... was "d'uh.. people published that stuff, what did they expect?". I still think Google should cache as much as they want, especially if they follow robots.txt. (Not following robots.txt is a bit rude, but if you're not implementing acess controls for the general public, why should Google be any different?)
Then I remembered one of my fraternity Brothers. At MIT, Freshmen (things change drastically in the Fall of 2002) pretty much decide durring their first week in Boston where they're going to live for the next 4 years, this includes pledging fraternities. To make things less chaotic, each MIT fraternity sends an information packet out to each incomming freshman male and print out lots more to have on hand durring the week of rush. The information packet needs to be finished by the end of the term. One Brother (let's call him Joe) was too busy at the end of the term to put much thought into the personal bio blurb required from each Brother. He thought he'd force the editors to completely rewrite hsi bio from scratch by making it too awful to print. He listed his interests as "Chinese eating, Chinese sleeping, midget tossing, anorexic women with low self esteem, and bovine necrophillia". The editors called his bluff and put his bio, unedited, in the Rush mailer. The rush mailler got transferred into electronic form. Luckily, I jut checked Google for his bio and got no hits. His name only shows 30 hits, half of which are him. It's not really bad, but might cause some flags to go up with sme potential employers/potential inlaws, particualrly since all of the other bios were completely serious and normal. Some stuff you write as a joke may someday end up in big glossy pages and online where it seems in context, but is totally out of context.
Some day you may wish something about you was never online. Oh well, you can't do anything about it.
Ehh... any idea how long most Mac or MS software is in development before it's made public? The main difference you see is that 1) most projects at Sourceforge are public even in the pre-pre-pre-alpha stage and 2) the realease-soon, release-often model looks like continuous change when compared to the snapshots you get every two years from MS. In reality, very few software projects, commercial or otherwise, ever declair themselves to be the pinacle of software completeness. Also, very few OSS projects ever officially cancel themselves, so you may see a lot of dormant projects that would have been canceled pre-beta at MS.
I love python. Java could definately learn some things from python.
However, Java is faster than at least TCL, PHP, and Python. So your "slow" comment is
definately wrong if you use those languages as a base.
As far as your claim about switch, if you're doing a switch statement on a string, you really don't want a switch statement, you want if-elses. A Java switch is implemented as
an opcode followd by a variable-length table of ints and branch ofsets. Adding in string comparisions would make it a completely different beast that should at least go by a dffernt name so that ignorant programmers aren't fooled into thinking they are getting a
really fast construct (switch) when they're getting a slow construct (nested string comparisons). Java has done a pretty good job of keeping the core constructs orthogonal.
If you want a switch(String) construct, go use a third-party preprocessor.
Java certainly has its stupid features (ints don't have the minimum subset of fetures to add them to Vectors, so they get changed to Integers). Templates also heve their advantages, but you don't want to get into STL hell where one error doesn't get caught until it's seven layers deep in templates. I think Java should also have a built-in preprocessor.
Read more carefully. I made no claims about what Mono intended to do. I made a claim about what they would do. I should have added the word "fully".
I doubt there will be a time where all progras compiled for the latest version of MS.NET will run properly on the latest version of Mono.
Perhapse tons of libraries was an overstaement. I'm just saying that the history of trying to implement an OSS JVM is probably going to repeat itself.
I wish Miguel et al. the best luck in the world. I would love to see Mono be a better.NET than MS.NET. I'm just saying that Microsoft's claim that.NET is multiplatform is false. The Free BSD shared source verion is not a full version. I fear that running VS.NET compiled stuff in Mono is always going to be like viewing IE-specific webpages under
Netscape. In any case, godspeed all you coders, godspeed.
Traffic lights and pacemakes getting cracked shouldn't be a problem if they're designed with hardware enforcemet of representaional invariants.
A pacemaker should never have a pulse rate outside of 40 bpm to 120 bpm. Sure, it'd be inconvenient if your pulse went down to 40 bpm, but you wouldn't die... maybe you'd pass out. Sure, some athletes have resting heart rates of 30, but if you need a pacemaker,
a pulse of 40 to 120 is completely functional. If the CPU tells the hardware to beat outside that range,
the hardware should put out a pulse rate of 72 bpm - the average resting heart rate for an adult male, slightly slow for the average female but it's easier on the heart.
Stoplights should have a finite state machine in hardware. One of the inputs sould be a hardware timer that goes low after a state transition and goes high 3 seconds later. The CPU can control some of the inputs to the FSM, but there are no unsafe states and no unsafe transitions. (i.e. "red, green, red, green" can only go to "red, yellow, red, yellow", "r,g,r,y", or "r,y,r,g"). This way, the CPU has no "words in its languge" to describe a transition to the state "green, green, green, green" or from "red, green, red, green" to "green, red, green, red" without going through 3 seconds of "red, yellow, red, yellow" (i.e. if the timer input is low, all of the state transitions for that state return to itself). This way, it's imposible for the light to be in an unsafe state or make an unsafe state transition.
You have the hardware check representational invariants (40 <= heart_rate <= 120) and go into a safe fallback state if the rep. invar. is violated (heart_rate = 72). Otherwise,
if the states are simple enough, you have a finite state machine containing only safe states and safe transitions. (If the FSM is too complicated, it's easy to screw it up, so
you should have a rep. invar. check to back you up.) If you use one of these techniques, a terrorist can at best inconvinience you, even if s/he replaces ALL of the web-connected CPU's software. A small FPGA or CPLD to do this enforcement costs less than a couple of bucks and the programming is pretty streight forward for simple invarients like those used in stoplights and pacemakers.
