Mosfet is a emotionally unstable GUI hacker. His knowlege of the long history and tradition of UNIX administration is pathetic. He ignores simple observables like PATH searches are more expensive than bin lookups. One executable dir per App would be FAR SLOWER than 2000 executables in a single dir.
This is another classic example of not letting programmers, especially GUI progrmmers, be involved in OS design.
For those of you who might be swayed by his foolish arguemnts, please read LHS, and the last decade of USENIX papers and LISA papers. Unix systems organization has been openly and vigorously debated for 15years. It has not be dictated by mere programmers from high on above like MS. And RedHat is to be applauded for properly implementing the FHS which is a standard, others like SUSE should be encouraged to become compliant (/sbin/init.d... mindless infidels:).
Read a book, get a clue, whatever. "cryptography as we know it is obsolete" puhlease.
Nobody has touched RSA, or RSA key exchange. The problem is the crytographic protocols and implementation. The protocols often use strong (re: compute intesive) cryptography to exchange weak (re: fast) cryptographic keys. It's the weak shit that is getting cracked. Also Protocols and implementations let information to leak about the weak keys. This alows cracks to exploit the leaked info, to find the private keys faster (re: polynomial time) than the standard approaches.
Someday these corporate dumbasses, who want to limit fundementally unlimited resources, will get a clue and start creating secure implementations. That is when you should be afraid.
Maybe they'll start developing their protocols and implementations in an open process. This would have a much better chance of flushing out the bugs . The only thing they need to really control is the master keys. A smart plan would be to develope the code in a open manner, then and only then start encrypting all their music, video, articles, with the super-secret "Master" keys.
We are lucky that they are following the security thru obscurity developement process. How many times will they fail before they start to wake up?
Sony is working on encryption to the monitor. Others are working on encryption to the speakers (usb speakers with DSPs in them). Combine that with a cluefull open developement process, and we are SCREWED in a squeal-for-me-boy kinda way.
Then we start hoping someone with the real keys reveal them (the whistle blower way). Or the big corperate dumbasses leave the keys on a not-quite-so-secure system and hackers release them.
The cryptographic algorithms aren't being attack it is all the dressing PROTOCOLS, IMPLEMENTATIONS, and PEOPLE.
ext2 doesn't have a 2GB file size limit. That was a operating system limit which went away somewhere in the middle 2.2.x stable series.
Further, ext3 is not the-next-version-of-EXT. It is an extention of ext2 which is fully compatible with ext2. Think of ext2 as two things: the format of bits on the disk, and the code to read/write those bits. Ext3 keeps the same format (actually with compatable extentions), but mostly it changes the code for reading/writing to the disk (journelling).
The ext2 filesystem is tried and true. You can go back and forth between ext2 and ext3 with no reformating or issueing of commands other than the mount command.
ReiserFS is a more "sophisticated" filesystem than ext[23], and XFS is a more "sophisticated" filesystem than ReiserFS. But I keep "sophisticated" in quotes because the utility, reliability, and speed of a FS relies more on your usage patterns, than on the genius of the filesystem designers/coders.
I use ext3 at home. Good speed, no need to tar up all my files..reformat drives..untar all my data, journelling, mainline kernel support, tried and true.
One place I would seriously consider ReiserFS is for home directories. The place it really shines is constantly reading and writing lots of "small" files (small ~50k). For Gnome and KDE config files, Mozilla disk caches, CVS checkouts, and untaring of source, ReiserFS is going to be a leader of the benchmarking pack. You'll notice the difference.
But don't get into holy wars over FS, and don't think that Linux is whole generations behind Commercial Unixen. Linux Kernel is dramaticallly ahead in some areas and minorly behind in others. The only place it is dramatically behind is places where the computer you are running the OS on cost more than a half million dollars.
Re:Comment from a real PERL programmer
on
Perl6 for Mortals
·
· Score: 1
I just had to throw something into my post to keep the "my-favorite versus your-favorite" fight going:).
I follow java because it is my favorite bondage and discipline language. (BTW I use "B&D language" as the inverse pejorative to "scripting language").
Ignoring the idea that the number of lines means anything (you or I could have streatched out the perl code). But if you read that Java 1.4 example of word counting with regex and char strings, you should have notice how complicated Sun has made this kind of thing. I was stunned. CharacterSet("ISO-8995-1") WTF!?! Creating a immutibale object that represents the number one and calling it ONE. Something went wrong in their language design.
I realize Java is growinig past it's initial quicky-design. I just wish all the OO-lusers would stop critisizing Perl for the same issues. Ruby is the one language that hasn't gotten crufty, yet. And that is cuz it's young.
Comment from a real PERL programmer
on
Perl6 for Mortals
·
· Score: 3, Insightful
I program large complex programs in perl. I use OO Perl all the time. I write small once off scripts in perl as well.
Some of this Perl 6 stuff scares me too. Mainly because I think perl can be abused to write bad code. I am thinking stuff that is REALLY obtuse. I've seen code with $|++. Which is stupid. Because if $| == 1, then the code doesn't do anything and the inverse $|-- fails to achieve your purpose when $| == 2. STDOUT->autoflush(1) is the clear way to write it.
Just because dumb-ass "programmers" CAN write obscure code in perl, doesn't invalidate the value of Perl. Any language with expressive power is vulnerable to having "Obfusicate-X" contestants write programs in that language. A wise quote: "Fortran programmers can write Fortran in any language".
Perl 6 is looking to be the exact opposite of LISP. In my view, LISP has little or no syntax; just Lots of Incessant Silly Parenthesis. Well it looks like perl 6 is going to be nothing else but syntax.
This might be valid perl6:
%b{@a}:= ^-x @a ^_ '.c';
I like perl by this might be to much for me.
