Not to be a troll, but this guy seems to be claiming he knows what he's doing. From his writeup, I very much doubt that.
Yeah, I thought the same thing. He goes out to test XBox as a cluster and then mainly bitches about how hard it was to setup (minor things like having to install gcc himself; btw, it's Mandrake, that means urpmi gcc should have worked?!). And he complains about modding being bothersome. After that I expected him at least to mention the outsight that installing will be easy in the future, if the XBox Linux project succeeds.
For the actual performance results he reserved less than a paragraph. Sad. Also, I'd like to see the MPI program he used for testing. Results for a "normal" PC cluster and for a single machine cluster would have been interesting for comparison, too.
PS: Regarding the kernel compile time - maybe DMA for the disk was not enabled?
I do not look forward to the day where I have to talk somebody through a command-line problem.
I don't see your point. How is this different from tweaking the registry or having to run ping in the console (what I see Windows users doing here regularly on trouble-shooting)?
I just told you, you have a GUI for most issues with recent distros. Sure, some problems still have no GUI solution, especially if the software is brand new (but end-users shouldn't use alpha software anyhow), but that is the same for Windows, IMNSHO.
[...] The monopoly agrument is very stale.
I expected that rebuttal. That was why I wrote there are several reasons.
I did mention it, because it is a real concern for us. If we could do as we want, we would be a complete Linux/FreeBSD/Mac shop. But half of our desktop computers/notebooks have Microsoft Windows installed solely or as dual boot because we have to.
Every time I suggest Mac or Linux when I install a computer for an aquaintence, I get to hear, they want Windows because that is what everyone uses and that are the computer skills employers ask for. So it may an old argument, but at least for my company and me personally, it is not stale, but going bothering every day.
The UI of recent Linux distros are not that bad anymore. They just differ from what most people know, i.e. Microsoft Windows.
Both KDE and Gnome, though they are significantly better now, are still inferior to Windows when it comes to UI. (as of Redhat 7.3, I have not used the ones with 8.0, I've heard they're supposed to be better.)
Well, that's what recent was about.
It suffers from serious lack of refinement. It's hard to find anything (In KDE, nearly every app begins with K),
I can only speak for Mandrake and Debian, because that's what I regularly use. In Mandrake, there is a menu which let's you choose the application by the task you want to accomplish, not by name. It is not perfect yet, but it's a good start.
and I've had no luck doing any real configuration. You still have to use the CLI to change desktop resolution, for example.
Last time I needed to use the shell for that was several years ago. If Redhat really has no GUI for this, it sucks more than I thought. But it sounds more probable that you are out of date.
Installing something is a cruel joke. The RPM Packager thing-a-ma-bob is not a very clear UI, nor is it immediately obvious why you need to go through that. Coming from the Windows world where you use setup.exe and pick where you want your stuff to go, RPM is full of un-necessities.
I can only tell for Mandrake, but there installing a packet does not involve rpms in any visible way. You simply choose in the GUI the packet to install, it will ask you to insert the CD or it will download it and that's it. No unnecessities involved.
Btw, Windows installer regularly asks question in such a cryptic way, that the uninitiated don't know what the question is about at all (ever hade a code version conflict - it happens regularly with non-English versions).
So no, it's not a matter of 'just being different', [...]
Oh yes, it is. You just proved it yourself. You are obviously coming from a Windows background and comparing to that.
My point was about people who are not used to any OS. I am the "computer guy" here and from the cries for help I get and the misconfiguration people make, I can tell you that Microsoft Windows is a lot more confusing than you admit - for quite some people.
Microsoft has put a great deal of research effort into making a usable UI.
Yes. I never said anything contrary. Microsoft has done a lot to further usability. I just claimed that Linux distros have catched up in recent versions. And that the stuff people still complain about is mostly that it simply does not behave as they are used to. Different. Not more difficult. See?
Like or hate it, millions upon millions of people are able to make use of it.
The amount of users does not imply anything regarding usability, since usability is only one of several reasons to buy software. Microsoft being a monopoly another one amongst others.
I remember once mysql said they would never implement stored procs or triggers or referential integrity,
I am interested in MySQL since when 3.21.x was current (1997?) and cannot remember such a statement. Surely, once they said, that they have no plans to do so. But that is not the same: don't having it on the list of things to be done is quite different from claiming to never implement it.
Also they often said, that when any of that feature is to come, it may not make existing stuff slower or it will be at least optional (and probably disabled by default).
AFAICT, the facts they created in the meantime do not contradict themselves yet.
it looks like they changed their minds.
OTOH, there would be nothing wrong with changing your mind once every 5 years or so.:-)
but for someone who it isn't they want a clean and clear interface that affords usage in obvious ways (for a media player that of course is for VCR like functionality)
Now you all understand where Windows users come from. A little bit of instability is a small price to pay for an interface that does what you expect.
I grant you this for Microsoft Windows users. They don't know better. But the original poster spoke about clean and clear interfaces, and I can tell you that not one of my technically non-savvy acquaintances considered the GUI of MS Windows clean and clear when first confronted with it.
Most of them found it very confusing and some of them still do not dare "playing" around, because always "strange, unexpected things" happen.