There may be reasons to give net connectivity to stoplights. (I can definately imagine giving them 802.11 with IPSEC so that ambulances can change the lights ahead of them witout having to have the 911 dispatch center do it for them.) As long as you have proper hardware enforcement, these things aren't a problem. If the terrorist has the time and acess to pull out the FPGA and re-burn it with some unsafe states, s/he might as well clip the wires to the lightbulbs and cross-wire the lights. On a similar note, if a terrorist has the ability to take the FPGA out of the pacemaker and reprogram it and put it back in the person without killing them. (Maybe for blackmail purposes.) Why not implant a remote
drug O.D. injector or a remote triggered half kilo of semtex in the abdomen?
It turns out that the DEC Alpha folks did exactly what I was talking about, and then some.
It's called F32!X. It does JIT compilation the first time it sees a foreign binary and saves the results, optimizing the saved binary on eachsucessive execution of the program. All of the system calls get translated. You can run Linux x86 binaries on Linux Alpha and WinNT x86 binaries on WinNT Alpha, but you can't do cross-os migration. This was exactly what I had in mind, plus adding the saving of the results. The DEC people realized it's what they needed in order to get a foothold.
I don't think Bochs can or will do it, but Plex86 (old site) might if it ever gets finished.
I'm fully aware of what Plex86 does. I've been wrestling with the CVS version of plex86 for the past few months. The config files in CVS issue commands that no loner exist, and
fail to load modules that are required to pass internal tests. Thier developer's list was also hitby a plague of Turkish SPAM a few months back, causing lots of people to unsubscribe.
I was talking about running Linux x86 apps on PPC Linux. The plex86
approach cannot work cross-platform. Basically, it scans a section of
code for operations it doesn't like, alters those it doesn't like, then lets
it all run. This is faster than emulation, but won't work for x86 apps on PPC.
I'm too lazy to bring up the./ article, but there were some benchmarks less than a year ago showing that for most applications (graphical I/O being the notable exception),
the latest IBM JIT JVM outperforms C++ using the MS VC++ compiler with the default optimizations.
I'll agree that sometimes the JVM takes forever to load, but the latest IBM JIT JVM
continuously profiles your code and then does the equivalent of compiling the most
commonly run parts with all of the optimizations turned on. I would guess that C++
does better relative to Java on non-x86 platforms, at least if you're using one of the older JITs. This is becuase the register-starved x86
looks pretty much like a stack-based machine in comparison to say the IBM POWER, HP PaRISC, or Sun SPARC CPU families. However, the latest IBM JIT probably does wonders
to naorrow the gap on machines with 16 or more general purpose registers.
Re:Don't scream
on
.NET for Apache
·
· Score: 1, Informative
...and ystem-neutral [sic] platforms like.NET
.NET may very well be yestem-neutral, but it's only going to run
on Windows. Mono is only going to implement the VM, you most likely
are still going to need tons of libraries.
It's all about the libraries. I think the massive ammounts of libraries
are the main reason Kaffe only supports Java 1.1. Do we have any idication
that the.NET library set will be smaller than the Java 1.2 library set?
Do you think the.NET upgrade cycle is going to go throug versions any slower
than Java? (Remember, selling new library/VM versions makes MS more money.)
In short, the CLR will be cross-system..NET probably will not work well at all
cross platform unless the program writers strictly discipline themselves to use a
small subset of the.NET libraries.
I think it's great that there's another Itanium compeditor (although the target markets are different). However, I'd much rather see x86 compatability through software. Last I heard, the Itanium runs x86 code about as fast as a 600 MHz PII. This shouldn't be that
hard to beat in software on a 2 GHz chip, particularly if they use a JIT compiler.
Are there any ports of bochs that pass system calls through to the native system so that none of the actual OS is running inside Bochs? This would allow you to, say, run x86 Linux code on Linux PPC or Win x86 apps on Win ia64. This assumes, of course, that the system call numbers and arguments are the same across architectures. Maybe it would require too much OS-awareness in Bochs in order to fix the endianess, but it would be nice to move away from hardware x86 decoders.
Please someone tell me that all of the 64-bit mode instructions are the same length. (Maybe the caryover instructions from x86 need to be padded with nops.) Varaible-width instructions absoutely kill hardware or software decoding speed, especially if you're trying to parallelize it. Maybe we can all migrate to pure x86-64 instructions and slowly rid ourselves of the old x86 instructions?
Ideally, AMD would come out with a RISC cpu with an open source x86 emulator for the OS vendors to integrate with thier OS. I would love to be able to have comodity RISC or VLIW chips on pricewatch. x86 decoding is a waste of heat and chip realestate.
Re:No... a 64bit chip doesn't have to be 'slower'
on
AMD's 64-Bit Chip
·
· Score: 2
I think the parent's parent was half right. It's not that 32-bit compatability is a huge drag, it's that x86 compatability is a huge waste of transistors if you don't
need it, or are willing to get the x86 compatability from emulation.
One of the big problems with the x86 instruction set is the multiple
width instructions that need to be partially decoded before their length is known,
this makes the decoding step partially inherently serial.. very hard on high performance
superscalar architectures... Without the P4's translation cache or the AthlonXP's
three decoders (two of which must always wait for the one to partially decode their op
before they know where to start), the ALUs would get instruction starved pretty quickly.