Of course the real reason I use perl is two fold; it's expressive power (unlike bondage and discipline languages) and CPAN (the killer feature).When I look at other languages like python or ruby, I look for their CPAN equivalent. Right now their is none, but maybe soon.
BTW, for the JAVA fans out there the following url is the same code as:
$/ = undef;
$wc{$_}++ for split(/\W+/,);
print($_, " = ", $wc{$_}) for sort keys %wc;
48 lines (took out comments and empty lines) versus 3.
BTW, This is as obfusicated as my code gets. I did it mostly for brevity.
Re:what a predicament ...
on
Linux Turns 10
·
· Score: 1
You need to understand one thing from reading the Kernel Mailing List.
The problem is that when _ALL_ the RAM is consumed and _ALL_ the swap is consumed ON 2GB+ RAM machines, THEN the computer goes into conniptions.
That is a very specific recipe. Linux should not have this problem; it should just degrade performance gracefully; but it does have this problem. There is a fundemental problem here. They kernel guys may be able to tune the VM to avoid the problem. But depending on the tuneing a malicious program could still cause the catelepsy.
My suggested work-arounds for the whinners with more than a GB of RAM:
- DON'T EXAUST ALL YOUR MEMORY AND ALL YOUR SWAP. If you start hitting swap consistantly, stop!
- Make sure users run with ulimits. Make them real big, but less than everything.
Disclaimer: I've never had a clearance, and this information isn't really a secret.
The Hubble space telescope was built using the super-structure of a line of spy satalites. A fella-who-knows told me that when they first discovered the optics problem, many people's first suspicion was that they installed a mirror from the spy satellites. This wasn't the case.
BTW, the difference btw the spy satellites and the Hubble is that the Hubble is focused for inifinity and the spy satellites are focused for a few hundred miles or so (and towards the earth, which would burn the Hubbles sensors).
Another thing, (prolly just a myth) is that the Hubble is the 13th of this line of spy satellite; Fun for the numeralogically inclined:).
Last thing, the US regularly announces satellite launches that don't exist, and skips numbers for the satellites. "SS1...SS2...SS3...SS5..Hey! how'ed they slip SS4 passed us?". They didn't their just fucking with you.
Try not to think of it as a Shield. What Bush is proposing is not even remotely comprehensive defense from large numbers of missles. It is meant as a point defense against less than 5 launches.
Futher, the missle defense systems are most likely to target the boost phase of the missle launch. We're supposed to be sitting near by the bad-missile-launching-country at see in military vessles carrying boost-phase anti-missle wepons.
The kind of defenses envisioned in Bush's VERY LIMITED anti-missle system would have no applicability to anti-meteor defense. Basically, unless we get a few years advanced warning about metor impact, we're screwed big time.
More money in Washington means more corruption, lobbying etc.
The way to end it is not to pass another set of useless regulations
This is the kind of all-regulation-is-bad thinking that I wanted to distance myself from. It isn't the existance of regulations that is bad. It is the kind of regulation AND how much responsibility people think the FEDERAL government should take. I pointed to the Clean Air Act as an example of good regulations. resource/environmental management is a federal responsibility. Regulations which take advantage of markets are a good kind of regulation; as opposed to obviating markets by creating one-size-fits-all autocratic mandates.
See big companies wanted no limits on their ability to pollute. Pollution has real costs on our economy (health care, loss of economically usable forests/land, etc) and quality of life. But the companies that create the pollution are not the companies paying for the pollution (insurance companies, forresters, and farmers pay the costs of pollution). Federal government regulation is the ONLY way to push the costs back to the companies who are otherwise being subsidized by the ability to pollute.
About John McCain's legislative initiatives:
His proposed regulation of campaign finances are not going to be "useless regulations". They may not be a complete solution to the disparity in VOLUME of speech and influence by wealthy parties (big companies not little ones), but they are a good try.
His advocacy for health care reform is most certainly pro-free-market. HMOs are specifically excempt from law suits. Non-HMO health insurance companies are NOT excempt. This is not fair to the traditional insurance companies.
This disparity is not pro-free-market. Further, law suits are a nessesary part of contract law. Contract law is abso-fucking-lutely nessesary to Free Markets. This doesn't means that we don't need Tort reform. I believe we do. Companies can be blown away by a single lawsuit on an issue where they were trying to bahave in good faith (example: Dow-Corning and Silicone brest implants). This isn't healthy for a free market. On the other had I do think companies should be killed by law suits when they bahave in grevious bad faith. For instance selling a product they know will and is killing people while constantly lieing and commiting fraud to deny what they have already proven to themselves; Tabbaco and certain car design issues come to mind (not the Firestone tires; I'm thinking the Gremlin issue). Lawsuits allow the share holders to suffer, because they are the source of authority in companies.
Don't be such a simple minded partisan.
BTW, I am registered Democrat so I can improve the breed, because I think they are better stock to work from. However, in my home state, the Free Socialist Republic of Maryland, I am looking forward the opportunity to vote for a Republican governor that doesn't snarl and drool. And I vote against the stealth senator Sarbanes even if the Republican candidate were to snarl and drool.
Actually, I am a registered Democrat. I think there is a sufficiantly large contingent of Democrats that aggree with pro-free market policies, that the Democrats are not to far gone. Unfortunately, the "Blue Dog Democrates" are being decimated by the southern states' shift away from conservative Democrats in favor of Christian-Right Republicans who are whole heartedly in favor of government involvement social and religious affairs.
Republicans really aren't very good on these issues either. Bush's tax plan is basically lame as an economic policy. There was no effort to get capital gains tax reductions. The Republican attitude towards regulatory reform is basically pro-big-business. They fought long and hard against the Clean Air Act, (Yes Bush Sr. signed it, but as a comprimise) which is a model of Free Market based regulation. The Clean Air Act established a market for "carbon credits" and let market forces deal with how, when, and where the Carbon emmissions were reduced; in less then a decade the last half century of carbon pollution was rolled back. At the same time the US went thru amazing economic growth. Regulation, the environment, and economic growth are NOT incompatable. It is a matter of "How".
