Sometimes I think we watch too much Star Trek (or, more likely, Star Trek, specifically, ST:NG reflects something that I don't like about our culture). Anyway, have you ever noticed that a ridiculously large proportion of ST:NG episodes feature some conflict that is resolved by Mr. LaForge saying something like "maybe, if we reverse the polarity on the field generator..." or Dr. Crusher saying "maybe, if we alter the microbes' DNA..."
My point is (and this is perhaps not so directly applicable to the article, but is reminiscent, anyway): Why is it that we so often look for technology to provide a quick fix for what is really a very complex and difficult social problem? Obviously, it's the easy way out, but does it really work? Think about all the various technical schemes for locking down copyrighted content that we have been discussing lately.
If we can come up with some technical way to reduce the CO2 output of smokestacks, then great. However, I still worry that unless we (and particularly we Americans) wake up and take up the difficult task of addressing all the causes of our massive CO2 output, this new method of scrubbing CO2 is not going to be enough.
Anyway, go cyanobacteria, because every little bit helps.
I know a lot of people here are very anti-regulation, but I think it would be great if case law established that web robots must obey the Robots Exclusion Standard. Since it's a widely-known standard, I think it can be fairly argued that robots that choose to disregard/robots.txt are in danger of tresspassing to chattels. Using the standard also would allow bots to fulfill their helpful role, while providing a clear distinction between what is and what is not acceptable.
Sure, one might argue that people might be unaware of the standard, but that is seldom an excuse. I may be unaware of fire/electrical codes, but I'll still get in trouble if I don't adhere to them, because I'm putting others at risk and thereby imposing a cost upon society (fire trucks and insurance don't come free). Web crawlers that index data in violation of the Robots Exclusion Standard impose a cost on companies and society just as well, in the end requiring people to by bigger pipes, faster servers, and so on (thereby using more power, dumping more old computer components into landfills and more chipmaking chemicals into the environment).
My point is that web crawler operators live in a society, just like everyone else, and they too must be held accountable for the consequences of their actions, particularly when they willfully disregard the requests of web site operators as expressed in/robots.txt
Is it just me, or does the article have this backwards? Sure, I'm just nit-picking, but it seems that one would install such metal fibers to promote conduction and thus prevent the build-up of static charge. Consider conductive anti-stat floor mats, wrist straps, et cetera, which make this seem not-so unusual (except that it's built into the carpet).
Oh well, anyway, it's still all pretty cool, and I envy the folks who get to work there (mainly for the nice cabling setup).:)
for training people up to be able to use a computer for running charities or getting a job you'd be far better off teaching them how to use Windows/Word/Excel than some open source solution which is more difficult and less well supported
For running charities, perhaps, but for getting a job? Take a look at the SANS 2000 Salary Survey Summary:
"The average for UNIX folks was $70,080, while the average for Windows folks was $61,233."
Granted, this only applies to SysAdmins (so I probably would agree with you in general), but given the currently high demand for "UNIX folks" (and despite the fact that Linux is not technically UNIX), it seems that training people in Linux is not such a bad idea. Speaking more anecdotally: I was able to get my current job as a UNIX admin because I got so sick of working in retail that I decided to train myself with Linux (plus it was interesting, fun, and free). I know of several other people who have done similar things, and I even have friends involved in hiring that say that Microsoft certifications actually make a resume look worse (one said that he threw away a resume as soon as he saw the letters "MCSE").
I should also add that Linux is an excellent platform on which to learn programming, which opens up yet another career.
Augh...wait a minute; what am I saying? Train no one in Linux! The fewer people there are who know how to use it, the more in-demand those of who do know will be. (Now you know the real reason why people report that Linux is hard to use;)
One thing that I feel that Linux is missing sorely (as opposed to the commercial Unices) is good process accounting.
For one thing, it's really shameful that lines like these two from acct.c:
ac.ac_io = encode_comp_t(0/* current->io_usage */);
ac.ac_rw = encode_comp_t(ac.ac_io / 1024); have been hanging around for so long (and in linux/kernel/, at that!). As I recall, a patch for this was submitted for 2.3.x, but I guess it went nowhere.