I'd be darn happy if I could get an x86-64 chip with the x86 decoder replaced by a decoder for instructions that are much closer to the chip's internal instruction set. The only software on my machine that I didn't apt-get is the Sun JVM, so if the chip were even mildly popular, there's a good chance all of the stuff I use would be ported. Without the x86 royalties being paid to Intel (parts of the instruction set are patented) and without the extra chip realestate (==chip cost) and work required (== heat and power consmption) for x86 compatability, I'd be much happier. Clock rates might not go higher, but the chips might be able to dipatch more instructions per cycle while using less power and fewer transistors.
Itanium and "majority of its speed" in the same sentance! Hehe... Why the heck didn't they
give the Itanium a full FPU instead of causing some of the floating point instructions to raise a floating point software assist exception for the kernel to handle? (This is a big performance kill for FP, and also the reason why at least some FP operations are not
permitted while running in kernel mode in the Linux ia64 port.) Register windowing was also a bad idea. It doesn't surprise me that the thing runs so hot and slow and uses so
much realestate. They added some whiz-bang features (register windowing, anyone seen it used effectively in SPARC?) instead of some useful ones (like a hardware-only FPU).
Up until 4.0, NT was multi-platform. Hopefully M$ had enough foresight to keep the code portable just in case they ever wanted to be multi-architecture again. As much as I hate the MS monopoly, I dislike the "everything is x86" mentality more. Linux ia64 sets the CPU up to be big endian, right? And Win ia64 sets it up as elttil naidne?
<OT> BTW, other than being able to use a single integer load opcode and still being able to increase register size, what are the advantages of elttil niadne architectures? (Obviously, programmer familiarity is also a factor, but the switch isn't that bad.)
Network byte order is defined as big endian, so switching everything back and forth stinks.
(I know, X11's solution is beautiful, but that's beside the point.) I noticed that
both Itanium and XScale CPUs have the endianess as part of the CPU state. MIPS machines
come in both endianesses. So, why do we come out with new architectures that have the
little-endian option? (Obviously x86 compatability reuires it in this case.)
MIPS comes in mips and mipsel flavors. Can you get a PPCel chip? PaRISC and SPARC are big-endian only, and I think the IBM POWER family are as well. Giving the chip the
option of ether endianess adds complexity and transistors, although not many.
Either way, most file-sharing is theft - plain and simple.
First of all, I definately go through periods where I whipe all of the
songs I don't own from my computer and I go through periods where I say
"hmm... deleting the songs from my computer for 6 months didn't cause me to buy them...
so my having them doesn't hurt sales." I'm on the fence about the morality of copying music.
In any case, we need to stop using industry-supplied propaganda terms like
"theft" and "piracy". These kids aren't taking other kid's lunch money or
forcably taking control of seagoing vessels. They are infringing on
copyright, just as if they'd snuck a camera into an art museum. Using
horrably innacurate terms doesn't help Joe Public or Jim Congressman think
more clearly about the issues. (Once, I'd like to see someone use a
Britney song turned up full blast, or a bured MS iso, to forcably take
controll of a seagoing craft in international waters. That's
music/software piracy.)
As far as solutions, the music industry needs to shift from being
property based to being service based. Tours, custom content,
subcriptions to a website with the latest pre-mastered songs and
outakes, latest news on their favorite band, etc. (Yes, and T.V. ads.)
I admit, I don't know enough to say exactly how it should work, butI do know that
the current state fo thing is decidedly bad for the economy. It's
highly artificial. Imagine if the cottage weavers guilds had
succesfully gotten the government to limit power looms to the same output
speed as human weaver. This is the situation we have today.
The new, more efficient method of content distribution is being
legally stifled becuase a lot of people have a lot invested in the
old system.
I'd also like to point out that telling people they're comitting theft
or piracy by downloading music kills your credibility. It's like
telling kids that pot is as bad for them as crack. People may not be able to
articulate their feelings, but most people know deep down that theft isn't the
right way to thik about it. Copyright infringement of this sort is illegal,
and perhapse morall wrong, but let's call it what it is.
(People lose respect for the law when the law doesn't make sense. I think we'd have fewer pot smokers if tobacco were illegal. I don't smoke pot or tobacco, but I think if we had to pick one of them to make illegal,
we picked the wrong one. Do the worst pot addicts smoke more than the equivalent of a
pack per day? Nicotene is pretty nasty stuff. I looked into chemically extracting it
and creating the sulphate for use as a skin contact fast acting mouse poison. Look it
up in the MERK index. Nicotene's primary natural function is killing insects that eat
tobacco leaves. It's a pretty efficient poison. Many long-term smokers have heart problems from having elivated
blood pressure and heart rates for most of the day. In many ways, copyright laws are just as screwed up. If the laws were more consistant, I
think they'd be respected and obeyed more.)
If the bands can't accept direct payment for songs, then don't pay for songs. Just make it a donation for the well being of the band. Problem solved. The RIAA is unreasonable and powerful, but I doubt that they would/could use thier power to prevent
no-strings-attached gifts. Then again, the US has been heading south for a few years.
2. Decryption takes no longer with 128 bit wep than it does with 64bit wep. This depends on the actual RC4 implementation...
In fact, the way RC4 works, the key and IV are concatenated and repeated over and over to create a 256-byte(2048-bit) key. The way I usually code up RC4 (I've done it a few times
before I memorized RC5... now to memorize AES.) is that I don't expand the key, just
loop over it, so a longer key means fewe conditional branches and actually faster rekeying.
Encryption speed should be identical. The key only sets the initial state (the permutation of 0..255), which evolves durring encryption/decryption. Unless the programmer is completely retarded, the actual encryption and decryption speeds for RC4 should be copletely independant of key size. (Yes, a keystream lookup table for short keys is cmpletely retarded.)