Further Republicans have shifted away from Free Market economic policies. The Barry Goldwater, John McCain (Goldwater's hand picked successor for his Senate seat), Gramm Rudman, etcetra branch of the Republican party are getting pushed out of the GOP in favor of the Social Conservatives (who have a very harsh all or nothnig attitude towards members of their own party).
Dashelle is not a freaky liberal, however he does seem to be to far to the "traditional" Democratic side of the Party than I care for. He is one sly bastard that is going to kick the GOP's ass; and I am not sure if that is good or bad.
Actually, waging war on the large corporations that are getting these stupid laws passed in Washington would benefit the majority of us, who work for the much smaller corporations that are the ones usually being hurt by these stupid laws.
Here! Here!
I figure no one will actually read this cuz I am posting it late, but here is my RANT...
Most of the feeble-minded "pro-business" crowd fail to recongnize this fact: 2/3 of the U.S. economy is small to medium sized businesses. There are government definitions of what this is, but basically they are less than 1000 employees and under a $100million in revenue. Consistantly the growth in the US economy comes from these buisnesses. They employ the majority of americans.
I'd argue (with ample agreement of economists) that this is THE single distinguishing feature of the US economy versus the European and Japanese economies. Japan is really the best case example of large corperations stiffling their economy. It isn't so much that large corperations are dumb and slow by nature, it is just that the aggragate behavior of many players in the market creates the pressure for efficiency gain and inovation thru a variety of mechanisms.
This all means that "monopoly-abuse-isn't-bad" and "big-companies-aren't-a-threat" ideas are wrong. These ideas are just ignorant and inaccurate.
My nit pick of the day is that the term "Capitalism" refers to the specific use of publicly traded stock and securities to fund investment in corperation building. The term that should be used is "Free Markets". Free Markets require reliable and liberal contract and regulatory law and the ability of most/all individuals in the society to participate in the market at any level.
Pro-big-buisness laws actually hurt Free Markets, because they typically create barriers to entry for smaller players.
This in a nutshell is why we have Anti-Trust laws. This is why I support campaign finance reforms (they favor big businesses over small-medium businesses; look at the "sponsers" of the Democratic and Republican Conventions).
If you want a stong economy, support small and medium buisnesses. To this end support:
Consistant Regulation (not nessesarily none). This allows for buisness planning.
Less paper-work and procedures for regulation. Beauracracy favors the large corperations and lowers effiency.
Generally less regulation (again: not none). More regulation, creates bariers to entry.
Lower capital gains taxes. This reduces the friction of capital to move to new endevours.
Open reporting of regulatory processes, legistlation, and public corperation finances (SEC filings). BTW, this is the one true arguement coming from all those anti-WTO protests. The WTO and other multi-national trade organizations are not as open to democratic review as they SHOULD be.
BTW, I am a self professed liberal, but I strongly believe in democratic free markets. Free Markets are the best way to help the most people; Free Markets require strong (but minimal) democratic regulatory environments to flourish (despite Anne Rand spouting Libertarian fairy-tales)
I have dipped into bad behavior before. Now I manage techies.
First of all, those who won't comment code and document design should be beaten severly about the head with a LART. That is bad coding/design practice and is completely unacceptable. Put them on undocumented code. When they bitch have them explain why it is bad and how the company should fix it. Listen to the response and implement. And you might notice their attitude changing. If it doesn't replace them.
Second, give them responsibility. I was once a camp counseler. I took the jokers/ring-leaders and put them in charge; small things at first then bigger things. They dig the power and are usually the most effective leaders. Again, if it doesn't work replace them.
Last, and most imporant. Have a frank conversation with them and respect their opinions. There is nothing as powerful as a little respect. Small reasonable acts of control and general respect are the best way to get people with the program. If it doesn't work, what's left? Fire them.
Re:Object Oriented programming is overrated
on
Why not Ruby?
·
· Score: 1
Exception handling has nothing to do with OOP; it is an entirely orthogonal feature!
You are wrong. Subtyping doesn't have anything to do with exception handling
Man! You've sliced and diced me with cogent argument. I am humbled.
You have to set and identifier of the exception to throw in order to catch different exceptions. You can do this this at least two ways I can think of. 1) A single list of identifiers, ala IOCTLs or 2) a namespace of identifiers.
The first is really not scalable in a large distributed collection of libraries. People are going to start stepping on each others exception numbers.
The second could be implemented in a non-OO language. But amazingly enough the namespace oriented typing of OO languages is a perfect match for needs of catch clauses in exception handling. WOW! That is just like what I originally said (closely paraphrased) "sub-typing lends itself to Exception Handling".
You know, examples and arguement are pretty cool tools in communication. Sure, you can't write a complete thesis in a slashdot comment, but you can throw enough out there for people to get your drift.
Re:Object Oriented programming is overrated
on
Why not Ruby?
·
· Score: 1
Hello McFly!
I said "one element of OOP is the sub-typing. That actaully lends itself to another feature common to OO languages". Did I say Exception handling is a nessesary part of OO-language? Nope. Did I say that a element of OOP, sub-typing, is complementary to exception handling? Yes.
The US estimated 100,000 Iraqi casualties, mostly due to feul-air bombs dropped on the fix fortifications along the Saudi border and systematic destruction of the Iraqis fleeing Kuwait along the "Highway of Death".
A pro-arab journelist in Turkey thought the US was grossly underestimating the casualties. He did a fairly professional investigation. He was suprised by his own findings. The US grossly over-estimated the casualties. His conclusions were that the true number of casualties we around 20,000 Iraqi deaths. The reason was that the US pamphleted the the fixed fortifications and the big traffic jam on the highway out of Kuwait. Then they would bomb. Analysis showed that people took the pamphlets seriously and got the hell out of there.