The UI of recent Linux distros are not that bad anymore. They just differ from what most people know, i.e. Microsoft Windows.
Nothing? OK, not quite nothing, but even Mozilla (one of the finest examples of recently complete software) is still being released 'for testing purposes only.'
Ouch. One of the worst examples you could have chosen. Mozilla is not intended to be ever released as end user software. That they leave to others. Mozilla is a browser and also a toolkit to base other software on (like Netscape).
So yes, their releases will always be for testing purposes only, because that is their stated goal. If they were a closed-source entity, they would never make public releases, the same as you don't see public releases of the general base your BIOS is specialized from.
Stop packaging Linux as a crappy Windows-clone (see Mandrake), and people will stop making the comparison.
Huh? To stay a bit more on-topic, i.e. related to the article: Because they are already packaged, on Mandrake it is a matter of some mouse clicks to install any of the named players. Very crappy, yes. (the crappy GUIs of the players are a different story)
So the installation hurdle JWZ complained about does not exist in recent distros (I simply presume that RedHat, Debian and SuSE do as well, meahwhile). In other words: using an older distro to judge the current state of video on Linux is more than a bit strange.
Oracle, as most commercial DBMSs, doesn't let you export the database in SQL format. Of course, you can write scripts to do that, but it shows how the commercial companies are always trying to find ways to lock you in.
[...] There is no "SQL format" for "exporting databases."
Huh? Why everyone in this thread seems to concentrate on the wording? Anyone using databases should know that he simply meant a SQL dump. For the clueless: creating SQL commands, which will create a copy of the database in question, when run through an appropriate tool (for running SQL commands in batch) and writing them to a file.
You'll need to write scripts in any case.
Huh? Why? Any open source database comes with the tools to run such a SQL dump. And most commercial ones, too, I presume.
I've spent so much time lately in the (relatively) flat-table world of MySQL that I had forgotten about inherited tables, subselects, constraints in table definitions,
Umm. Constraints are there with InnoDB for what? 2 years?
and oh yes, vacuuming.;) Looks like it is time to revisit postgres, especially for some db-agnostic PEAR apps I'm building. For me, it's the subselects that really make it worth the effort.
Well, or you might have a look at the current alpha (version 4.1) were basic sub-selects and derived tables are already running. Though, without much optimization yet.
That is not meant as shameless plug (although it probably is;-), but only because you sounded interested in MySQL I thought I update you on the fact that those features you are missing are under full work currently.
> If you are robbed, you lose some property. And you get tied up and beat. And your wife is murdered. And your daughter is raped. But its only property, right?
Well, as was already said in another comment, a gun mostly surprises and scares the unprepared. If "they" get into your house with the intend to rape and murder, I think there is good reason to believe they do not come unprepared.
I agree, in a worst case scenario you can get things only better by defending your family in any way possible, including a gun.
But most cases are not worst cases and there is a good chance that using a gun lets things escalate unnecessarily. That is what the parent post was about, IMHO.
Isn't it funny, that Texas, a state where anyone can carry a gun as long as its *plainly visable* has the lowest amount of fatal gun violence in the U.S.A. But Maryland, where carrying a gun is illegal for private citizens, is one of the highest ?
I guess you are mixing up cause and effect. I don't know any details, but I would bet, that carrying a gun in Maryland has been outlawed *because* it has one of the highest amount of fatal gun violence. Not the other way around.
Take for example Germany, where I live: You need a license in order to own most weapons (not limited to guns). It is non-trivial to get a license and becomes the harder the more dangerous the weapon is - some weapons are banned completely. E.g. there is no way at all to get a license for a machine gun, AFAIK.
Most guns are owned and used in shooting clubs for sportive reasons (and only licensed for that use) and are specially crafted for sportive purposes.
Even if you own a (normal) license, it is illegal to carry a *loaded* gun. You have to transport gun and munition seperated. There are only few people who possess a license to carry loaded weapons. Mostly police and such. It is very unusual that security personal carries guns, except maybe for bodyguards.
I don't know anybody owning a gun and I don't remember when the last time was I saw someone with a gun (police officers excluded) and I don't think I ever saw somebody with a drawn gun (police officers included). Except in TV, of course.;)
If someone would hold a gun, I would be quite scared, too, because it is so uncommon here.
I don't mean to claim that allowing guns necessarily means that there will be more violence. But saying the opposite is nonsense. There are a lot of countries (England, France,...) which show that you can live relatively safe without everyone being allowed to own a gun.
E.g. in one year, in Germany about 400 people are killed by guns (not limited to crimes, i.e. including people killed by police). And about 1200 people die by homicide (not only guns). USA, which has about 3 times the population has about 11.000 people killed by guns only.
But when I tried it (first quarter this year), I got too frustrated with setting up. If there is an easy way to avoid the part about enforcement, I missed it.
Maybe the problem is only that the documentation is too centric about that and ignores that some people would try Aegis for other reasons. Maybe I just was too unpatient.
But in the end, I dropped arch, because I got the impression that "skipping and ignoring" the steps got too burdensome. Probably this is, because Aegis emphasis too much (for my needs) to be more than a revision control system.
What turned me away from Aegis is the fact that you have to adapt a policy at all. Aegis sounds great if you want to enforce a process.
