was to cut Netscape's "air supply" (the Microsoft memos and e-mails said so). In the OSS case the objective is to share the technology around, not to destroy competition.
The objective of OSS is to drive innovation and invention.
This is interesting, in fact. With a proper and efficient DB as a backend rather than a traditional FS, you could not only make fast searches, you could reorganize the way your data is presented to you.
/2003/June/1[5-8]/*
Now you are entering data in directories and files, but what if you want to find out what you worked on a particular series of days? You could reorganize your data on the fly so that your directory structure would now be presented in a timeline fashion. E.g. you would be looking in for files in
A nice database combined with version control could potentially find for you the changes to a particular file you made that day. All pretty much instantly.
You could decide to organize all the data any which way that would make sense to you at the time. You would not be stuck with a fixed directory structure anymore.
If this is what MS has in mind this could be a real advance.
Hi,
> All depends on how you define "best" I guess. I'm
> only after a single language and platform, and I'm
> judging based only on execution speed of the code
> produced. gcc is not in the running.
The word you are looking for is `uniformly'. Gcc is not uniformly better than the sum of all other C/C++ compilers. Some compiler smaller or faster code, some are more ISO compliant, some come with a GUI.
However as a single piece of software Ccc is arguably the best compiler because it compiles very reasonable code on just about any platform known to Man, from your lowly Palm to IBM mainframes and various supercomputers via your current platform of choice. You would not use it for extreme performance but you can use it as your benchmark. If a compiler is not better than Gcc in some useful way it is not worth much.
Cheers.
Have you read the article? No Sunday, no Saturday, no overtime, no compensation, nothing. Just 6*84=504h of straight work with not even enough time to sleep.
Jeez you are an angry person.
This is not a job, this is a Death March. Have you done one of them yourself? After two weeks of 12/7 regimen they will not recognize their own family members. After that there will be actual sickness (depression, heart attack) and possibly suicides if the Death March is managed efficiently.
This is scary. Everybody is rooting for the mighty anti-competitive giant of the 80's. What is going on?
Will we be all rooting for Microsoft in 20 years' time?
In Australia even the plans for businesses are laughable. This is actually a huge problem, how can internet services get even started without bandwidth?
The only entities with good access are universities and the CSIRO. They layed down AARNET in Australia 25 years ago, and now they are enjoying 100mb+ accesses. On AARNET the typical 600MB CD-ROM gets
downloaded in 10-20 minutes...
Xv may have been leapfrogged but not by eog. Come to think of it there is not a single viewer in the whole of RH9 distribution that I like.
Simple things such as pixel colour information are universally missing.
Right, they only (used to) say that `Windows will be a better Unix than Unix).
I don't want to make it a religious issue, but my interpretation is that if you want to use Isp, it should be in (s), and you can use both metric and Imperial measurements. In Imperial you must use pounds of thrust and pounds of mass, in metric your formula is slightly different because you must use Newtows and kilograms, but there is a (g) correcting factor so that eventually everything evaluates to the *same* values on both sides of the Atlantic.
Do you agree with that?
Sorry I should have made it clearer, that wasn't my point. Even if you run your path-optimized executable on the *test* data, which is basically the most advantageous figure case, you don't get that much of a speedup over a normally optimized executable, if the test case is sufficiently complex to represent reality (if it's a toy test case, then sure you do get some speedup, but who cares?).
I haven't seen the documentary that you speak of, but I have seen one called `Citizen Welles', which tells a similar story.
My reading of the `Citizen Kane' story is different from yours. Welles was a very bright young man when he started on that film. He had negociated and obtained unheard-of conditions to make his film, and used them to the fullest. He made his film, finished it pretty much to his satisfaction and then had a hell of a time to get it distributed and recognized thanks to Hearts' efforts.
However it did not finish him, by a long shot. According to IMDB He directed another 36 films after CK, produced another 12 and acted in another 110 films! This is just the sort of guy he was. A cocky, overconfident little brat who had no qualms about making a film on the super-rich media magnate of his time. This achievement brought him an immense deal of respect *outside* of Hollywood at first, and eventually within again.
In 1942, Orson Welles himself was nominated for best picture, best actor and best director for CK. He WON the best original screenplay Oscar. In 1943 he directed `The magnificent Anderson' and that picture won the Oscar for best film.
in 1971 Hollywood awarded him a `lifetime achievement' oscar.
Among his other film, there are `touch of evil', `the 3rd man', an adaptation of Kafka's `the trial', `the lady of Shanghai', classic renditions of Macbeth and Othello, which all took awards and are considered great films.
On the other hand sure Hearst managed to blockroad CK as much as he can. He spent all of his influence making sure the film was never going to be a success, and he FAILED. CK is regularly voted among the 10 best films of ALL TIMES. People everywhere have seen this film and critics never fail to mention that Kane's character is based on W.R. Hearst and how he tried to have this film banned. Now Hearst is seens as one of greatest bastards of all times, and this film effectively ruined his business and his life.
Sure CK cost a lot to Welles, but you don't get that sort of success (because CK was eventually a huge success) without pain. With 20/20 hindsight and the wisdom of an older man, Welles admits that it might have been an error to have made that film, or that he should have waited until he had an unassailable position in Hollywood to make it, but he had this film in him, and it made him, and it unmade him at the same time.
But what a story! How many people have one like that to tell? Would have Welles been happier if he had been more calculating and more patient? maybe, but who can tell for sure?
Yes sir, I wasn't very clear. With due respect Microsoft was not convicted (confirmed on appeal) of obtaining a monopoly, but of illegally leveraging the monopoly that they have to effectively kill other competing companies. The case in point was Netscape but it was in no way the only one.
We can all agree that the `punishment' was extremely weak, although being convicted of illegal business practices is never nice and did hurt Microsoft in the PR department.
Ah, "the JIT knows better" argument.
In fact, in a C/C++ environment you can optimize a particular program knowing exactly which paths and variables it is going to use, this is called `profile-based optimization'. You can try this with the Intel C/C++ compiler or with a recent gcc.
Now you need a test sample of the data your application is going to work on, but then the optimizer will make use that data exactly as a JIT compiler would do.
If you try this with a realistic application (not a toy 10 liner) you will find that with few exceptions that super-duper optimization strategy does not do significantly better than normal optimization (it does get a few percent faster). What I mean is if you try this method on an already somewhat big, mature and tested application you will get few benefits.
This is because in a realistic application, things get optimized by the interaction between programmers and testers already. Often inefficient code will have been removed already, and what is left is there for a reason so there is not much to work on. Try it for yourself sometimes.
Thanks,
I wish I could moderate you higher than you are now.
Cheers
> If that statement ever shows up in court he is
> royally fucked.
Not necessarily. A lot of patents really are ridiculous, you only have to show that. I'm sure that the community would help footing his legal bill. We might end up with a saner patent system as a result.
Bring it on.
> No matter what you say, improving weapons is a
> noble goal
Yeah right, until you are at the receiving end of such weapons.
So the new weapons make war a possibility again, right? and that is noble?
Which one is best: huge weapons that kill lots of people that are so bad that they can't be used (strategic nukes) or little weapons that kill the people you want only -- mostly, so they can be used (so called smart bombs)?
with the former no one gets killed, with the latter lots do get killed, including some innocents. I find your reasoning higly dubious.
The US weapons program is not being developed with humanistic goals in mind. It is developed so that the US can continue to wage war against less powerful opponents with minimal bad press. The question of whether such wars are justified or not is never addressed.
Now to fight against such weapons the ennemy has no choice but to fight dirty. Remember this next time a tower crumbles because maybe you will be a victim of the collateral damage.