Any use of a nuclear device on Iraqi command and control would have automatically set off Iraqi Chemical and Biological warhead SCUD missles. Those missles were targeted at Rhiad and Israel. There are terribly inaccurate, but with Chemical and Biological warheads they don't need to be accurate.
The US had tactical nuclear missles pointed at Sadam's head and the Iraqis had weapons of mass distruction pointed at civilians. Yet another game of MAD (mutually assured distructions. I actually think MAD is just a gamble that we've won sofar, but you only have to fail once.
The "Gulf War", aka "Sadam's Ass-Kickin'", would have gone out of Bush Sr. control, if the US had used Nuclear wepons, because of the default response by the Iraqis (C&C is not needed for this response).
So basically, I am saying you are full of shit. Nuclear wepons are weapons of mass distruction even small ones (that is why they are so usefull). Once you open war up to weapons of mass distruction, you open Pandora's Box.
FYI,/etc/init.d is the LSB specified place to put init scripts. The FHS doesn't have anything to say about init.d .
My personal argument would be that I might run a service that was install in/opt or/usr/local, but it would make no sense to install the script into/sbin for a server that is install in/usr/local/sbin .
Think about running a distro from CDROM that mounts/etc and/var as a RW ram disk and mounts/usr/local from NFS. You keep most of the dirs off of the root dir from needing to be writeable.
Re:Where is the bottleneck, really?
on
Benchmark Madness
·
· Score: 1
I agree that the level of FS knowlege by the Slashdot dweebs (aka people with User IDs over 100,000), but you are not completely correct.
You can Journel meta-data and data with Journeling file systems. Most just journel the meta-data. However, Ext3 developers found that due to their design the could journel meta-data and data. So that is a counter example to your assertion.
Second, B-trees are good in general but they have serious design conflicts with continously and simultaneously accessed file system type environments. To make this more concrete think about keeping b-trees balanced in the middle of constant use. Further consider the difficulties with shrinking b-trees while multiple processes are using a directory and keeping it consistant with the dcache. These are solvable problems but they extract a price in complexity and unexpected stalls for processes doing work.
One caveat to using Jouneling file systems on RAID5 disk systems, is that RAID5 is particularly slowed on writes, while Journeling file systems WRITE TO DISK TWICE; once to the journel and again to the file system meta-data.
I'd like to see these benchmarks on RAID5 systems.
Futher, I'd like to see these benchmarks on data sets bigger than memory. I think the reason they had dramatically different results on the first run versus subsequent runs, is because alot of inode data got cached in the dcache. This is also a probably why alot of the benchmarks were so similar.
Basically, these were poor benchmarks on unrealistic systems. Single disk systems are mostly used on workstations and home systems. Journeling is mostly advertised as an Enterprise level feature and Enterprise level machines use SMP and RAID5.
I think I gave up on the idea that Linux would ever know what it was doing with filesystems when I saw this abomination . That's right, Linux not only has no concept of vnodes, it actually uses a union of every single filesystem type for its inode data structure. Search lxr for "vnode" if you don't believe me. So much for generic, to say nothing of modular.
This has been discussed on the LKML. Using a Union for all the various FS specific inode info is a performance hack. You could have a pointer there to a FS specific inode, but you are going to lose memory locality benefits, and cause cache line thrashing. Inodes are touch OFTEN, that makes thes memory issues important.
Further this is a performance hack that takes advantage of the fact that the Linux Kernel is open source, and easily compiled. You can't get this advantage in closed source without arbitrary FS specific inode info size limits.
But, you are right there is an ugliness to it, but don't make the mistake that Linux Kernel developers are idiots or poor coders. This has been considered.
Remember the truism: The Perfect is the enemy of the Good.
Are you standing in support of the claim Perl is hard to learn?
So you've used sed, awk, bourne-shell, cut, sort, and grep alot...right? You know that bourne-shell uses backticks for sub shelled output? You know that '/' is called slash and '\' is called back-slash? (inverting this distinction is a big child-of-windows give away.) You've programmed in C...so you know argc, argv and envp right?
All these things translate one-for-one into PERL and you find it hard to understand PERL?
If you didn't understant all the above unix references, you aren't from a unix background. You are a cage-monkey the knows how to reboot unix computers. (Sorry for insulting all the cage-monkeys out there; love yah!)
Perl is easy to learn if you are coming from a Unix background. All the good stuff in built into the language (grep, sort, hashes, arrays). It has got the best from awk, sed, C, and shell.
For the Children of Windows (much like Children of the Corn), you are screwed. If you want an easy-to-use/familliar programming language try Logo. A more serious suggestion is Rexx from OS/2.
Has anybody noticed that all these Python/KDE/PHP/GUI loving and Unix whining posters come from User IDs over 200,000 ? What you people want in Windows 2000. Go get it from a warez site and call it Freeware.
The *BSD family's/usr/ports/ is a great idea and works great in practice. The source is brought in, dependancies as needed, and compiled for your box (See http://www.freebsd.org/ports/ for more info on that)
Few questions (I really don't know the answers):
Does the ports system allow for removal of a package and all it's files?
Can you query a package to find out if files have been modified?
How are the dependencies defined?
Re:Finally, something resembling clustering for Li
on
Mosix 1.0 Released
·
· Score: 3
Why do you feel compelled to post on a subject correcting other people's ignorance and yet being so profoundly ignorant yourself?
Repeat after me "THERE ARE MANY KINDS OF CLUSTERS". Again...again...again....
Now we will play a game: match the clustering technology description to a popular name. Match the letter to the number.