If you want to return all of your JPEGs to them via fax machine,
base64 encode them first. Make sure to use a really large font to
help their OCR. Put a rendering of the image afterwards to help with
error correction.
Anyone with a phone in the same area code as their fax machine want to set
up a web gateway?
Hmm... I was looking for a good python practice program...
Declair bullet-proof shields for your T.V. to be in the public domain before someone patents them!
Now, more than ever, people will want to be able to shoot thier T.V. Hmm...
too badthe PS2 doesn't have an RF input, or you could write an app to
allow you to "shoot" the image on the T.V.
Oh well. I guess TiVo will come out with a "letterbox" option to
get rid of the add and re-center the image. And they'll find themselves back
in court.
The number of detected attacks is rising sharply for linux and slightly falling for Windows. What percentage of Windows houses install an IDS solution? What about Linux houses? What percentage of Linux break-ins get reported to someone? and Windows break-ins? There very well may be a point this year when attacks against Linux outnumber attacks against Windows, but I think it's more likey that the vast vast majority of attacks againt Windows machines go unnoticed.
Also, nimda and code red scans are attacks. If those got counted, allong with every virus email, the story would be very different.
If you were given the IP address or a vulnerable WinXP box, a vulnerable Linux box, and a vulnerable OpenBSD box and your life depended on owning one of the boxes without getting detected, which one would you chose given no other information? Only the suicidal would pick OpenBSD... the probability of there being another OpenBSD dedicated IDS box nearby is pretty high.
Let's not forget that a Linux shop can do a minimal install on a retired PII (or maybe even a 486) server and use it as a dedicated IDS box... no MS liscence fee. MS isstill goingto charge you for every running x86 box, regardless of OS, if you have an MS site liscence, so no negligible-cost dedicated IDS boxes for Windows shops.
I'm biased. I sure am... but it's mostly due to experience... I was a residet computer consultant for my fraternity for 3 years. Sure we had the one guy that talked another guy into trying out Mandrake and didn't bother to tell him to keep it up to date, but for the vast majority of the Brothers, the Linux guys could hold thier own. Several of the Windows guys were accidently running "Are you sure? What is IIS? Why is that bad?". And then there were the windows alerts popping up once per minute on all of the Win32 boxes in the house because one guy decided to test his UPS. These are very smart guys, but they
gave me a vey bad impression of Windows users. I doubt the general populace
can do better than my fraternity.
(Yes, the house GPA was in the 75th percentile fr fraternities and the average fraternity GPA is above the on-campus GPA at MIT. Even the management and bio majors could kick your ass in diferential equations, so no "stupid drunk frat boys" comments. They get tiring... very very tiring... especially comming from people that can't integrate thier way out of a paper bag.)
In summery, let's not forget that Linux and Windows often get deployed in very different environments.
Well, if I were to design the system, I'd design the default behavior to ask about formatting whenever you inserted blank removable media. This would format it as UFS or FAT12, not sure which I'd chose as the default. If I wanted to cange the default behavior
for CDR drives to not format, I'd insert a special case to not even ask.
If I wer a betting man, I'd put money on Solaris at least attempting to put a filesystem on the CDR if you told it to format the blank media. If they changed the default behavior,
it wouldn't make sense to put the special case after the dialog popped up. They probably didn't see a need to change the default behavior, or they changed teh default behavior for
a writable CD drive to use that standard CD-RW filesystem (the name escapes me at the moment... UDFS???).
The poster does seem a bit paranoid... but lightning strikes are much longer in duration and a lot of the EM power generated is low enough frequency that the plane's skin acts as a faraday cage. The Russians did some research on this stuff and created some crazy stuff. Supposedly for a small fraction of a second, some of their devices woudl generate more power than all the rest on man-made electrical energy on the planet at that time. Of course, the Russian devices were quite involved. They used an explosively pinched inductor (stored energy remains constant as inductance rapidly approaches zero -> current rapidly approaches infinate) to charge a specially shapped copper chamber and then vaporized this huge capacitor in a controlled manner (using TNT). Basically, my understanding is that this second stage was a Marx generator. (Copper vapor filling the spark gaps. Not sure what they used for the large resistors.)
In any case, you one can generate much more power than a lightning strike without using as much energy. Also, if designed properly, some portion of the mechanical energy from the chemical emplosion gets converted into additional electrical output. A short durration and short distance arc also ends up generating much more of its EM radiation in a frequency range more destructive to aircraft.
If one were to have the TNT to make an EMP bomb, one could think of much simpler ways to take out jetliners, not all of which would require bypassing airport security. Supposedly Al Qaeda looks at action movies and the internet for ideas, so let's no enumerate the possibilities here. Let's just take it for granted that if someone has 5 or 25 lbs. of TNT, EMP bombs are not the biggest threat the FBI should be looking at.
This is somewhat offtopic, but it's stillin the military topic. (My second paragraph is more on-topic.) I've read the book Fulcrum by a Soviet pilot that stole a MIG-29 from the then USSR and landed in Turkey. Interesting book. Very intersting. I recomend you all read it. The author wanted to steal a MIG-29 and flyit to the west without killing anyone. He walked through the planning stages of figuring out how to drug all of his friends to keep them from being called up to shoot him down when Moscow figured out what was going on. He got shot by the soldier guarding the plane he stole, but not too badly because he had a clear enoug head to remember that an AK-47 pulls up and left while firing full-auto. You read the book and you can tell the author is probablya boarderline genius. More relavently, the author explains and dispells a lot of the Western misconceptions about the Soviet military.