But if you are fine to do without a rule-based process (maybe you have a small team, or very disciplined developers, or whatever), Aegis looks as if it brings a lot of unnecessary load with it.
Aegis is a source management system (and enforces it), while I only want a revision control system.
This alone is why Linux is never going to be a mainstream OS... WHO WANTS TO DO ALL THAT SHOT JUST TO WATCH A QUICKTIME!!!!???
First understand that this is not a Linux problem per se, but an Apple one. I would be curious what you would do on Microsoft Windows, if the QT plugin would be only published for Linux. Don't laugh: consider what could be in 5, 10 or 20 years. So my real point is, within some years at most, Linux versions for apps will be published by default.
Second, understand that you have to go through these install steps, because QT support in mplayer is brand new. Soon enough, it will be as easy to install as any other package.
Using Mandrake Cooker (that's the dev version) you already can enjoy this. I had to do nothing except
urpmi mplayer
or alternatively, if you like doing it with your mouse, simply check the box right of mplayer in the graphical installer and press the Install button (btw, you have to have non-USA aka PLF in the sources list).
The remark about mplayer hanging (for me it's at 19%, not 0%) and the need to use wget still stands, but that is only a simply bug in an prerelease which will not last for long.
So in short: be cautious when you use the word "never". It may turn out to mean "soon".
Patches for even earlier kernels are available over at that site: http://www.kernel.org/pub/linux/kernel/v1.3 /patch- 1.3.99.gz
It's an odd number, but the development release thing only started with the 2.x kernels.
Wrong. That started at least with 1.3 as development kernel. Among other things, you can see that by the fact that the LATEST link points to 2.0. AFAI can remember, even 1.0 was already a stable and 1.1 a development kernel.
I was not able to quickly come up with a reference to an announcement of that time, but the fact that there are only so few releases for 1.0 does make it look a lot like a stable tree.
Granted there hasn't been a NEW patch for 1.3 since 1996, but that's because 1.3 is DONE. There's nothing left to fix. I defy anyone to find a bug.
A slew of all-new features were implimented and it became 2.0.
Well, that's equally wrong. 1.3.x has been a devopment kernel, indeed (see above). And the fact, that there was no patch since 1996 is because the stable, 2.0.0, has been released in 1996. If you want to have a bug for 1.3, simply look for the last security bug in 2.0. Easy as that.
According to the link you cited, Microsoft Windows 2000 will be supported for 5 years and some days. 2 extra years under "extended support". One could argue that this was what the original poster meant.
Regardless, although the original comment was about release date, the real matter is: if I buy Microsoft Windows 2000 Server today (there is no XP equivalent yet, AFAIK), it will not be supported 5 years from now, even under "extended support".
-rw-r--r-- 1 536 536 6573183 Dec 16 1997 linux-2.0.33.tar.gz
Still using Linux 2.0.33?
That came out five years ago. Want to count "service packs?" Okay. Still using 2.0.x?
You're not?
Sure. On several servers.
With Linux, if you wanted new filesystem support, SMP, faster I/O, etc., you needed to upgrade.
Yeah, but most of that I could only use when I upgrade the hardware. No reason to install a newer kernel (major release), if the server does its job fine.
Another, more important point. 2.2.x is almost 4 years old. Ask the same question this way: Still using 2.2.x? How much is the chance that it is still used 14 months from now? I guess a lot of people would answer that question positive.
But how do you come to the conclusion that this rule affects the private key?
A private key is part of the linker.
I presume the private key is not something that is normally "distributed [...] with the major components (compiler, kernel, and so on) of the operating system"
Some proprietary system vendors distribute a binary-signing key to authorized developers along with the rest of the official compiler toolchain.
Well, "authorized developers" does not go well with "normally distributed". IMHO, the GPL section is quite clear on that: As exception, you are not required to distribute what commonly comes with your operating system. The exact wording:
The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable. However, as a special exception, the source code distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable.
IMHO, you can only claim the exception, if anyone who gets/busys the Standard Linker, also receive that key. If this is indeed the case, the private key is nothing special anymore. At most one could try to restrict the use of the key by some strange license.
However, if I take that word by word, a BIOS is not running on any operating system, therefore the exception rule does not apply and you may only distribute a GPL'ed BIOS if you distribute the complete tool chain to create it under the GPL.;-)
but it seems that the GNU GPL, section 3, specifically excludes any software that came with the OS or the compiler toolchain from the requirement of distribution of source code.
Correct.
Because the linker signs the app, those who distribute signed binaries of GPL'd software do not need to distribute...
the linker.
...the system vendor's private key.
But how do you come to the conclusion that this rule affects the private key? I presume the private key is not something that is normally distributed [...] with the major components (compiler, kernel, and so on) of the operating system.;-) [GPLv2, 3.]
I am not yet positive, whether a private key used to sign a binary is to be regarded source code for that binary (in the widest sense that could mean signed RPMs break the GPL), but the reasoning you provided above is simply broken.
Re:I was lucky enough to see one of these in actio
on
Transmeta Astro Processor
·
· Score: 2, Insightful
I am sad that I have no moderator points currently. That is as much troll as it can get. Well, in order to reduce the chance that anybody believes that stuff...