A. Message passing clusters used primarily for low bandwidth parralelel computation. B. Load balanced single protocol network clustering. C. Hardware takeover / hardware redudency for Hi-Availability clustering. D. Load balanced, homogeneous platform, with process migration clustering.
1. Veritas Cluster Server with Sun Multipath IO devices. 2. Arrowpoint-type web load balancer. 4. Mosix. 3. Beowulf.
For extra credit:
Is the above listing of clustering technologies comprehensive? [Y]es [N]o
Answers available from those with a clue after class.
How am I supposed to convince my management to move Oracle off a 30-gig VxFS and onto RedHat if I still have to deal with fsck?
You don't. Linux isn't always better that Solaris or HP/UX . And there is no filesystem on linux as cool as VxFS. Journelling != Logging . Journeling means that every write is repeated a second time. Logging only writes once. Gee what is the slowest operation on RAID-5 ummmmm...WRITING maybe.
I love Linux and think it will take over the world, but it ain't ready to run E10k with 1TB filesystems using Oracle in all it's tweaked out
glory.
Slashdot Mirror
Mosfet is a emotionally unstable GUI hacker. His knowlege of the long history and tradition of UNIX administration is pathetic. He ignores simple observables like PATH searches are more expensive than bin lookups. One executable dir per App would be FAR SLOWER than 2000 executables in a single dir.
... mindless infidels :).
This is another classic example of not letting programmers, especially GUI progrmmers, be involved in OS design.
For those of you who might be swayed by his foolish arguemnts, please read LHS, and the last decade of USENIX papers and LISA papers. Unix systems organization has been openly and vigorously debated for 15years. It has not be dictated by mere programmers from high on above like MS. And RedHat is to be applauded for properly implementing the FHS which is a standard, others like SUSE should be encouraged to become compliant (/sbin/init.d
Read a book, get a clue, whatever. "cryptography as we know it is obsolete" puhlease.
Nobody has touched RSA, or RSA key exchange. The problem is the crytographic protocols and implementation. The protocols often use strong (re: compute intesive) cryptography to exchange weak (re: fast) cryptographic keys. It's the weak shit that is getting cracked. Also Protocols and implementations let information to leak about the weak keys. This alows cracks to exploit the leaked info, to find the private keys faster (re: polynomial time) than the standard approaches.
Someday these corporate dumbasses, who want to limit fundementally unlimited resources, will get a clue and start creating secure implementations. That is when you should be afraid.
Maybe they'll start developing their protocols and implementations in an open process. This would have a much better chance of flushing out the bugs . The only thing they need to really control is the master keys. A smart plan would be to develope the code in a open manner, then and only then start encrypting all their music, video, articles, with the super-secret "Master" keys.
We are lucky that they are following the security thru obscurity developement process. How many times will they fail before they start to wake up?
Sony is working on encryption to the monitor. Others are working on encryption to the speakers (usb speakers with DSPs in them). Combine that with a cluefull open developement process, and we are SCREWED in a squeal-for-me-boy kinda way.
Then we start hoping someone with the real keys reveal them (the whistle blower way). Or the big corperate dumbasses leave the keys on a not-quite-so-secure system and hackers release them.
The cryptographic algorithms aren't being attack it is all the dressing PROTOCOLS, IMPLEMENTATIONS, and PEOPLE.
ext2 doesn't have a 2GB file size limit. That was a operating system limit which went away somewhere in the middle 2.2.x stable series.
Further, ext3 is not the-next-version-of-EXT. It is an extention of ext2 which is fully compatible with ext2. Think of ext2 as two things: the format of bits on the disk, and the code to read/write those bits. Ext3 keeps the same format (actually with compatable extentions), but mostly it changes the code for reading/writing to the disk (journelling).
The ext2 filesystem is tried and true. You can go back and forth between ext2 and ext3 with no reformating or issueing of commands other than the mount command.
ReiserFS is a more "sophisticated" filesystem than ext[23], and XFS is a more "sophisticated" filesystem than ReiserFS. But I keep "sophisticated" in quotes because the utility, reliability, and speed of a FS relies more on your usage patterns, than on the genius of the filesystem designers/coders.
FFS-style: ext2,ufsFFS+journel: ext3, ufs+
B+tree directories, B+tree block layout, Journelling: ReiserFS
B+tree directories, B+tree block layout, extents, Journelling : XFS, JFS
Loggin FS: VxFS (my favorite)
I use ext3 at home. Good speed, no need to tar up all my files..reformat drives..untar all my data, journelling, mainline kernel support, tried and true.
One place I would seriously consider ReiserFS is for home directories. The place it really shines is constantly reading and writing lots of "small" files (small ~50k). For Gnome and KDE config files, Mozilla disk caches, CVS checkouts, and untaring of source, ReiserFS is going to be a leader of the benchmarking pack. You'll notice the difference.
But don't get into holy wars over FS, and don't think that Linux is whole generations behind Commercial Unixen. Linux Kernel is dramaticallly ahead in some areas and minorly behind in others. The only place it is dramatically behind is places where the computer you are running the OS on cost more than a half million dollars.
I just had to throw something into my post to keep the "my-favorite versus your-favorite" fight going :).
I follow java because it is my favorite bondage and discipline language. (BTW I use "B&D language" as the inverse pejorative to "scripting language").
Ignoring the idea that the number of lines means anything (you or I could have streatched out the perl code). But if you read that Java 1.4 example of word counting with regex and char strings, you should have notice how complicated Sun has made this kind of thing. I was stunned. CharacterSet("ISO-8995-1") WTF!?! Creating a immutibale object that represents the number one and calling it ONE. Something went wrong in their language design.
I realize Java is growinig past it's initial quicky-design. I just wish all the OO-lusers would stop critisizing Perl for the same issues. Ruby is the one language that hasn't gotten crufty, yet. And that is cuz it's young.