Process accounting may not seem like a big deal to the typical desktop user, but for sysadmins (especially on big systems), it can be really important.
Maybe they figure that not telling users is a form of damage control. Sure, maybe some people will get upset if this sort of thing happens and you don't inform those who may have been affected. Then again, if you send out a message to all your customers telling them that some hacker has their credit card number, you know your phone is going to be ringing off the hook. I certainly wouldn't be happy to sign my name to that sort of letter. Maybe they feel that they would lose more customers by admitting fault than by keeping mum.
This is not to say that I think keeping silent is right (taking responsibility for your mistakes is the Right Thing To Do), but it is certainly understandable.
Very well, you're right. You happen to have caught me in a grumpy mood today, and I admit I was probably looking for a fight. I certainly thought that you were trying to offer one, but I see that I was mistaken.
Wow. Amazing. Well, since you took the time to be sarcastic, snide, and insulting, why don't I return the favor?
Whether or not Jupiter has a solid core is still being debated. Most notable astrophysicists agree that there is no way of knowing if Jupiter has a core mass or not. The last time I heard anyone argue FOR the solid core was 1989. Of course it's impossible to know for certain if Jupiter has a rocky core without direct measurement, and some people may not believe until they reach out and touch it with their hands, but that's not happening anytime soon, and is irrelevant. As for the '1989' date, you seem to be implying that no one believes in the rocky core anymore. Perhaps, then, you should try the following experiment. Go to Google. Enter "jupiter rocky core". Press "I'm feeling lucky." See JPL mention rocky core. Or, just press "Google Search" and read links that deal with the concept more directly. This, of course, does not prove that the core exists, but rather goes to refute the idea that you present above, which is that scientists no longer believe that it exists.
Maybe that's just me. You're right, though, it is believed to have a rocky core. But it is also believed to not have one. You tell me to disregard Galileo's data but you don't have a problem quoting implied data from Pioneer 10 and 11. If you meant that YOU believe it has a rocky core, I would like some more justification. You're right. While you (I assume you wrote the post to which I replied, AC) can come out and state that 'there is no "solid" mass on Jupiter' and give no justification whatsoever, mere mortals such as I must give it. And I didn't say that one should disregard Galileo's data, merely that not finding the expected abundance of water along the probe's entry path (which infrared images revealed to be a relatively unclouded area) does not imply a total absence of water on Jupiter. I never argued that there was as much water as we originally thought, merely that there is some water.
What "rocky core" do you propose can sustain it's solid phase at the awesome pressure and temperature at the core of Jupiter, after being immersed in liquid metallic hydrogen? Sure, the temperature is an issue, but the "awesome pressure" would encourage most materials to go into solid phase (except for weird ones like water, of course). And what the heck does the liquid metallic hydrogen sitting above it have to do with anything? Are you proposing that it would dissolve the core or do you just think the words "liquid metallic hydrogen" are going to convince me that you know what you're talking about?
Impact events would make no difference one way or another, unless you are implying that Jupiter collided with another planet. That could give it a core mass. Oh, sure, that's scientific. So, an asteroid couldn't do it, but a planet could? What about a really big asteroid, or a really small planet? My point is: what magic line determines that a body is big enough for you to dub it a "planet" and simltaneously allow your imagination to grant it the power to give Jupiter a rocky core, where all other impacts "would make no difference"? Give me a break.
Where did this theoretical planet come from? Extrasolar? Oh, sure... the only bodies around while the solar system was forming are the ones we see today, is that what you are saying? Your argument keeps getting better and better.
Most of the extrasolar planets we have ever discovered are gas giants themselves, and larger than Jupiter to boot. Right, this is irrelevant, but I suppose you believe that those numbers are in no way skewed by the detection methods currently in use? Obviously if we look for planets by looking at the gravitational wobble of stars, we're going to detect massive ones that orbit close to their stars. Heck, that much is even in the article to which this story referred, if you were unable to figure it out for yourself using basic physics. If there were any previous indication that you were scientifically qualified, this last paragraph is squashing that.
Which would mean Jupiter used to BE an extrasolar planet, and collided with a smaller planet with the orbit Jupiter now holds. If that is your theory, I advise you to examine the feasibility of it. Wow. Where are you getting these ideas and whatever it is that you are smoking? That's quite a ridiculous theory that you put in my mouth, and I will shortly spit it back at you.
