The SMS is stored by the operator (in case receiver is not online). At least this is what happens in GSM, I have no clue about Verizon.
The storage is calles Short Message Service Centre, and they have been known to fail and/or create delays under heavy load (e.g. during new years eve).
The behaviour is very different from "normal" call, comparing bucks per byte makes no sense.
I think the parents point is that there cannot be universal rules for "passing their requirements".
Unless, of course, "they" are not accountable (i.e. cannot be sued). Which, of course, just paves way for bribery, mistreatment, unjustice, nepotism, etc.
First, Nyquist *IS* about faithful reproduction. But it is a mathematical concept (and therefore "assumes" e.g. infinitely accurate measurements). Nyquist states you can get infinidesimally close to f/2 (but not to f/2).
In reality there are such things as near-perfect digital low-pass filters. They are in pretty much every (decent) CD player.
Filtering with digital filters up to 22kHz is not impossible. Filtering up to 20kHz is trivial (with less than 0.1dB attenuation and no aliasing).
Actually the filters are quite cheap. They are called up-sampling (or oversampling) converters. They do most (or all) of the filtering in digital domain (before D/A).
They can get very close to 22kHz (if not better, I have not checked the specs lately) and the filter will be brick wall and phase error free.
Then you would start whining about something else. Like "OOo is soo slow and it is the biggest showstopper" or "there is no identical-to-Outlook replacement and it is the..." or whatever. Just look at the comments here, there are a lot of whiners.
You clearly have two possibilities: stay with Windows or don't. Whining is not going to change your options. If the games are so important you rather use Windows that is your problem, not Linux's.
To me open file formats and interoperable (free) applications are much more important.
Trac is somewhat limited. Having used Quality Center I appreciate integration of requirement specifications and test cases (to bug database).
After all, a defect should be findable by a test case (else you should create new one), and calculating defect count for a feature/requirement is a good thing to have.
Integration to SVN is a very good feature (which QC lacks).
I do agree that the kernel needs to improve. IMHO a nice compromise is to keep KBI stable during major versions[1]. It does not mean there cannot be any changes. It means the old KBI is kept (not removed, not changed) and new things are just added to it.
I think keeping "old" interfaces "for a while" is nowadays not such a big hindrance - they most likely can be kept as a loadable modules. The USB is still a complete mess (both in Linux and Solaris - AFAIK).
[1] my rant of losing something from 2.4 to 2.6 is therefore just a rant without much substance. I just wanted to point out that "having code in kernel" is not a guarantee of anything.
To make my point: I do not give a shit about nVidia.
But I do care when some *beep* gets "binary-only" under his/her skin so badly that he/she makes *on purpose* my life much harder than necessary. And this is what Linux kernel developers are doing!
As can be seen, nVidia is doing pretty much what you propose: they are testing their SW for every release.
But OSS people are not. Give me mtp006 driver for 2.6! (it was included in 2.4 kernel tree). Alas, you cannot.
See, OSS developers do not have time - they graduate - they lose interest. Drivers are left without testers and developers. I, the end user, became tester. My life gets harder. Just because some *beep* thinks it is okay to annoy nVidia.
As somebody (Rich0) put it "The whole point of the no-binary-interface to the kernel is to make your life as big of a pain as possible." I do not think this is acceptable strategy.
Besides, I think it is naive to think nVidia can *in reality* (i.e. it makes business sense) get all the code they use with OSS license. But as said - I do not care about nVidia. I care about *me* (and OSS driver developers).
(I do not need mtp006 anymore because the motherboard died, but I would have needed it back then)
Again you ignore testing completely. Are you implying testing is not necessary?
You do have it right that having code in kernel tree does not stop nVidia (from improving). It stops OSS developers (there are devices like mtp006 which were not ported from 2.4 to 2.6). Licenses and contracts stop nVidia.
At the same time the job of OSS driver writers is made "as big of a pain as possible". That discourages me to write drivers for Linux (I have no intent writing binary-only drivers).
If the code isn't open it doesn't really benefit him. Why did he use Bitkeeper? Double standards, perhaps?
1. I do not want to recompile NVidia drivers after every (kernel) upgrade. There wont be a situation when there is no binary-only drivers, not in my life time. 2. If my devices X and Y are not in kernel tree (or even if they are), I cannot know which versions of the kernel they are tested with. Or if I know the versions might not be same. What to do?
Besides the link has outright lies - unless you ignore testing completely. Getting a driver into kernel tree does not give any quality improvements - quite contrary as the kernel developers do not have the HW to *test* with. Changes without testing? No thanks!