The lack of sse2 support greatly hindered this chip in any fps demo [...] can only run at agp 2x, which again can hinder game performance.
It obviously did not occur to you that Transmeta chips are mainly for notebooks and notebooks usually are not intended for heavy gaming. And so, that Transmeta maybe is not targeting power gamers.
Btw, such a notebook would be more usable for games than my desktop PC (Athon 700, 256MB) - and I have no problems with current games regarding a GeForce 4200 64MB.
it was quite unstable, requiring a reboot in windows98se after just 2 hours
Using Window98SE as reference platform for CPU stability. *rotfl*
Aside from that, the chip is in development. Ever heard that this may mean that it may be more unstable than the final version?
I have also heard, from reliable sources
I hope that they are not as reliable as the conclusions in your posting.
For power desktop use forget about using this chip
Who claimed that it is intended for power desktop use. Well, however, you may have found the single one usage it is not applicable resp. thought for.
Look at the Ecplise IDE however if you want to see a client side graphical Java app running just as fast as C.
Huh? I have used it under both, QNX 6.2.0 and MS Windows 2000. Under QNX it is barely usable (hardly optimized JVM, I guess), under MS Windows 2000 it is still definitely slower than a native application. Above-second reaction times (e.g. switching perspectives) are surely not "as fast as C" - although, of course, one can write slow code in any language.
[...] linux will seriosuly dent microsoft when someone can use linux completely, satisfactorially, on a daily basis, in all aspects of use and NEVER HAVE TO TOUCH A COMMAND LINE INTERFACE FOR A SECOND. or even know one exists! [...]
[... the footer:] C:\>tracert life.liberty.pursuit-of-happiness
if someone breaks into an ftp server, they might as well replace the md5 signatures, too. a better solution would be signing the sources with a gpg key.
Btw, it's enough to sign the md5sums. That gives the best of two worlds: the speed of a "simple" checksum (md5) and the security of real cryptographic signature.
If those people redistribute the binary then they have to make the source available - not people further up the chain.
It depends.
1) If you gave them the source when you gave them the binaries, they have to make the source available, if they distribute it further.
2) If you do not give them the source along with the binaries, but a note explaining how you guarantee that they can retrieve the source from you, your note has to be valid for all third parties, too.
3) If you only received a note according to 2) and you do a non-commercial distribution, you may pass along the note in order to comply (this, btw, requires the note to be valid for third parties).
For completeness, here are the relevant GPL excerpts:
regarding 1) above: 3. You may copy and distribute the Program [...] in [...] executable form [...] provided that you also do one of the following:
a) Accompany it with the complete corresponding machine-readable source code, [...]
regarding 2) above: 3. b) Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, [...]
regarding 3) above: 3. c) Accompany it with the information you received as to the offer to distribute corresponding source code. (This alternative is allowed only for noncommercial distribution and only if you received the program in object code or executable form with such an offer, in accord with Subsection b above.)
The problem with "loser pays" is that it favors the rich. A rich entity can sue a poor entity, and under "loser pays," at worst, the rich entity will have to pay the poor entity's legal bills, which are going to be tiny compared to the large entity's bills. But if a poor entity brings a valid suit against a rich entity, but gets out-lawyered, then the poor entity now has to pay a gigantic amount of money, which it does not have, since it is poor.
I cannot speak for the whole of Europe, but only Germany. Here, AFAIK, it depends on the kind of case how the costs are distributed.
But more important, how much costs a party can make accountable, is quite fixed, i.e. if you decide to pay for 10 layers instead of one, that is your problem. If you decide to pay your laywer 10-fold of what is common for such a case, again, that is your problem. For all legal costs, you want the other party to pay for in addition, you have to proof that they were necessary (e.g. if you hired a private investigator: if his proof was not needed to decide the case, it's your costs. if it was needed, because the other party lied, it is the other party's cost). Of course, the actual rules are more complicated, but you get the idea.
In most private cases the costs are distributed according to "who owes how much". If you sue for EUR 1000, but only get EUR 700, the other party has to pay 70% of the legal costs and you have to pay 30%, usually. That is, because you were with 70% right, that the other party owed you EUR 1000 and the other party was with 30% right, that it did not owe you that much.
That way, if someone decides to sue me, but has no stand at all, my risk is only my time and nerves. This system still leaves the advantages to the rich, because they don't mind the costs as much as I, i.e. the risk is the same, but the perceived risk is lower for the rich.
The advantage of this system is that it is not as easy to scare someone with only a threat to sue (I mean, if it wouldn't hold in court), because defender hires a laywer and knows, there is not much trouble and the costs will be paid for by the offender at the end.
For real cases, I mean cases, where both side really think they are right, the risk is about the same as in the US, of course.
But mod_php is installed so often, because PHP is so popular
The above is ripped out of context. Which would be rather obvious, if you had quoted the original (Why did you took the effort to change the case manually and therefore forge it to look like a stand-alone statement?).
mod_php is installed so often because it is easier and cheaper for ISPs. The problem with mod_perl is that it is too powerfull[...] In short to provide mod_perl hosting ISP have to give each its customer dedicated Apache server.