Well that was fun. Let me take some time out at the end here, leaving all the problems with what you just said behind, to explain the problems with your initial argument. The bottom line is that you have made statements that cannot be reasonably defended.
You say that "there is no water on Jupiter" [emphasis mine]. Now, I could talk about solar oxygen levels and the presumed abundance of oxygen in the early solar system and so on, but I don't have to. I'm even willing to believe that Jupiter has a relatively low oxygen abundance in its atmosphere. Nevertheless, we all saw a bunch of chunks of frozen water slam into Jupiter a few years back. Discarding for the sake of argument the idea that there was any water present on Jupiter before the impacts, if one single molecule of water from the impacts survived, then there is water and it's on Jupiter, so you lose; end of argument.
Similarly, you say "there is no 'solid' mass on Jupiter". Again, if there is a single speck of dust on Jupiter, you lose; end of argument.
The smart thing to do would have been to admit that you were exaggerating (that's why I gave you that "out" at the top of my post), and then we could have settled into a more reasonable discussion of planetary formation and elemental abundances. Instead you went off on these weird tangents, became insulting, and capped it off with ridiculous statements like the one about extrasolar planets mostly being giants, which thoroughly discredits you, made it a joy to write this reply, and makes it certain that neither I nor anyone else will pay any mind to anything more you have to say.
Alright that's it. It's time to start working on a poor man's massive IDE RAID solution. I just got back from talking with someone else about this, and now I'm psyched.
There is no water on Jupiter. Ummm, that's just plain wrong. Sure, our probe found less water than expected, but it dropped into a hot spot. Besides, do you think Shoemaker-Levy contained no water at all?
If there is any oxygen at all there, it's in very trace amounts Well sure, everything that's not hydrogen or helium on Jupiter you might call "trace," since that's not really a quantitative term. Still, Jupiter is big enough that there is plenty of H2O and CO. I would advise against treating the Galileo probe's measurements as authoritative for Jupiter's entire atmosphere. There very well may be concetrations of oxygen in certain portions of that vast ocean of gas.
There is no "solid" mass on Jupiter, unless you consider pressurized hydrogen a solid mass. Also incorrect. Jupiter is believed to have a rocky core. Do you really think that in the enitre history of Jupiter, nothing rocky ever hit it and sank through all that gas?
"But if you were sane, you'd just buy backup tapes instead. Or for that matter, use RAID-5."
RAID by itself is no substitute for backups. Sure, it protects against the failure of a disk, but what about other sorts of disasters, like bad RAID controller firmware, fire, flood, filesystem corruption, hostile intrusion, and so on?
Of course, there are still the biggest use of backups: user carelessness.
Of course, with the falling price of IDE drives, I have been thinking that it might be possible to build some IDE-RAID-based storage system that would rival tape libraries in cost/terabyte, but be much faster. (Maybe this is what you meant.) You could still make tapes for offsites and archives and such. Unfortunately, all the commercial products out there that do something like IDE RAID -> SCSI (which one might need to keep things simple) are way more expensive than their components would seem to justify, and I have no money, so it remains a pipe dream.
I suppose you could just get tons of pentium systems doing software RAID and NFS.... hmmmm.... this sounds like a job for those folks at NASA who brought us Beowulf (and all the "imagine a Beowulf cluster of [ ]" on Slashdot.
Imagine trying to do revision with a speech recognition package. It's completely unsuited to the draft-revision-draft-revision-ad infinitum process used for serious writing. Limited usefullness at best. A good secretary will rewrite your dictated memos and edit them for clarity. It'll take more than cpu horsepower to get a computer to produce readable english prose - it'll take major advances in AI.
Sadly (and I don't mean to be cruel here), it looks like Anand could really benefit from a copy editor or secretary with a good understanding of english grammar and punctuation rules. Some of the sentences in the article are downright obfuscated.
What we really need is an english compiler (gec?):
Warning: Syntax error on line 789 (missing comma?)
Warning: ANSI english forbids implicit verb on line 1031.