Some of this Perl 6 stuff scares me too. Mainly because I think perl can be abused to write bad code. I am thinking stuff that is REALLY obtuse. I've seen code with $|++. Which is stupid. Because if $| == 1, then the code doesn't do anything and the inverse $|-- fails to achieve your purpose when $| == 2. STDOUT->autoflush(1) is the clear way to write it.
Just because dumb-ass "programmers" CAN write obscure code in perl, doesn't invalidate the value of Perl. Any language with expressive power is vulnerable to having "Obfusicate-X" contestants write programs in that language. A wise quote: "Fortran programmers can write Fortran in any language".
Perl 6 is looking to be the exact opposite of LISP. In my view, LISP has little or no syntax; just Lots of Incessant Silly Parenthesis. Well it looks like perl 6 is going to be nothing else but syntax.
This might be valid perl6:
I like perl by this might be to much for me.Of course the real reason I use perl is two fold; it's expressive power (unlike bondage and discipline languages) and CPAN (the killer feature).When I look at other languages like python or ruby, I look for their CPAN equivalent. Right now their is none, but maybe soon.
BTW, for the JAVA fans out there the following url is the same code as:
48 lines (took out comments and empty lines) versus 3.
BTW, This is as obfusicated as my code gets. I did it mostly for brevity.
You need to understand one thing from reading the Kernel Mailing List.
The problem is that when _ALL_ the RAM is consumed and _ALL_ the swap is consumed ON 2GB+ RAM machines, THEN the computer goes into conniptions.
That is a very specific recipe. Linux should not have this problem; it should just degrade performance gracefully; but it does have this problem. There is a fundemental problem here. They kernel guys may be able to tune the VM to avoid the problem. But depending on the tuneing a malicious program could still cause the catelepsy.
My suggested work-arounds for the whinners with more than a GB of RAM:
- DON'T EXAUST ALL YOUR MEMORY AND ALL YOUR SWAP. If you start hitting swap consistantly, stop!
- Make sure users run with ulimits. Make them real big, but less than everything.
Disclaimer: I've never had a clearance, and this information isn't really a secret.
:).
The Hubble space telescope was built using the super-structure of a line of spy satalites. A fella-who-knows told me that when they first discovered the optics problem, many people's first suspicion was that they installed a mirror from the spy satellites. This wasn't the case.
BTW, the difference btw the spy satellites and the Hubble is that the Hubble is focused for inifinity and the spy satellites are focused for a few hundred miles or so (and towards the earth, which would burn the Hubbles sensors).
Another thing, (prolly just a myth) is that the Hubble is the 13th of this line of spy satellite; Fun for the numeralogically inclined
Last thing, the US regularly announces satellite launches that don't exist, and skips numbers for the satellites. "SS1...SS2...SS3...SS5..Hey! how'ed they slip SS4 passed us?". They didn't their just fucking with you.
Basically, No.
Try not to think of it as a Shield. What Bush is proposing is not even remotely comprehensive defense from large numbers of missles. It is meant as a point defense against less than 5 launches.
Futher, the missle defense systems are most likely to target the boost phase of the missle launch. We're supposed to be sitting near by the bad-missile-launching-country at see in military vessles carrying boost-phase anti-missle wepons.
The kind of defenses envisioned in Bush's VERY LIMITED anti-missle system would have no applicability to anti-meteor defense. Basically, unless we get a few years advanced warning about metor impact, we're screwed big time.
This is the kind of all-regulation-is-bad thinking that I wanted to distance myself from. It isn't the existance of regulations that is bad. It is the kind of regulation AND how much responsibility people think the FEDERAL government should take. I pointed to the Clean Air Act as an example of good regulations. resource/environmental management is a federal responsibility. Regulations which take advantage of markets are a good kind of regulation; as opposed to obviating markets by creating one-size-fits-all autocratic mandates.
See big companies wanted no limits on their ability to pollute. Pollution has real costs on our economy (health care, loss of economically usable forests/land, etc) and quality of life. But the companies that create the pollution are not the companies paying for the pollution (insurance companies, forresters, and farmers pay the costs of pollution). Federal government regulation is the ONLY way to push the costs back to the companies who are otherwise being subsidized by the ability to pollute.
About John McCain's legislative initiatives:
His proposed regulation of campaign finances are not going to be "useless regulations". They may not be a complete solution to the disparity in VOLUME of speech and influence by wealthy parties (big companies not little ones), but they are a good try.
His advocacy for health care reform is most certainly pro-free-market. HMOs are specifically excempt from law suits. Non-HMO health insurance companies are NOT excempt. This is not fair to the traditional insurance companies.
This disparity is not pro-free-market. Further, law suits are a nessesary part of contract law. Contract law is abso-fucking-lutely nessesary to Free Markets. This doesn't means that we don't need Tort reform. I believe we do. Companies can be blown away by a single lawsuit on an issue where they were trying to bahave in good faith (example: Dow-Corning and Silicone brest implants). This isn't healthy for a free market. On the other had I do think companies should be killed by law suits when they bahave in grevious bad faith. For instance selling a product they know will and is killing people while constantly lieing and commiting fraud to deny what they have already proven to themselves; Tabbaco and certain car design issues come to mind (not the Firestone tires; I'm thinking the Gremlin issue). Lawsuits allow the share holders to suffer, because they are the source of authority in companies.
Don't be such a simple minded partisan.
BTW, I am registered Democrat so I can improve the breed, because I think they are better stock to work from. However, in my home state, the Free Socialist Republic of Maryland, I am looking forward the opportunity to vote for a Republican governor that doesn't snarl and drool. And I vote against the stealth senator Sarbanes even if the Republican candidate were to snarl and drool.