Compare this to Solaris. I downloaded NVidia driver (and Marvel gigabit driver). I install them. They work. I upgrade from S10 to S10 06/06. Then from S10 06/06 to S10 06/10. They work. Most likely I can upgrade to S11 and they still work (with perhaps some limited functionality/performance). The developers can use their time on improvements, not recompiling and testing every two weeks for every imaginable distribution. The testing is non-trivial, as there is huge amount of different (SMP etc) machines.
My experience, Asus A8V motherboard. Solaris did not (do not?) support the (Marvel) gigabit controller out of box but it was easy to find (alas binary only) drivers for it. Similarly it was easy to download and install NVidia drivers.
The only problems I encountered: 1. No support for my scanner (Epson - Avasys makes binary-only drivers for Linux). 2. No support (back then) for virtual consoles. I *need* them. 3. Mouse stutt-tters. Horribly. A show stopper. Either GeForce FX5200 is not properly supported by the driver or there is something broken. I have not heard anybody else having this problem...
I have heard there is support for the VCs now, so I might try again. Besides, I probably will upgrade the display adapter to 6600 or like.
BTW, ZFS was *excellent*, nothing I have used before comes close.
Fact: nobody disagrees that streams modules are easier to write than Linux kernel modules.
Still even the idea of accepting streams into Linux kernel is abhorred.
The reason? It is not "fast enough" for IP. Though Solaris IP, using (heavily modified) stream modules is an excellent performer.
Unfortunately Linux is more guilty of NIH than many others.
And who you would turn to, if not some kind of government?
The less fortunately is exactly what *I* am talking about - and I still, after forty-something years, cannot be sure I am fortunate. I cannot definitely say any of my "choices" are bad, but I am not smart enough to say whether the stocks (etc.) I have are going to strong enough after another 40 years and be enough to pay my potential hospital bills *at that age*.
Communism... it so silly straw man that I am not touching it... just as silly as saying "liberatism is Somalia".
And it is true, you cannot provide less fortunate if you don't adequately manage everybody else. It is just what is "adequate", and what is "everybody else".
Somebody is always going to pay the bills - either in taxes or in a "war" - and as you know, the innocents are going lose most in a war.
And you honestly believe that Mono implementing Silverlight will actually make a difference as to whether Silverlight succeeds or fails? No, you do. Seriously. If you didn't you would not announce it.
Besides, I am absolutely certain you cannot implement the DRM which is required for "getting access to content".
The "hope" (b) approach worked perfectly for ActiveX, it worked, despite your "heroic" efforts, with.NET 1.1 and I see no reason why it will not work for.NET 2.0 (and Silverlight). Java and Flash are just better, especially in OSS sense.
You have fallen too heavily in love with your pet project even to accept it has huge problems: cross platform problems with GUI, patents, DRM and perhaps even GPL3.
Quite often some of the intermediate results are interesting, either for debugging, backup or other reasons.
And if you do not need the intermediate at all then why limit yourself to a shell, why not pick a real programming language? What is the point of the shell?
Apparently any young having e.g. leukemia is "a lardass".
Your belief that you can be completely responsible of your destiny is admirable. I do not share that belief, I am certain if a lot of people are left outside the society the will, sooner or later, revolt.
BTW, I am educated, I am a "net payer" (pay more taxes than get in "return"), but still would never think about cutting the "safety net" from less fortunate.
What is wrong with Cygwin? Compared to what? cmd.exe of bash in *nix?
To latter: Installation: default installation doesn't install even cwd.exe. Home directory is not \Documents\ and\ setting\me but/home/me (fixable with symlink) X is still not very usable. Terminal is "cygwin", vt52/vt100 emulation is very poor (ssh+screen to other hosts breaks). Handling of network disks is awfull ("ln -s/cygdrive/c/C", etc. helps a *LOT*). Escapes like "C:\\Program\ Files" are PITA, as is keeping $PATH up to date. unzip does not handle permission properly (exe/dll => should be executable, but no => winzip packed exes gets broken). File locking ("cannot access - is in use") is humongously brain dead (not fault of Cygwin, per se). Symbolic links (and hardlinks) do not work properly (again, more Windows fault). From Perl (and Python,...) you need two version: native and cygwin. Which are not 100% compatible.
Slashdot Mirror
The SMS is stored by the operator (in case receiver is not online). At least this is what happens in GSM, I have no clue about Verizon.
The storage is calles Short Message Service Centre, and they have been known to fail and/or create delays under heavy load (e.g. during new years eve).
The behaviour is very different from "normal" call, comparing bucks per byte makes no sense.