Regardless, if PHP would not be popular, no ISP could satisfy its customers by only installing mod_php. The one goes with the other. In fact, there seems to be no problem with mod_perl: see my sister post, which I suppose you have seen when you read mine. It looks like either your conclusion is wrong, or the installation base is not bound to ISPs. Whatever.(Take the treatment of stats from my sister post, too.)
Slashdot Mirror
Not to be a troll, but this guy seems to be claiming he knows what he's doing. From his writeup, I very much doubt that.
Yeah, I thought the same thing. He goes out to test XBox as a cluster and then mainly bitches about how hard it was to setup (minor things like having to install gcc himself; btw, it's Mandrake, that means urpmi gcc should have worked?!). And he complains about modding being bothersome. After that I expected him at least to mention the outsight that installing will be easy in the future, if the XBox Linux project succeeds.
For the actual performance results he reserved less than a paragraph. Sad. Also, I'd like to see the MPI program he used for testing. Results for a "normal" PC cluster and for a single machine cluster would have been interesting for comparison, too.
PS: Regarding the kernel compile time - maybe DMA for the disk was not enabled?
I do not look forward to the day where I have to talk somebody through a command-line problem.
I don't see your point. How is this different from tweaking the registry or having to run ping in the console (what I see Windows users doing here regularly on trouble-shooting)?
I just told you, you have a GUI for most issues with recent distros. Sure, some problems still have no GUI solution, especially if the software is brand new (but end-users shouldn't use alpha software anyhow), but that is the same for Windows, IMNSHO.
[...] The monopoly agrument is very stale.
I expected that rebuttal. That was why I wrote there are several reasons.
I did mention it, because it is a real concern for us. If we could do as we want, we would be a complete Linux/FreeBSD/Mac shop. But half of our desktop computers/notebooks have Microsoft Windows installed solely or as dual boot because we have to.
Every time I suggest Mac or Linux when I install a computer for an aquaintence, I get to hear, they want Windows because that is what everyone uses and that are the computer skills employers ask for. So it may an old argument, but at least for my company and me personally, it is not stale, but going bothering every day.
Well, that's what recent was about.
It suffers from serious lack of refinement. It's hard to find anything (In KDE, nearly every app begins with K),
I can only speak for Mandrake and Debian, because that's what I regularly use. In Mandrake, there is a menu which let's you choose the application by the task you want to accomplish, not by name. It is not perfect yet, but it's a good start.
and I've had no luck doing any real configuration. You still have to use the CLI to change desktop resolution, for example.
Last time I needed to use the shell for that was several years ago. If Redhat really has no GUI for this, it sucks more than I thought. But it sounds more probable that you are out of date.
Installing something is a cruel joke. The RPM Packager thing-a-ma-bob is not a very clear UI, nor is it immediately obvious why you need to go through that. Coming from the Windows world where you use setup.exe and pick where you want your stuff to go, RPM is full of un-necessities.
I can only tell for Mandrake, but there installing a packet does not involve rpms in any visible way. You simply choose in the GUI the packet to install, it will ask you to insert the CD or it will download it and that's it. No unnecessities involved.
Btw, Windows installer regularly asks question in such a cryptic way, that the uninitiated don't know what the question is about at all (ever hade a code version conflict - it happens regularly with non-English versions).
So no, it's not a matter of 'just being different', [...]
Oh yes, it is. You just proved it yourself. You are obviously coming from a Windows background and comparing to that.
My point was about people who are not used to any OS. I am the "computer guy" here and from the cries for help I get and the misconfiguration people make, I can tell you that Microsoft Windows is a lot more confusing than you admit - for quite some people.
Microsoft has put a great deal of research effort into making a usable UI.
Yes. I never said anything contrary. Microsoft has done a lot to further usability. I just claimed that Linux distros have catched up in recent versions. And that the stuff people still complain about is mostly that it simply does not behave as they are used to. Different. Not more difficult. See?
Like or hate it, millions upon millions of people are able to make use of it.
The amount of users does not imply anything regarding usability, since usability is only one of several reasons to buy software. Microsoft being a monopoly another one amongst others.
Anything about table inheritance or triggers?
m l suggest that triggers will come somewhere in 5.0 probably. Heard nothing about table inheritance, though.
:-)
Well, http://www.mysql.com/doc/en/ANSI_diff_Triggers.ht
I remember once mysql said they would never implement stored procs or triggers or referential integrity,
I am interested in MySQL since when 3.21.x was current (1997?) and cannot remember such a statement. Surely, once they said, that they have no plans to do so. But that is not the same: don't having it on the list of things to be done is quite different from claiming to never implement it.
Also they often said, that when any of that feature is to come, it may not make existing stuff slower or it will be at least optional (and probably disabled by default).
AFAICT, the facts they created in the meantime do not contradict themselves yet.
it looks like they changed their minds.
OTOH, there would be nothing wrong with changing your mind once every 5 years or so.
I grant you this for Microsoft Windows users. They don't know better. But the original poster spoke about clean and clear interfaces, and I can tell you that not one of my technically non-savvy acquaintances considered the GUI of MS Windows clean and clear when first confronted with it.
Most of them found it very confusing and some of them still do not dare "playing" around, because always "strange, unexpected things" happen.
The UI of recent Linux distros are not that bad anymore. They just differ from what most people know, i.e. Microsoft Windows.