Actually, I am a registered Democrat. I think there is a sufficiantly large contingent of Democrats that aggree with pro-free market policies, that the Democrats are not to far gone. Unfortunately, the "Blue Dog Democrates" are being decimated by the southern states' shift away from conservative Democrats in favor of Christian-Right Republicans who are whole heartedly in favor of government involvement social and religious affairs.
Republicans really aren't very good on these issues either. Bush's tax plan is basically lame as an economic policy. There was no effort to get capital gains tax reductions. The Republican attitude towards regulatory reform is basically pro-big-business. They fought long and hard against the Clean Air Act, (Yes Bush Sr. signed it, but as a comprimise) which is a model of Free Market based regulation. The Clean Air Act established a market for "carbon credits" and let market forces deal with how, when, and where the Carbon emmissions were reduced; in less then a decade the last half century of carbon pollution was rolled back. At the same time the US went thru amazing economic growth. Regulation, the environment, and economic growth are NOT incompatable. It is a matter of "How".
Further Republicans have shifted away from Free Market economic policies. The Barry Goldwater, John McCain (Goldwater's hand picked successor for his Senate seat), Gramm Rudman, etcetra branch of the Republican party are getting pushed out of the GOP in favor of the Social Conservatives (who have a very harsh all or nothnig attitude towards members of their own party).
Dashelle is not a freaky liberal, however he does seem to be to far to the "traditional" Democratic side of the Party than I care for. He is one sly bastard that is going to kick the GOP's ass; and I am not sure if that is good or bad.
Here! Here!
I figure no one will actually read this cuz I am posting it late, but here is my RANT...
Most of the feeble-minded "pro-business" crowd fail to recongnize this fact: 2/3 of the U.S. economy is small to medium sized businesses. There are government definitions of what this is, but basically they are less than 1000 employees and under a $100million in revenue. Consistantly the growth in the US economy comes from these buisnesses. They employ the majority of americans.
I'd argue (with ample agreement of economists) that this is THE single distinguishing feature of the US economy versus the European and Japanese economies. Japan is really the best case example of large corperations stiffling their economy. It isn't so much that large corperations are dumb and slow by nature, it is just that the aggragate behavior of many players in the market creates the pressure for efficiency gain and inovation thru a variety of mechanisms.
This all means that "monopoly-abuse-isn't-bad" and "big-companies-aren't-a-threat" ideas are wrong. These ideas are just ignorant and inaccurate.
My nit pick of the day is that the term "Capitalism" refers to the specific use of publicly traded stock and securities to fund investment in corperation building. The term that should be used is "Free Markets". Free Markets require reliable and liberal contract and regulatory law and the ability of most/all individuals in the society to participate in the market at any level.
Pro-big-buisness laws actually hurt Free Markets, because they typically create barriers to entry for smaller players.
This in a nutshell is why we have Anti-Trust laws. This is why I support campaign finance reforms (they favor big businesses over small-medium businesses; look at the "sponsers" of the Democratic and Republican Conventions).
If you want a stong economy, support small and medium buisnesses. To this end support:
BTW, I am a self professed liberal, but I strongly believe in democratic free markets. Free Markets are the best way to help the most people; Free Markets require strong (but minimal) democratic regulatory environments to flourish (despite Anne Rand spouting Libertarian fairy-tales)
Sorry for the misunderstanding. My list was not meant as a cronological order. It was three different methods, some based on specific symtoms.
You are absolutely right. You start with talking and listening. And you do it for everyone, not just the trouble makers.
Icchy! I feel like Oprah Winfrey. Lets all hold hands and sing Kum-by-yah.
My real point is "Take no shit, give no shit. And mind-fuck them if you have to.". Ahhh! now I feel better.
I have dipped into bad behavior before. Now I manage techies.
First of all, those who won't comment code and document design should be beaten severly about the head with a LART. That is bad coding/design practice and is completely unacceptable. Put them on undocumented code. When they bitch have them explain why it is bad and how the company should fix it. Listen to the response and implement. And you might notice their attitude changing. If it doesn't replace them.
Second, give them responsibility. I was once a camp counseler. I took the jokers/ring-leaders and put them in charge; small things at first then bigger things. They dig the power and are usually the most effective leaders. Again, if it doesn't work replace them.
Last, and most imporant. Have a frank conversation with them and respect their opinions. There is nothing as powerful as a little respect. Small reasonable acts of control and general respect are the best way to get people with the program. If it doesn't work, what's left? Fire them.
You are wrong. Subtyping doesn't have anything to do with exception handling
Man! You've sliced and diced me with cogent argument. I am humbled.
You have to set and identifier of the exception to throw in order to catch different exceptions. You can do this this at least two ways I can think of. 1) A single list of identifiers, ala IOCTLs or 2) a namespace of identifiers.
The first is really not scalable in a large distributed collection of libraries. People are going to start stepping on each others exception numbers.
The second could be implemented in a non-OO language. But amazingly enough the namespace oriented typing of OO languages is a perfect match for needs of catch clauses in exception handling. WOW! That is just like what I originally said (closely paraphrased) "sub-typing lends itself to Exception Handling".
You know, examples and arguement are pretty cool tools in communication. Sure, you can't write a complete thesis in a slashdot comment, but you can throw enough out there for people to get your drift.
Hello McFly!
I said "one element of OOP is the sub-typing. That actaully lends itself to another feature common to OO languages". Did I say Exception handling is a nessesary part of OO-language? Nope. Did I say that a element of OOP, sub-typing, is complementary to exception handling? Yes.
Your comment was orthogonal to my point.
The US estimated 100,000 Iraqi casualties, mostly due to feul-air bombs dropped on the fix fortifications along the Saudi border and systematic destruction of the Iraqis fleeing Kuwait along the "Highway of Death".