I think the parents point is that there cannot be universal rules for "passing their requirements".
Unless, of course, "they" are not accountable (i.e. cannot be sued). Which, of course, just paves way for bribery, mistreatment, unjustice, nepotism, etc.
I insist on using imperial units! Then a terabyte would be 2.34 thousand milebytes.
First, Nyquist *IS* about faithful reproduction. But it is a mathematical concept (and therefore "assumes" e.g. infinitely accurate measurements). Nyquist states you can get infinidesimally close to f/2 (but not to f/2).
In reality there are such things as near-perfect digital low-pass filters. They are in pretty much every (decent) CD player.
Filtering with digital filters up to 22kHz is not impossible. Filtering up to 20kHz is trivial (with less than 0.1dB attenuation and no aliasing).
Actually the filters are quite cheap. They are called up-sampling (or oversampling) converters. They do most (or all) of the filtering in digital domain (before D/A).
They can get very close to 22kHz (if not better, I have not checked the specs lately) and the filter will be brick wall and phase error free.
No, it would not.
..." or whatever. Just look at the comments here, there are a lot of whiners.
Then you would start whining about something else. Like "OOo is soo slow and it is the biggest showstopper" or "there is no identical-to-Outlook replacement and it is the
You clearly have two possibilities: stay with Windows or don't. Whining is not going to change your options. If the games are so important you rather use Windows that is your problem, not Linux's.
To me open file formats and interoperable (free) applications are much more important.
Trac is somewhat limited. Having used Quality Center I appreciate integration of requirement specifications and test cases (to bug database).
After all, a defect should be findable by a test case (else you should create new one), and calculating defect count for a feature/requirement is a good thing to have.
Integration to SVN is a very good feature (which QC lacks).
The whole war is illegal, by the international law. So, in principle, you can think every single soldier breaking was breaking "the law".
So, "breaking the law" is not the question, it is "breaking the moral".
I leave to the reader to think whether the soldiers do break the moral. I myself am 100% certain none of the soldiers (on any "side") have any moral.
I do agree that the kernel needs to improve. IMHO a nice compromise is to keep KBI stable during major versions[1]. It does not mean there cannot be any changes. It means the old KBI is kept (not removed, not changed) and new things are just added to it.
I think keeping "old" interfaces "for a while" is nowadays not such a big hindrance - they most likely can be kept as a loadable modules. The USB is still a complete mess (both in Linux and Solaris - AFAIK).
[1] my rant of losing something from 2.4 to 2.6 is therefore just a rant without much substance. I just wanted to point out that "having code in kernel" is not a guarantee of anything.
To make my point: I do not give a shit about nVidia.
But I do care when some *beep* gets "binary-only" under his/her skin so badly that he/she makes *on purpose* my life much harder than necessary. And this is what Linux kernel developers are doing!
As can be seen, nVidia is doing pretty much what you propose: they are testing their SW for every release. But OSS people are not. Give me mtp006 driver for 2.6! (it was included in 2.4 kernel tree). Alas, you cannot. See, OSS developers do not have time - they graduate - they lose interest. Drivers are left without testers and developers. I, the end user, became tester. My life gets harder. Just because some *beep* thinks it is okay to annoy nVidia.
As somebody (Rich0) put it "The whole point of the no-binary-interface to the kernel is to make your life as big of a pain as possible." I do not think this is acceptable strategy.
Besides, I think it is naive to think nVidia can *in reality* (i.e. it makes business sense) get all the code they use with OSS license. But as said - I do not care about nVidia. I care about *me* (and OSS driver developers).
(I do not need mtp006 anymore because the motherboard died, but I would have needed it back then)
You do have it right that having code in kernel tree does not stop nVidia (from improving). It stops OSS developers (there are devices like mtp006 which were not ported from 2.4 to 2.6). Licenses and contracts stop nVidia.
At the same time the job of OSS driver writers is made "as big of a pain as possible". That discourages me to write drivers for Linux (I have no intent writing binary-only drivers). If the code isn't open it doesn't really benefit him. Why did he use Bitkeeper? Double standards, perhaps?
I would not call the model "very successful", just because the shortcomings I pointed out.
Besides, you ignored the fubared users. Just for the convenience of the kernel developers all the users are inconvenienced.
Everybody else is perfectly capable of creating stable kernel binary interface (within minor revisions). Why not Linux?
Oh, yes *I* do.
c id=18955733 - by someone else. There are a lot more similar experiences.
1. I do not want to recompile NVidia drivers after every (kernel) upgrade. There wont be a situation when there is no binary-only drivers, not in my life time.