Warning: What the hell are you trying to say on line 1532?
It would have to be a lot better than Word's stupid "I don't understand the subjuctive mood" green squiggles, though.
The companies only make provisions like this (in the case of the EULA) because it is necessary.
Is it really? Let's look at the three statements that the article identifies as being most common in an EULA.
1) "First, there is always an assertion of the company's copyright and ultimate ownership of the intellectual property represented by, and associated with the game"
Is this really necessary? Either the company owns the code or it doesn't. What the end user happens to think about the matter is of no effect. If someone steals my car, will I not be able to press charges unless I had previously made the thief agree that the car was mine? Proof of ownership is not a matter for EULAs, but for property law, unless you think I should put a sign on the car door saying "By touching this car, you agree that...."
2) "Second, there is always a promise that the game (or its associated documentation) will not be copied or shared" Again, this is not a matter for EULAs, but rather for copyright law. I am not prevented from photocopying a book or the work of a professional photographer by an EULA or even a copyright symbol, but rather by copyright law.
3) "The third promise every game obtains from the user is that they will not try to reverse engineer or modify the product in any way using the program they've received." This is a more debatable matter, and one I'm not going to get into in this post, except to bring up the "car with the hood welded shut" example and to note that while such provisions are of course beneficial to the software company, they have questionable value to society as a whole.
Why is it that the software market enjoys this unique priviledge of a prohibition on reverse engineering? (Yes, I know that computer make it a little easier, but that is beside the point.) I understand that software developers don't want people ripping off their ideas, but is this really good for society? What if, for example, Gillette had sued Shick for ripping off the idea of the disposable razor and thereby locked up the disposable razor market? "Your honor, we submit that it would have been impossible for Schick to develop a disposable razor without reverse-engineering ours. We even have evidence that a janitor at Schick corporation once saw a Gillette razor!"
Instead, Gillette has to rely on making better and better razors, which I would argue is good for the consumer and society as a whole. (Just think of all the extra trees we'd have to cut down to make toilet paper to dab up the blood from the nicks and cuts from straight razors.)
Anyway, this is all a tangent (it looks like I "got into it" anyway), and I suddenly feel the need to shave.
To the law, clicking "I agree" is different than buying a toaster, even if you wait to flip through the warranty papers back in your easy chair at home: you've still bought that toaster. (The difference has been that most appliance stores would take your toaster back if it turned out you disliked the warranty for some reason: computer retailers have often been less understanding.)
The other (real) difference is that the warranty on the toaster amounts to saying "If you do certain things, we don't have to replace this toaster", whereas a toaster with an EULA might say "You may not allow your friends to make toast with this toaster. You may not toast anything but white bread in this toaster. Once you plug this toaster into one outlet, you may not move it to another outlet. By opening the box to this toaster, you agree that if this toaster explodes and burns down your house, we are not liable, even if we knew about the exploding-toaster bug. This toaster and any toast you make with it remain the property of Hyper-mega-toaser-co, Inc."... and so on.
The typical EULA is nothing like the typical Warranty. The former attempts to retrict the ways in which you may use a product, while the latter just says that if you do stupid things with the product, the company is not liable for any damage you might cause.
Are you replying to my post specifically or just the idea of circumventing RFID chips in general? It seems the latter.
You'll nead to carry a portable radar amplifier with you to zap these chips - ther're pretty tough.
Presumably a "portable EMP generator" might incorporate this (and not be just a 9V battery and a piece of wire)? Since I'm just talking hot air on this one, I'm going to say my imaginary EMP generator would.;) How about this: since the tag is right there, why not deliver a high-power pulse directly to the chip or its ink antenna? Here's something simpler. Locate the little piece of paper with the chip on it. Unbend a paper clip, poke it into the tag and pop the chip off its paper antenna. There are a thousand other ways to accomplish the same thing. The idea was to disable the tag in some way (and it is impossible to prove that it can not be disabled, so don't pick nits) and to take advantage of the automated checkout to look to the cameras like you're buying stuff but to cause selected high-ticket items not to be rung up.
Two - they can be used in conjuction with proven anti-theft technologies.