A pro-arab journelist in Turkey thought the US was grossly underestimating the casualties. He did a fairly professional investigation. He was suprised by his own findings. The US grossly over-estimated the casualties. His conclusions were that the true number of casualties we around 20,000 Iraqi deaths. The reason was that the US pamphleted the the fixed fortifications and the big traffic jam on the highway out of Kuwait. Then they would bomb. Analysis showed that people took the pamphlets seriously and got the hell out of there.
Any use of a nuclear device on Iraqi command and control would have automatically set off Iraqi Chemical and Biological warhead SCUD missles. Those missles were targeted at Rhiad and Israel. There are terribly inaccurate, but with Chemical and Biological warheads they don't need to be accurate.
The US had tactical nuclear missles pointed at Sadam's head and the Iraqis had weapons of mass distruction pointed at civilians. Yet another game of MAD (mutually assured distructions. I actually think MAD is just a gamble that we've won sofar, but you only have to fail once.
The "Gulf War", aka "Sadam's Ass-Kickin'", would have gone out of Bush Sr. control, if the US had used Nuclear wepons, because of the default response by the Iraqis (C&C is not needed for this response).
So basically, I am saying you are full of shit. Nuclear wepons are weapons of mass distruction even small ones (that is why they are so usefull). Once you open war up to weapons of mass distruction, you open Pandora's Box.
FYI, /etc/init.d is the LSB specified place to put init scripts. The FHS doesn't have anything to say about init.d .
/opt or /usr/local, but it would make no sense to install the script into /sbin for a server that is install in /usr/local/sbin .
/etc and /var as a RW ram disk and mounts /usr/local from NFS. You keep most of the dirs off of the root dir from needing to be writeable.
My personal argument would be that I might run a service that was install in
Think about running a distro from CDROM that mounts
I agree that the level of FS knowlege by the Slashdot dweebs (aka people with User IDs over 100,000), but you are not completely correct.
You can Journel meta-data and data with Journeling file systems. Most just journel the meta-data. However, Ext3 developers found that due to their design the could journel meta-data and data. So that is a counter example to your assertion.
Second, B-trees are good in general but they have serious design conflicts with continously and simultaneously accessed file system type environments. To make this more concrete think about keeping b-trees balanced in the middle of constant use. Further consider the difficulties with shrinking b-trees while multiple processes are using a directory and keeping it consistant with the dcache. These are solvable problems but they extract a price in complexity and unexpected stalls for processes doing work.
One caveat to using Jouneling file systems on RAID5 disk systems, is that RAID5 is particularly slowed on writes, while Journeling file systems WRITE TO DISK TWICE; once to the journel and again to the file system meta-data.
I'd like to see these benchmarks on RAID5 systems.
Futher, I'd like to see these benchmarks on data sets bigger than memory. I think the reason they had dramatically different results on the first run versus subsequent runs, is because alot of inode data got cached in the dcache. This is also a probably why alot of the benchmarks were so similar.
Basically, these were poor benchmarks on unrealistic systems. Single disk systems are mostly used on workstations and home systems. Journeling is mostly advertised as an Enterprise level feature and Enterprise level machines use SMP and RAID5.
This has been discussed on the LKML. Using a Union for all the various FS specific inode info is a performance hack. You could have a pointer there to a FS specific inode, but you are going to lose memory locality benefits, and cause cache line thrashing. Inodes are touch OFTEN, that makes thes memory issues important.
Further this is a performance hack that takes advantage of the fact that the Linux Kernel is open source, and easily compiled. You can't get this advantage in closed source without arbitrary FS specific inode info size limits.
But, you are right there is an ugliness to it, but don't make the mistake that Linux Kernel developers are idiots or poor coders. This has been considered.
Remember the truism: The Perfect is the enemy of the Good.
You come from a unix background. That is nice.
Are you standing in support of the claim Perl is hard to learn?
So you've used sed, awk, bourne-shell, cut, sort, and grep alot...right? You know that bourne-shell uses backticks for sub shelled output? You know that '/' is called slash and '\' is called back-slash? (inverting this distinction is a big child-of-windows give away.) You've programmed in C...so you know argc, argv and envp right?
All these things translate one-for-one into PERL and you find it hard to understand PERL?
If you didn't understant all the above unix references, you aren't from a unix background. You are a cage-monkey the knows how to reboot unix computers. (Sorry for insulting all the cage-monkeys out there; love yah!)
For the Children of Windows (much like Children of the Corn), you are screwed. If you want an easy-to-use/familliar programming language try Logo. A more serious suggestion is Rexx from OS/2.
Has anybody noticed that all these Python/KDE/PHP/GUI loving and Unix whining posters come from User IDs over 200,000 ? What you people want in Windows 2000. Go get it from a warez site and call it Freeware.
Few questions (I really don't know the answers):
Repeat after me "THERE ARE MANY KINDS OF CLUSTERS". Again...again...again....
Now we will play a game: match the clustering technology description to a popular name. Match the letter to the number.
A. Message passing clusters used primarily for low bandwidth parralelel computation.
B. Load balanced single protocol network clustering.
C. Hardware takeover / hardware redudency for Hi-Availability clustering.
D. Load balanced, homogeneous platform, with process migration clustering.
1. Veritas Cluster Server with Sun Multipath IO devices.
2. Arrowpoint-type web load balancer.
4. Mosix.
3. Beowulf.
For extra credit:
Is the above listing of clustering technologies comprehensive? [Y]es [N]o
Answers available from those with a clue after class.
You don't. Linux isn't always better that Solaris or HP/UX . And there is no filesystem on linux as cool as VxFS. Journelling != Logging . Journeling means that every write is repeated a second time. Logging only writes once. Gee what is the slowest operation on RAID-5 ummmmm...WRITING maybe.
I love Linux and think it will take over the world, but it ain't ready to run E10k with 1TB filesystems using Oracle in all it's tweaked out glory.