The author makes a good point that the U.S. expected the same tactics out of Iraq that it saw from countries that paid for bargain-basement military training from the Soviet Union. The author was appearently an invaluable consultant in the Gulf War in correcting misconceptions the U.S. had on Soviet tactics and helped them guess what the Iraqis got for thier training dollar. Do you thnk the Soviets allowed their best training and tactics to be exported? Also, there were two main branches of the Soviet Air Force. One group that did long-range intercepts and one that appearently trained for air-to-air combat and nuclear strikes. The one branch was mustly full of bureacrats' kids looking for cushy jobs that let them drink themselves to sleep every night of the week. They had a reputationfor being sloppy, unprofessional flying missle platforms: take off.. let your ground controller vector you in, launch your missle, then land. It was these people that shot down the Korean airliner to cover up the fact they lied to Moscow about having fixed storm damage to an early-warning radar installation. These people wer also much more easily recruited by the CIA, so most of the U.S. intelligence was about this branch. It turns out that the other branch of the Soviet Air Force was comparable to the U.S.A.F> in terms of discipline and tactics.
Also, the Su-27 uses technologies not seen on us fighters at that time. Super light aluminum-Lithium alloys are expensive, yet the Soviets apopearently spared no expense on teh Su-27. It had a helmet-mounted targting system to backup the normal nose-mounted system. If a pilot could see a target, he could get missle lock, and the thrust vectoring on the missles pretty much allowed the pilot to fire as soon as he got lock. Not having to
put the endemy plane in front of you is a huge advantage in a dog fight.
I always thought that spyware on a windoze box had the power to do whatever it felt like.
Unlss you've set up seperate users' accounts, the default user has Administrator privledges. Any installer s/he runs can do anything it wishes to the system given a little ingenuity.
Single user systems are evil. I'm the only one that ever uses the machine I'm currently running, yet it has 3 accounts... root, serious work, and a leasure account. This way, I protect myself from myself.
Macromedia Fireworks (I can'tremember which version) can't run in an unpriveledged account. I set up my GF's machine properly and then started pounding my head on the table because Macromedia was forcing her to do everything I just told her not to do. It makes me wonder what thier QA people do all day. Appearently they never tested it on a multi-user system. Oh well.
Slashdot Mirror
The episode with the monster called "skuzlebut" and the hunting? "No, it's not poaching if you first yell 'Look out, it's comming right for us.'"
Then I remembered one of my fraternity Brothers. At MIT, Freshmen (things change drastically in the Fall of 2002) pretty much decide durring their first week in Boston where they're going to live for the next 4 years, this includes pledging fraternities. To make things less chaotic, each MIT fraternity sends an information packet out to each incomming freshman male and print out lots more to have on hand durring the week of rush. The information packet needs to be finished by the end of the term. One Brother (let's call him Joe) was too busy at the end of the term to put much thought into the personal bio blurb required from each Brother. He thought he'd force the editors to completely rewrite hsi bio from scratch by making it too awful to print. He listed his interests as "Chinese eating, Chinese sleeping, midget tossing, anorexic women with low self esteem, and bovine necrophillia". The editors called his bluff and put his bio, unedited, in the Rush mailer. The rush mailler got transferred into electronic form. Luckily, I jut checked Google for his bio and got no hits. His name only shows 30 hits, half of which are him. It's not really bad, but might cause some flags to go up with sme potential employers/potential inlaws, particualrly since all of the other bios were completely serious and normal. Some stuff you write as a joke may someday end up in big glossy pages and online where it seems in context, but is totally out of context.
Some day you may wish something about you was never online. Oh well, you can't do anything about it.
It doesn't work for me in Konq, so I log in as 10101/10101. Does it work for anyone else, or did the NYT catch on?
Ehh... any idea how long most Mac or MS software is in development before it's made public? The main difference you see is that 1) most projects at Sourceforge are public even in the pre-pre-pre-alpha stage and 2) the realease-soon, release-often model looks like continuous change when compared to the snapshots you get every two years from MS. In reality, very few software projects, commercial or otherwise, ever declair themselves to be the pinacle of software completeness. Also, very few OSS projects ever officially cancel themselves, so you may see a lot of dormant projects that would have been canceled pre-beta at MS.
As far as your claim about switch, if you're doing a switch statement on a string, you really don't want a switch statement, you want if-elses. A Java switch is implemented as an opcode followd by a variable-length table of ints and branch ofsets. Adding in string comparisions would make it a completely different beast that should at least go by a dffernt name so that ignorant programmers aren't fooled into thinking they are getting a really fast construct (switch) when they're getting a slow construct (nested string comparisons). Java has done a pretty good job of keeping the core constructs orthogonal. If you want a switch(String) construct, go use a third-party preprocessor.
Java certainly has its stupid features (ints don't have the minimum subset of fetures to add them to Vectors, so they get changed to Integers). Templates also heve their advantages, but you don't want to get into STL hell where one error doesn't get caught until it's seven layers deep in templates. I think Java should also have a built-in preprocessor.
I doubt there will be a time where all progras compiled for the latest version of MS.NET will run properly on the latest version of Mono.
Perhapse tons of libraries was an overstaement. I'm just saying that the history of trying to implement an OSS JVM is probably going to repeat itself.