2. If my devices X and Y are not in kernel tree (or even if they are), I cannot know which versions of the kernel they are tested with. Or if I know the versions might not be same. What to do?
Besides the link has outright lies - unless you ignore testing completely. Getting a driver into kernel tree does not give any quality improvements - quite contrary as the kernel developers do not have the HW to *test* with. Changes without testing? No thanks!
Compare this to Solaris. I downloaded NVidia driver (and Marvel gigabit driver). I install them. They work. I upgrade from S10 to S10 06/06. Then from S10 06/06 to S10 06/10. They work. Most likely I can upgrade to S11 and they still work (with perhaps some limited functionality/performance). The developers can use their time on improvements, not recompiling and testing every two weeks for every imaginable distribution. The testing is non-trivial, as there is huge amount of different (SMP etc) machines.
Also see http://linux.slashdot.org/comments.pl?sid=233087&
My experience, Asus A8V motherboard. Solaris did not (do not?) support the (Marvel) gigabit controller out of box but it was easy to find (alas binary only) drivers for it. Similarly it was easy to download and install NVidia drivers.
...
The only problems I encountered:
1. No support for my scanner (Epson - Avasys makes binary-only drivers for Linux).
2. No support (back then) for virtual consoles. I *need* them.
3. Mouse stutt-tters. Horribly. A show stopper. Either GeForce FX5200 is not properly supported by the driver or there is something broken. I have not heard anybody else having this problem
I have heard there is support for the VCs now, so I might try again. Besides, I probably will upgrade the display adapter to 6600 or like.
BTW, ZFS was *excellent*, nothing I have used before comes close.
The reason? It is not "fast enough" for IP. Though Solaris IP, using (heavily modified) stream modules is an excellent performer.
Unfortunately Linux is more guilty of NIH than many others.
And who you would turn to, if not some kind of government? The less fortunately is exactly what *I* am talking about - and I still, after forty-something years, cannot be sure I am fortunate. I cannot definitely say any of my "choices" are bad, but I am not smart enough to say whether the stocks (etc.) I have are going to strong enough after another 40 years and be enough to pay my potential hospital bills *at that age*. Communism ... it so silly straw man that I am not touching it... just as silly as saying "liberatism is Somalia".
And it is true, you cannot provide less fortunate if you don't adequately manage everybody else. It is just what is "adequate", and what is "everybody else".
Somebody is always going to pay the bills - either in taxes or in a "war" - and as you know, the innocents are going lose most in a war.
Seriously.
If you didn't you would not announce it.
Besides, I am absolutely certain you cannot implement the DRM which is required for "getting access to content".
.NET 1.1 and I see no reason why it will not work for .NET 2.0 (and Silverlight). Java and Flash are just better, especially in OSS sense.
The "hope" (b) approach worked perfectly for ActiveX, it worked, despite your "heroic" efforts, with
You have fallen too heavily in love with your pet project even to accept it has huge problems: cross platform problems with GUI, patents, DRM and perhaps even GPL3.
Quite often some of the intermediate results are interesting, either for debugging, backup or other reasons.
And if you do not need the intermediate at all then why limit yourself to a shell, why not pick a real programming language? What is the point of the shell?
Apparently any young having e.g. leukemia is "a lardass". Your belief that you can be completely responsible of your destiny is admirable. I do not share that belief, I am certain if a lot of people are left outside the society the will, sooner or later, revolt. BTW, I am educated, I am a "net payer" (pay more taxes than get in "return"), but still would never think about cutting the "safety net" from less fortunate.
No, absolutely not - you missed my point completely.
Everything I said is based on simple statistics and human nature: people who think they're better off not paying taxes do not want to pay taxes.
Lemme make an out-of-ass wild trollish guess: you are young healthy white single male with no kids.
You should be a libertarian.
Lemme make another out-of-ass wild totally troll guess: you have absolutely no problem using public roads, etc. And that you oppose toll roads.
To latter:
Installation: default installation doesn't install even cwd.exe.
Home directory is not \Documents\ and\ setting\me but
X is still not very usable.
Terminal is "cygwin", vt52/vt100 emulation is very poor (ssh+screen to other hosts breaks).
Handling of network disks is awfull ("ln -s
Escapes like "C:\\Program\ Files" are PITA, as is keeping $PATH up to date.
unzip does not handle permission properly (exe/dll => should be executable, but no => winzip packed exes gets broken).
File locking ("cannot access - is in use") is humongously brain dead (not fault of Cygwin, per se).
Symbolic links (and hardlinks) do not work properly (again, more Windows fault).
From Perl (and Python,
Need more?