Nothing? OK, not quite nothing, but even Mozilla (one of the finest examples of recently complete software) is still being released 'for testing purposes only.'
Ouch. One of the worst examples you could have chosen. Mozilla is not intended to be ever released as end user software. That they leave to others. Mozilla is a browser and also a toolkit to base other software on (like Netscape).
So yes, their releases will always be for testing purposes only, because that is their stated goal.
If they were a closed-source entity, they would never make public releases, the same as you don't see public releases of the general base your BIOS is specialized from.
Stop packaging Linux as a crappy Windows-clone (see Mandrake), and people will stop making the comparison.
Huh? To stay a bit more on-topic, i.e. related to the article: Because they are already packaged, on Mandrake it is a matter of some mouse clicks to install any of the named players. Very crappy, yes. (the crappy GUIs of the players are a different story)
So the installation hurdle JWZ complained about does not exist in recent distros (I simply presume that RedHat, Debian and SuSE do as well, meahwhile). In other words: using an older distro to judge the current state of video on Linux is more than a bit strange.
Huh? Why everyone in this thread seems to concentrate on the wording? Anyone using databases should know that he simply meant a SQL dump. For the clueless: creating SQL commands, which will create a copy of the database in question, when run through an appropriate tool (for running SQL commands in batch) and writing them to a file.
You'll need to write scripts in any case.
Huh? Why? Any open source database comes with the tools to run such a SQL dump. And most commercial ones, too, I presume.
I've spent so much time lately in the (relatively) flat-table world of MySQL that I had forgotten about inherited tables, subselects, constraints in table definitions,
;) Looks like it is time to revisit postgres, especially for some db-agnostic PEAR apps I'm building. For me, it's the subselects that really make it worth the effort.
;-), but only because you sounded interested in MySQL I thought I update you on the fact that those features you are missing are under full work currently.
Umm. Constraints are there with InnoDB for what? 2 years?
and oh yes, vacuuming.
Well, or you might have a look at the current alpha (version 4.1) were basic sub-selects and derived tables are already running. Though, without much optimization yet.
That is not meant as shameless plug (although it probably is
> If you are robbed, you lose some property.
And you get tied up and beat.
And your wife is murdered.
And your daughter is raped.
But its only property, right?
Well, as was already said in another comment, a gun mostly surprises and scares the unprepared. If "they" get into your house with the intend to rape and murder, I think there is good reason to believe they do not come unprepared.
I agree, in a worst case scenario you can get things only better by defending your family in any way possible, including a gun.
But most cases are not worst cases and there is a good chance that using a gun lets things escalate unnecessarily. That is what the parent post was about, IMHO.
Isn't it funny, that Texas, a state where anyone can carry a gun as long as its *plainly visable* has the lowest amount of fatal gun violence in the U.S.A. But Maryland, where carrying a gun is illegal for private citizens, is one of the highest ?
;)
...) which show that you can live relatively safe without everyone being allowed to own a gun.
I guess you are mixing up cause and effect. I don't know any details, but I would bet, that carrying a gun in Maryland has been outlawed *because* it has one of the highest amount of fatal gun violence. Not the other way around.
Take for example Germany, where I live: You need a license in order to own most weapons (not limited to guns). It is non-trivial to get a license and becomes the harder the more dangerous the weapon is - some weapons are banned completely. E.g. there is no way at all to get a license for a machine gun, AFAIK.
Most guns are owned and used in shooting clubs for sportive reasons (and only licensed for that use) and are specially crafted for sportive purposes.
Even if you own a (normal) license, it is illegal to carry a *loaded* gun. You have to transport gun and munition seperated. There are only few people who possess a license to carry loaded weapons. Mostly police and such. It is very unusual that security personal carries guns, except maybe for bodyguards.
I don't know anybody owning a gun and I don't remember when the last time was I saw someone with a gun (police officers excluded) and I don't think I ever saw somebody with a drawn gun (police officers included). Except in TV, of course.
If someone would hold a gun, I would be quite scared, too, because it is so uncommon here.
I don't mean to claim that allowing guns necessarily means that there will be more violence. But saying the opposite is nonsense. There are a lot of countries (England, France,
E.g. in one year, in Germany about 400 people are killed by guns (not limited to crimes, i.e. including people killed by police). And about 1200 people die by homicide (not only guns). USA, which has about 3 times the population has about 11.000 people killed by guns only.
Well, could be that it is better by now.
But when I tried it (first quarter this year), I got too frustrated with setting up. If there is an easy way to avoid the part about enforcement, I missed it.
Maybe the problem is only that the documentation is too centric about that and ignores that some people would try Aegis for other reasons. Maybe I just was too unpatient.
But in the end, I dropped arch, because I got the impression that "skipping and ignoring" the steps got too burdensome. Probably this is, because Aegis emphasis too much (for my needs) to be more than a revision control system.
What turned me away from Aegis is the fact that you have to adapt a policy at all. Aegis sounds great if you want to enforce a process.
But if you are fine to do without a rule-based process (maybe you have a small team, or very disciplined developers, or whatever), Aegis looks as if it brings a lot of unnecessary load with it.
Aegis is a source management system (and enforces it), while I only want a revision control system.