I wish Miguel et al. the best luck in the world. I would love to see Mono be a better .NET than MS.NET. I'm just saying that Microsoft's claim that .NET is multiplatform is false. The Free BSD shared source verion is not a full version. I fear that running VS.NET compiled stuff in Mono is always going to be like viewing IE-specific webpages under
Netscape. In any case, godspeed all you coders, godspeed.
A pacemaker should never have a pulse rate outside of 40 bpm to 120 bpm. Sure, it'd be inconvenient if your pulse went down to 40 bpm, but you wouldn't die... maybe you'd pass out. Sure, some athletes have resting heart rates of 30, but if you need a pacemaker, a pulse of 40 to 120 is completely functional. If the CPU tells the hardware to beat outside that range, the hardware should put out a pulse rate of 72 bpm - the average resting heart rate for an adult male, slightly slow for the average female but it's easier on the heart.
Stoplights should have a finite state machine in hardware. One of the inputs sould be a hardware timer that goes low after a state transition and goes high 3 seconds later. The CPU can control some of the inputs to the FSM, but there are no unsafe states and no unsafe transitions. (i.e. "red, green, red, green" can only go to "red, yellow, red, yellow", "r,g,r,y", or "r,y,r,g"). This way, the CPU has no "words in its languge" to describe a transition to the state "green, green, green, green" or from "red, green, red, green" to "green, red, green, red" without going through 3 seconds of "red, yellow, red, yellow" (i.e. if the timer input is low, all of the state transitions for that state return to itself). This way, it's imposible for the light to be in an unsafe state or make an unsafe state transition.
You have the hardware check representational invariants (40 <= heart_rate <= 120) and go into a safe fallback state if the rep. invar. is violated (heart_rate = 72). Otherwise, if the states are simple enough, you have a finite state machine containing only safe states and safe transitions. (If the FSM is too complicated, it's easy to screw it up, so you should have a rep. invar. check to back you up.) If you use one of these techniques, a terrorist can at best inconvinience you, even if s/he replaces ALL of the web-connected CPU's software. A small FPGA or CPLD to do this enforcement costs less than a couple of bucks and the programming is pretty streight forward for simple invarients like those used in stoplights and pacemakers.
There may be reasons to give net connectivity to stoplights. (I can definately imagine giving them 802.11 with IPSEC so that ambulances can change the lights ahead of them witout having to have the 911 dispatch center do it for them.) As long as you have proper hardware enforcement, these things aren't a problem. If the terrorist has the time and acess to pull out the FPGA and re-burn it with some unsafe states, s/he might as well clip the wires to the lightbulbs and cross-wire the lights. On a similar note, if a terrorist has the ability to take the FPGA out of the pacemaker and reprogram it and put it back in the person without killing them. (Maybe for blackmail purposes.) Why not implant a remote drug O.D. injector or a remote triggered half kilo of semtex in the abdomen?
I don't think Bochs can or will do it, but Plex86 (old site) might if it ever gets finished.
I'm fully aware of what Plex86 does. I've been wrestling with the CVS version of plex86 for the past few months. The config files in CVS issue commands that no loner exist, and fail to load modules that are required to pass internal tests. Thier developer's list was also hitby a plague of Turkish SPAM a few months back, causing lots of people to unsubscribe.
I was talking about running Linux x86 apps on PPC Linux. The plex86 approach cannot work cross-platform. Basically, it scans a section of code for operations it doesn't like, alters those it doesn't like, then lets it all run. This is faster than emulation, but won't work for x86 apps on PPC.
I'm too lazy to bring up the ./ article, but there were some benchmarks less than a year ago showing that for most applications (graphical I/O being the notable exception),
the latest IBM JIT JVM outperforms C++ using the MS VC++ compiler with the default optimizations.
I'll agree that sometimes the JVM takes forever to load, but the latest IBM JIT JVM continuously profiles your code and then does the equivalent of compiling the most commonly run parts with all of the optimizations turned on. I would guess that C++ does better relative to Java on non-x86 platforms, at least if you're using one of the older JITs. This is becuase the register-starved x86 looks pretty much like a stack-based machine in comparison to say the IBM POWER, HP PaRISC, or Sun SPARC CPU families. However, the latest IBM JIT probably does wonders to naorrow the gap on machines with 16 or more general purpose registers.
It's all about the libraries. I think the massive ammounts of libraries are the main reason Kaffe only supports Java 1.1. Do we have any idication that the .NET library set will be smaller than the Java 1.2 library set?
Do you think the .NET upgrade cycle is going to go throug versions any slower
than Java? (Remember, selling new library/VM versions makes MS more money.)
In short, the CLR will be cross-system. .NET probably will not work well at all
cross platform unless the program writers strictly discipline themselves to use a
small subset of the .NET libraries.
Are there any ports of bochs that pass system calls through to the native system so that none of the actual OS is running inside Bochs? This would allow you to, say, run x86 Linux code on Linux PPC or Win x86 apps on Win ia64. This assumes, of course, that the system call numbers and arguments are the same across architectures. Maybe it would require too much OS-awareness in Bochs in order to fix the endianess, but it would be nice to move away from hardware x86 decoders.
Please someone tell me that all of the 64-bit mode instructions are the same length. (Maybe the caryover instructions from x86 need to be padded with nops.) Varaible-width instructions absoutely kill hardware or software decoding speed, especially if you're trying to parallelize it. Maybe we can all migrate to pure x86-64 instructions and slowly rid ourselves of the old x86 instructions?
Ideally, AMD would come out with a RISC cpu with an open source x86 emulator for the OS vendors to integrate with thier OS. I would love to be able to have comodity RISC or VLIW chips on pricewatch. x86 decoding is a waste of heat and chip realestate.