There is no "proven anti-theft technology" that cannot be disabled by a skilled and knowledgable person. Moreover, the problem is in using these with an automated cart-at-once checkout as seen in that (AT&T I think) commercial (this being the idea behind my original hastily-written post). That is, where you just roll your cart through the exit and it bills your credit card and prints out a receipt. No holding items up one by one to a camera or receiver or tag remover or deactivator.
Take a look at Sensormatics smartEAS - based on Intermecs chip.
I have. Those things are as big and bulky as current anti-theft clothing tags, meaning they would have to be removed at checkout. (The smaller versions look to be attached with adhesive, but that just makes them easy to remove.) I'm not sure how they can claim (as they do) that this will enable "automated inventory returns, active merchandising, self-checkout." I certainly wouldn't want one of those things in my undies poking me you-know where. Therefore, at some point (probably POS), it must be removed. (Sensormatic explicitly says this.) If they are going to sell "readers (fixed and hand-held), detachers and deactivators," what is to stop a disgruntled Wal-Mart employee from lifting a detacher/deactivator and turning it over for reverse-engineering and duplication? (I have, ahem, heard that this has already been done.) Anyway, this still prohibits the automated cart-at-once checkout idea, which was the topic of my original post, so these devices, others like them, and this entire paragraph are not really relevant.
Lastly - jamming won't work easily. The chips work woth frequency hopping interrogatorators.
Who said anything about jamming? Presumably Jamming would just be silly--it's a good way to get busted for shoplifting and get in trouble with the FCC. It's about as smart to disable the interrogators by kicking them over and ripping them out of the floor.
Anyway, the point of my original post was that attempts to use RFID tech to allow self-checkout leaves one open to abuse. Retailers know this, and that's why these RFID tags are probably just going to end up doing three things in the short term:
1) reducing the amount of time the clerk spends looking for the barcode and repeatedly trying to scan the same barcode-less side of an item,
2) providing better inventory control for retailers, though that is debateable (Why do they need to know which box of cereal sold? Because management says so.), and
3) creating more jobs for programmers to re-write inventory systems to handle the new data.
There is, of course, the possibility that retailers will find the level of shoplifting allowed by self-checkout and imperfect security acceptable in light of the ability to lay off a few clerks and shorten lines. Which, as I said, I think is a great idea!;)
I note two things about this article that might make shopping with automated checkouts fun.
First, they say that this BiStatix chip is programmable. It sure would be nice to reprogram the chip and get an Armani suit for the price of a pair of Armani socks. (Maybe even switch the 96-bit ID between the two and leave a nice surprise for somebody else.) Presumably there would have to be someone watching as there is at automated (bar-code) grocery checkouts now.
Second, it's a chip connected to an antenna. What happens when I walk around the store with a portable EMP generator and zap the chips on some portion of the goods I pick up? My cart is full of stuff--how are they going to notice if a few of the more expensive items in the middle of the cart don't register?
Still, very few people are likely to do this, and it may be like credit cards--VISA and MC accept a certain level of fraud in exchange for the business they get.
That said, I think these chips are a great idea!;)
What about rants is supposed to be fair? I thought the point was to blow off steam, not to present a balanced review.;)
Anyway, of course one can just tar up the package and untar it (I'm not an idiot). (You could also use a shared/usr/local, as is the case with our workstations, not that I'm even planning on doing that -- it was just an example to show why GUI installs are not always good, but I digress...) One can also write a script to automate the user end of setup, as I have done. Still, those both amount to bypassing and re-writing the installation program, and the point of the post was that the installation program sucks from a sysadmin point of view. The fact that one is tempted to replace this lauded java installer with a home-grown solution only proves the point. How, then, is this "unfair"?
As for the wrapper, the truth is that I don't want the setup program to be run at all. Hence the aforementioned script.
You want to see something else great and in the same vein as the "/net"? Take a look at the Xpdefaults file ($HOME/.Xpdefaults or $DEST/share/xp3/Xpdefaults):
[windows]
; This file contains the settings for the printer installation, which you
; edit with the program "spadmin".
; The following line specifies the standard printer. It must be equal to
; one of the lines in the section [devices].
device=fooprinter,SGENPRT PostScript,fooprinter
[devices] ...and it goes on like that. Looks just like win.ini. Sigh.