This alone is why Linux is never going to be a mainstream OS... WHO WANTS TO DO ALL THAT SHOT JUST TO WATCH A QUICKTIME!!!!???
First understand that this is not a Linux problem per se, but an Apple one. I would be curious what you would do on Microsoft Windows, if the QT plugin would be only published for Linux. Don't laugh: consider what could be in 5, 10 or 20 years. So my real point is, within some years at most, Linux versions for apps will be published by default.
Second, understand that you have to go through these install steps, because QT support in mplayer is brand new. Soon enough, it will be as easy to install as any other package.
Using Mandrake Cooker (that's the dev version) you already can enjoy this. I had to do nothing except
urpmi mplayer
or alternatively, if you like doing it with your mouse, simply check the box right of mplayer in the graphical installer and press the Install button (btw, you have to have non-USA aka PLF in the sources list).
The remark about mplayer hanging (for me it's at 19%, not 0%) and the need to use wget still stands, but that is only a simply bug in an prerelease which will not last for long.
So in short: be cautious when you use the word "never". It may turn out to mean "soon".
Patches for even earlier kernels are available over at that site:3 /patch- 1.3.99.gz
http://www.kernel.org/pub/linux/kernel/v1.
It's an odd number, but the development release thing only started with the 2.x kernels.
Wrong. That started at least with 1.3 as development kernel. Among other things, you can see that by the fact that the LATEST link points to 2.0. AFAI can remember, even 1.0 was already a stable and 1.1 a development kernel.
I was not able to quickly come up with a reference to an announcement of that time, but the fact that there are only so few releases for 1.0 does make it look a lot like a stable tree.
Granted there hasn't been a NEW patch for 1.3 since 1996, but that's because 1.3 is DONE. There's nothing left to fix. I defy anyone to find a bug.
A slew of all-new features were implimented and it became 2.0.
Well, that's equally wrong. 1.3.x has been a devopment kernel, indeed (see above). And the fact, that there was no patch since 1996 is because the stable, 2.0.0, has been released in 1996. If you want to have a bug for 1.3, simply look for the last security bug in 2.0. Easy as that.
In fact, Windows 2000 will not be supported five years after it's release date.
Yes it will.
According to the link you cited, Microsoft Windows 2000 will be supported for 5 years and some days. 2 extra years under "extended support". One could argue that this was what the original poster meant.
Regardless, although the original comment was about release date, the real matter is: if I buy Microsoft Windows 2000 Server today (there is no XP equivalent yet, AFAIK), it will not be supported 5 years from now, even under "extended support".
-rw-r--r-- 1 536 536 6573183 Dec 16 1997 linux-2.0.33.tar.gz
Still using Linux 2.0.33?
That came out five years ago. Want to count "service packs?" Okay. Still using 2.0.x?
You're not?
Sure. On several servers.
With Linux, if you wanted new filesystem support, SMP, faster I/O, etc., you needed to upgrade.
Yeah, but most of that I could only use when I upgrade the hardware. No reason to install a newer kernel (major release), if the server does its job fine.
Another, more important point. 2.2.x is almost 4 years old. Ask the same question this way: Still using 2.2.x? How much is the chance that it is still used 14 months from now? I guess a lot of people would answer that question positive.
But how do you come to the conclusion that this rule affects the private key?
;-)
A private key is part of the linker.
I presume the private key is not something that is normally "distributed [...] with the major components (compiler, kernel, and so on) of the operating system"
Some proprietary system vendors distribute a binary-signing key to authorized developers along with the rest of the official compiler toolchain.
Well, "authorized developers" does not go well with "normally distributed". IMHO, the GPL section is quite clear on that: As exception, you are not required to distribute what commonly comes with your operating system. The exact wording:
The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable. However, as a special exception, the source code distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable.
IMHO, you can only claim the exception, if anyone who gets/busys the Standard Linker, also receive that key. If this is indeed the case, the private key is nothing special anymore. At most one could try to restrict the use of the key by some strange license.
However, if I take that word by word, a BIOS is not running on any operating system, therefore the exception rule does not apply and you may only distribute a GPL'ed BIOS if you distribute the complete tool chain to create it under the GPL.
but it seems that the GNU GPL, section 3, specifically excludes any software that came with the OS or the compiler toolchain from the requirement of distribution of source code.
...the system vendor's private key.
;-) [GPLv2, 3.]
Correct.
Because the linker signs the app, those who distribute signed binaries of GPL'd software do not need to distribute...
the linker.
But how do you come to the conclusion that this rule affects the private key? I presume the private key is not something that is normally distributed [...] with the major components (compiler, kernel, and so on) of the
operating system.
I am not yet positive, whether a private key used to sign a binary is to be regarded source code for that binary (in the widest sense that could mean signed RPMs break the GPL), but the reasoning you provided above is simply broken.
I am sad that I have no moderator points currently. That is as much troll as it can get.
Well, in order to reduce the chance that anybody believes that stuff...
The lack of sse2 support greatly hindered this chip in any fps demo [...] can only run at agp 2x, which again can hinder game performance.
It obviously did not occur to you that Transmeta chips are mainly for notebooks and notebooks usually are not intended for heavy gaming. And so, that Transmeta maybe is not targeting power gamers.