One of the big problems with the x86 instruction set is the multiple width instructions that need to be partially decoded before their length is known, this makes the decoding step partially inherently serial.. very hard on high performance superscalar architectures... Without the P4's translation cache or the AthlonXP's three decoders (two of which must always wait for the one to partially decode their op before they know where to start), the ALUs would get instruction starved pretty quickly.
I'd be darn happy if I could get an x86-64 chip with the x86 decoder replaced by a decoder for instructions that are much closer to the chip's internal instruction set. The only software on my machine that I didn't apt-get is the Sun JVM, so if the chip were even mildly popular, there's a good chance all of the stuff I use would be ported. Without the x86 royalties being paid to Intel (parts of the instruction set are patented) and without the extra chip realestate (==chip cost) and work required (== heat and power consmption) for x86 compatability, I'd be much happier. Clock rates might not go higher, but the chips might be able to dipatch more instructions per cycle while using less power and fewer transistors.
Itanium and "majority of its speed" in the same sentance! Hehe... Why the heck didn't they give the Itanium a full FPU instead of causing some of the floating point instructions to raise a floating point software assist exception for the kernel to handle? (This is a big performance kill for FP, and also the reason why at least some FP operations are not permitted while running in kernel mode in the Linux ia64 port.) Register windowing was also a bad idea. It doesn't surprise me that the thing runs so hot and slow and uses so much realestate. They added some whiz-bang features (register windowing, anyone seen it used effectively in SPARC?) instead of some useful ones (like a hardware-only FPU).
<OT>
BTW, other than being able to use a single integer load opcode and still being able to increase register size, what are the advantages of elttil niadne architectures? (Obviously, programmer familiarity is also a factor, but the switch isn't that bad.) Network byte order is defined as big endian, so switching everything back and forth stinks. (I know, X11's solution is beautiful, but that's beside the point.) I noticed that both Itanium and XScale CPUs have the endianess as part of the CPU state. MIPS machines come in both endianesses. So, why do we come out with new architectures that have the little-endian option? (Obviously x86 compatability reuires it in this case.) MIPS comes in mips and mipsel flavors. Can you get a PPCel chip? PaRISC and SPARC are big-endian only, and I think the IBM POWER family are as well. Giving the chip the option of ether endianess adds complexity and transistors, although not many.
First of all, I definately go through periods where I whipe all of the songs I don't own from my computer and I go through periods where I say "hmm... deleting the songs from my computer for 6 months didn't cause me to buy them... so my having them doesn't hurt sales." I'm on the fence about the morality of copying music.
In any case, we need to stop using industry-supplied propaganda terms like "theft" and "piracy". These kids aren't taking other kid's lunch money or forcably taking control of seagoing vessels. They are infringing on copyright, just as if they'd snuck a camera into an art museum. Using horrably innacurate terms doesn't help Joe Public or Jim Congressman think more clearly about the issues. (Once, I'd like to see someone use a Britney song turned up full blast, or a bured MS iso, to forcably take controll of a seagoing craft in international waters. That's music/software piracy.)
As far as solutions, the music industry needs to shift from being property based to being service based. Tours, custom content, subcriptions to a website with the latest pre-mastered songs and outakes, latest news on their favorite band, etc. (Yes, and T.V. ads.)
I admit, I don't know enough to say exactly how it should work, butI do know that the current state fo thing is decidedly bad for the economy. It's highly artificial. Imagine if the cottage weavers guilds had succesfully gotten the government to limit power looms to the same output speed as human weaver. This is the situation we have today. The new, more efficient method of content distribution is being legally stifled becuase a lot of people have a lot invested in the old system.
I'd also like to point out that telling people they're comitting theft or piracy by downloading music kills your credibility. It's like telling kids that pot is as bad for them as crack. People may not be able to articulate their feelings, but most people know deep down that theft isn't the right way to thik about it. Copyright infringement of this sort is illegal, and perhapse morall wrong, but let's call it what it is.
(People lose respect for the law when the law doesn't make sense. I think we'd have fewer pot smokers if tobacco were illegal. I don't smoke pot or tobacco, but I think if we had to pick one of them to make illegal, we picked the wrong one. Do the worst pot addicts smoke more than the equivalent of a pack per day? Nicotene is pretty nasty stuff. I looked into chemically extracting it and creating the sulphate for use as a skin contact fast acting mouse poison. Look it up in the MERK index. Nicotene's primary natural function is killing insects that eat tobacco leaves. It's a pretty efficient poison. Many long-term smokers have heart problems from having elivated blood pressure and heart rates for most of the day. In many ways, copyright laws are just as screwed up. If the laws were more consistant, I think they'd be respected and obeyed more.)
If the bands can't accept direct payment for songs, then don't pay for songs. Just make it a donation for the well being of the band. Problem solved. The RIAA is unreasonable and powerful, but I doubt that they would/could use thier power to prevent no-strings-attached gifts. Then again, the US has been heading south for a few years.
In fact, the way RC4 works, the key and IV are concatenated and repeated over and over to create a 256-byte(2048-bit) key. The way I usually code up RC4 (I've done it a few times before I memorized RC5... now to memorize AES.) is that I don't expand the key, just loop over it, so a longer key means fewe conditional branches and actually faster rekeying. Encryption speed should be identical. The key only sets the initial state (the permutation of 0..255), which evolves durring encryption/decryption. Unless the programmer is completely retarded, the actual encryption and decryption speeds for RC4 should be copletely independant of key size. (Yes, a keystream lookup table for short keys is cmpletely retarded.)