I suppose it depends on your perspective. For a typical home single-user environment, I suppose Star Office's Previous/Next/Finish wizard installation routine might seem comfortingly Windows-ish.
From a system administrator's point of view, however, StarOffice has the most bone-headedly stupid install that I have ever seen. This evening, already tired, grumpy, and low on caffeine, I installed Star Office 5.2 on a machine running Solaris 7 (or whatever it is in Sun's current numbering scheme). Let me elaborate (and sorry about the length, but I'm still grumpy and ranty):
print <<"EndOfRant";
My first complaint is the stupid java-based and pointlessly-GUIed installation program. If I were to install this on our lab full of workstations, I would have to spend an entire day walking from workstation to workstation clicking "Next." (Maybe we're going to have to write an X version of Expect.) You call that easy to install? I call running a script to ssh into each workstation and run "make install" while I read Slashdot easy to install. This is just a waste of my time. Anyway, I digress.
The second thing that I noticed about the installation was that it defaults to being installed in a user's home directory. What? That's just insane! In order to get it to install itself in some reasonable place, the directions seem to indicate that I have to run "./setup/net" (which I suppose has something to do with an expectation that you will be using a shared network filesystem among single-user workstations), except that it's not "/net", it's "-net" (Apparently StarOffice was developed under DOS). That is if you can find the setup program, because it's not where the documentation says it is either (looks like they added multiple language support at the last minute -- it wasn't that hard to figure out, just sort of JV).
The third crazy thing is that the default "-net" install directory is "/office52". What a great idea! I'm just going to stop installing things in "/usr" and "/usr/local", and I'll put them all in one big "/" partition. Anyway, I changed the directory to "/usr/local/staroffice-5.2" because I like descriptive names and don't feel the need to limit myself to 8 characters (there's that DOS thing again).
At this point it gets a little blurry. I think I remember waiting for some little blue bar to move to the right while some sort of slide show went on. I think the slides were trying to sell me on this something called "StarOffice". It looked pretty neat, and I thought about getting it, but then I realized I was already installing it! What is that all about? Sigh.
After the little "complete" dialog came up, I thought I was done. "Not so," said StarOffice. Turns out that each of our 20,000 users was going to have to run a setup script for him- or herself that will install over 2.5 MB of files into some user-defined directory. What? I was in shock. Programs that blindly create little ".program" files and directories all over the place are annoying enough, but even Netscape is smart enough to make ".netscape" all by itself (unless you wrap it in a script that makes the directory and preferences files for Netscape).
This presents new problems:
First, I am obviously going to have to write some sort of wrapper that makes sure setup is run before Star Office is run. Since the destination directory is user-defined, I would have to create some separate tracking mechanism ("touch $HOME/.so52-setup-done" or some crap like that).
Second, the per-user install starts with a scary warning that the machine needs to be patched, as well as other messages which are obviously intended for the system administrator. (Yes, the box is already patched.) I can hear the phone ringing already.
Third, 2.5 MB? Many of our users are already bumping up against their quotas. Another 2.5 MB might just push them and their 99MB inboxes over the edge, and then they'd have to learn how to press "d" in pine or to use gzip, which would surely break their minds.:)
What right does StarOffice have to take up 2.5MB with user config files, anyway? (And what else is it putting there besides config files?) I can understand the directory growing after use, but 2.5MB right off the bat seems a little weird.
Fourth, in the process of playing with and configuring StarOffice my test account, I repeatedly came across messages that
"/usr/local/staroffice-5.2/foo/bar" is not writable. Well of course it isn't! Isn't that what the stuff in the home directory is for? I guess not. Odd. (Clearly this is more hold-over from StarOffice's single-user roots.)
EndOfRant
What makes this all so crazy is that this software is (well is now, anyway) from a division of Sun "The Network is the Computer", Inc. and was being installed on a machine and operating system designed and built by Sun, Inc. Why, then, is this software so poorly adapted to a multi-user/shared/distributed environment? It's just weird.
Anyway, it's not that big of a deal. I just wanted to point out that one person's "easy graphical wizard" is another person's "child of satan that makes me take my hands away from the keyboard to use a rodent."