Btw, such a notebook would be more usable for games than my desktop PC (Athon 700, 256MB) - and I have no problems with current games regarding a GeForce 4200 64MB.
it was quite unstable, requiring a reboot in windows98se after just 2 hours
Using Window98SE as reference platform for CPU stability. *rotfl*
Aside from that, the chip is in development. Ever heard that this may mean that it may be more unstable than the final version?
I have also heard, from reliable sources
I hope that they are not as reliable as the conclusions in your posting.
For power desktop use forget about using this chip
Who claimed that it is intended for power desktop use. Well, however, you may have found the single one usage it is not applicable resp. thought for.
Look at the Ecplise IDE however if you want to see a client side graphical Java app running just as fast as C.
Huh? I have used it under both, QNX 6.2.0 and MS Windows 2000. Under QNX it is barely usable (hardly optimized JVM, I guess), under MS Windows 2000 it is still definitely slower than a native application. Above-second reaction times (e.g. switching perspectives) are surely not "as fast as C" - although, of course, one can write slow code in any language.
[...] linux will seriosuly dent microsoft when someone can use linux completely, satisfactorially, on a daily basis, in all aspects of use and NEVER HAVE TO TOUCH A COMMAND LINE INTERFACE FOR A SECOND. or even know one exists! [...]
[... the footer:]
C:\>tracert life.liberty.pursuit-of-happiness
no comment.
if someone breaks into an ftp server, they might as well replace the md5 signatures, too. a better solution would be signing the sources with a gpg key.
Btw, it's enough to sign the md5sums. That gives the best of two worlds: the speed of a "simple" checksum (md5) and the security of real cryptographic signature.
If those people redistribute the binary then they have to make the source available - not people further up the chain.
It depends.
1) If you gave them the source when you gave them the binaries, they have to make the source available, if they distribute it further.
2) If you do not give them the source along with the binaries, but a note explaining how you guarantee that they can retrieve the source from you, your note has to be valid for all third parties, too.
3) If you only received a note according to 2) and you do a non-commercial distribution, you may pass along the note in order to comply (this, btw, requires the note to be valid for third parties).
For completeness, here are the relevant GPL excerpts:
regarding 1) above:
3. You may copy and distribute the Program [...] in [...] executable form [...] provided that you also do one of the following:
a) Accompany it with the complete corresponding machine-readable source code, [...]
regarding 2) above:
3. b) Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, [...]
regarding 3) above:
3. c) Accompany it with the information you received as to the offer to distribute corresponding source code. (This alternative is allowed only for noncommercial distribution and only if you received the program in object code or executable form with such an offer, in accord with Subsection b above.)
The problem with "loser pays" is that it favors the rich. A rich entity can sue a poor entity, and under "loser pays," at worst, the rich entity will have to pay the poor entity's legal bills, which are going to be tiny compared to the large entity's bills. But if a poor entity brings a valid suit against a rich entity, but gets out-lawyered, then the poor entity now has to pay a gigantic amount of money, which it does not have, since it is poor.
I cannot speak for the whole of Europe, but only Germany. Here, AFAIK, it depends on the kind of case how the costs are distributed.
But more important, how much costs a party can make accountable, is quite fixed, i.e. if you decide to pay for 10 layers instead of one, that is your problem. If you decide to pay your laywer 10-fold of what is common for such a case, again, that is your problem. For all legal costs, you want the other party to pay for in addition, you have to proof that they were necessary (e.g. if you hired a private investigator: if his proof was not needed to decide the case, it's your costs. if it was needed, because the other party lied, it is the other party's cost). Of course, the actual rules are more complicated, but you get the idea.
In most private cases the costs are distributed according to "who owes how much". If you sue for EUR 1000, but only get EUR 700, the other party has to pay 70% of the legal costs and you have to pay 30%, usually. That is, because you were with 70% right, that the other party owed you EUR 1000 and the other party was with 30% right, that it did not owe you that much.
That way, if someone decides to sue me, but has no stand at all, my risk is only my time and nerves. This system still leaves the advantages to the rich, because they don't mind the costs as much as I, i.e. the risk is the same, but the perceived risk is lower for the rich.
The advantage of this system is that it is not as easy to scare someone with only a threat to sue (I mean, if it wouldn't hold in court), because defender hires a laywer and knows, there is not much trouble and the costs will be paid for by the offender at the end.
For real cases, I mean cases, where both side really think they are right, the risk is about the same as in the US, of course.
But mod_php is installed so often, because PHP is so popular
The above is ripped out of context. Which would be rather obvious, if you had quoted the original (Why did you took the effort to change the case manually and therefore forge it to look like a stand-alone statement?).
mod_php is installed so often because it is easier and cheaper for ISPs. The problem with mod_perl is that it is too powerfull[...] In short to provide mod_perl hosting ISP have to give each its customer dedicated Apache server.
Regardless, if PHP would not be popular, no ISP could satisfy its customers by only installing mod_php. The one goes with the other. In fact, there seems to be no problem with mod_perl: see my sister post, which I suppose you have seen when you read mine. It looks like either your conclusion is wrong, or the installation base is not bound to ISPs. Whatever.(Take the treatment of stats from my sister post, too.)