Anyone with a phone in the same area code as their fax machine want to set up a web gateway?
Hmm... I was looking for a good python practice program...
Does anyone know of any documented cases of servers being exploited through specially formatted emails? (besides buffer overflows)
Now, more than ever, people will want to be able to shoot thier T.V. Hmm... too badthe PS2 doesn't have an RF input, or you could write an app to allow you to "shoot" the image on the T.V.
Oh well. I guess TiVo will come out with a "letterbox" option to get rid of the add and re-center the image. And they'll find themselves back in court.
Also, nimda and code red scans are attacks. If those got counted, allong with every virus email, the story would be very different.
If you were given the IP address or a vulnerable WinXP box, a vulnerable Linux box, and a vulnerable OpenBSD box and your life depended on owning one of the boxes without getting detected, which one would you chose given no other information? Only the suicidal would pick OpenBSD... the probability of there being another OpenBSD dedicated IDS box nearby is pretty high.
Let's not forget that a Linux shop can do a minimal install on a retired PII (or maybe even a 486) server and use it as a dedicated IDS box... no MS liscence fee. MS isstill goingto charge you for every running x86 box, regardless of OS, if you have an MS site liscence, so no negligible-cost dedicated IDS boxes for Windows shops.
I'm biased. I sure am... but it's mostly due to experience... I was a residet computer consultant for my fraternity for 3 years. Sure we had the one guy that talked another guy into trying out Mandrake and didn't bother to tell him to keep it up to date, but for the vast majority of the Brothers, the Linux guys could hold thier own. Several of the Windows guys were accidently running "Are you sure? What is IIS? Why is that bad?". And then there were the windows alerts popping up once per minute on all of the Win32 boxes in the house because one guy decided to test his UPS. These are very smart guys, but they gave me a vey bad impression of Windows users. I doubt the general populace can do better than my fraternity.
(Yes, the house GPA was in the 75th percentile fr fraternities and the average fraternity GPA is above the on-campus GPA at MIT. Even the management and bio majors could kick your ass in diferential equations, so no "stupid drunk frat boys" comments. They get tiring... very very tiring... especially comming from people that can't integrate thier way out of a paper bag.)
In summery, let's not forget that Linux and Windows often get deployed in very different environments.
If I wer a betting man, I'd put money on Solaris at least attempting to put a filesystem on the CDR if you told it to format the blank media. If they changed the default behavior, it wouldn't make sense to put the special case after the dialog popped up. They probably didn't see a need to change the default behavior, or they changed teh default behavior for a writable CD drive to use that standard CD-RW filesystem (the name escapes me at the moment ... UDFS???).
I meant more power than all other man made electrical power at the time. I didn't mean to imply that I was comparing Watts to Joules.
In any case, you one can generate much more power than a lightning strike without using as much energy. Also, if designed properly, some portion of the mechanical energy from the chemical emplosion gets converted into additional electrical output. A short durration and short distance arc also ends up generating much more of its EM radiation in a frequency range more destructive to aircraft.
If one were to have the TNT to make an EMP bomb, one could think of much simpler ways to take out jetliners, not all of which would require bypassing airport security. Supposedly Al Qaeda looks at action movies and the internet for ideas, so let's no enumerate the possibilities here. Let's just take it for granted that if someone has 5 or 25 lbs. of TNT, EMP bombs are not the biggest threat the FBI should be looking at.
The author makes a good point that the U.S. expected the same tactics out of Iraq that it saw from countries that paid for bargain-basement military training from the Soviet Union. The author was appearently an invaluable consultant in the Gulf War in correcting misconceptions the U.S. had on Soviet tactics and helped them guess what the Iraqis got for thier training dollar. Do you thnk the Soviets allowed their best training and tactics to be exported? Also, there were two main branches of the Soviet Air Force. One group that did long-range intercepts and one that appearently trained for air-to-air combat and nuclear strikes. The one branch was mustly full of bureacrats' kids looking for cushy jobs that let them drink themselves to sleep every night of the week. They had a reputationfor being sloppy, unprofessional flying missle platforms: take off.. let your ground controller vector you in, launch your missle, then land. It was these people that shot down the Korean airliner to cover up the fact they lied to Moscow about having fixed storm damage to an early-warning radar installation. These people wer also much more easily recruited by the CIA, so most of the U.S. intelligence was about this branch. It turns out that the other branch of the Soviet Air Force was comparable to the U.S.A.F> in terms of discipline and tactics.
Also, the Su-27 uses technologies not seen on us fighters at that time. Super light aluminum-Lithium alloys are expensive, yet the Soviets apopearently spared no expense on teh Su-27. It had a helmet-mounted targting system to backup the normal nose-mounted system. If a pilot could see a target, he could get missle lock, and the thrust vectoring on the missles pretty much allowed the pilot to fire as soon as he got lock. Not having to put the endemy plane in front of you is a huge advantage in a dog fight.
Unlss you've set up seperate users' accounts, the default user has Administrator privledges. Any installer s/he runs can do anything it wishes to the system given a little ingenuity.
Single user systems are evil. I'm the only one that ever uses the machine I'm currently running, yet it has 3 accounts... root, serious work, and a leasure account. This way, I protect myself from myself.
Macromedia Fireworks (I can'tremember which version) can't run in an unpriveledged account. I set up my GF's machine properly and then started pounding my head on the table because Macromedia was forcing her to do everything I just told her not to do. It makes me wonder what thier QA people do all day. Appearently they never tested it on a multi-user system. Oh well.