Whew. That's a long post. Tomorrow morning I'm going to write some little perl-based installation routine to bypass all of the GUI crap for each user. (I'll be nice and pop up the license, but the rest can be skipped.) By then, I will have probably discovered some "--skip-java-gui" and "--dont-hog-my-home-directory" flags for the install, but for now the "GUI rage" is still fresh on my CLI fingertips.
Alright, before I hear any more complaints about the data rate, keep in mind that Galileo is 50 light minutes from Earth. It's one thing to get 100Mbps over twisted pair copper between machines a kilometer apart, but quite another to do it by radio over about 900,000,000 times that distance (by some quick back of the envelope calculations) with a broken antenna.
Bottom line: don't mess with NASA -- chances are one of their engineers wrote your network driver.
Slashdot Mirror
"Senator, there are exactly 23 card-carrying members of the Communist Party at the FSF."
Sometimes I think we watch too much Star Trek (or, more likely, Star Trek, specifically, ST:NG reflects something that I don't like about our culture). Anyway, have you ever noticed that a ridiculously large proportion of ST:NG episodes feature some conflict that is resolved by Mr. LaForge saying something like "maybe, if we reverse the polarity on the field generator..." or Dr. Crusher saying "maybe, if we alter the microbes' DNA..."
My point is (and this is perhaps not so directly applicable to the article, but is reminiscent, anyway): Why is it that we so often look for technology to provide a quick fix for what is really a very complex and difficult social problem? Obviously, it's the easy way out, but does it really work? Think about all the various technical schemes for locking down copyrighted content that we have been discussing lately.
If we can come up with some technical way to reduce the CO2 output of smokestacks, then great. However, I still worry that unless we (and particularly we Americans) wake up and take up the difficult task of addressing all the causes of our massive CO2 output, this new method of scrubbing CO2 is not going to be enough.
Anyway, go cyanobacteria, because every little bit helps.
I know a lot of people here are very anti-regulation, but I think it would be great if case law established that web robots must obey the Robots Exclusion Standard. Since it's a widely-known standard, I think it can be fairly argued that robots that choose to disregard /robots.txt are in danger of tresspassing to chattels. Using the standard also would allow bots to fulfill their helpful role, while providing a clear distinction between what is and what is not acceptable.
/robots.txt
Sure, one might argue that people might be unaware of the standard, but that is seldom an excuse. I may be unaware of fire/electrical codes, but I'll still get in trouble if I don't adhere to them, because I'm putting others at risk and thereby imposing a cost upon society (fire trucks and insurance don't come free). Web crawlers that index data in violation of the Robots Exclusion Standard impose a cost on companies and society just as well, in the end requiring people to by bigger pipes, faster servers, and so on (thereby using more power, dumping more old computer components into landfills and more chipmaking chemicals into the environment).
My point is that web crawler operators live in a society, just like everyone else, and they too must be held accountable for the consequences of their actions, particularly when they willfully disregard the requests of web site operators as expressed in
Is it just me, or does the article have this backwards? Sure, I'm just nit-picking, but it seems that one would install such metal fibers to promote conduction and thus prevent the build-up of static charge. Consider conductive anti-stat floor mats, wrist straps, et cetera, which make this seem not-so unusual (except that it's built into the carpet).
:)
Oh well, anyway, it's still all pretty cool, and I envy the folks who get to work there (mainly for the nice cabling setup).
For running charities, perhaps, but for getting a job? Take a look at the SANS 2000 Salary Survey Summary:
"The average for UNIX folks was $70,080, while the average for Windows folks was $61,233."
Granted, this only applies to SysAdmins (so I probably would agree with you in general), but given the currently high demand for "UNIX folks" (and despite the fact that Linux is not technically UNIX), it seems that training people in Linux is not such a bad idea. Speaking more anecdotally: I was able to get my current job as a UNIX admin because I got so sick of working in retail that I decided to train myself with Linux (plus it was interesting, fun, and free). I know of several other people who have done similar things, and I even have friends involved in hiring that say that Microsoft certifications actually make a resume look worse (one said that he threw away a resume as soon as he saw the letters "MCSE").
I should also add that Linux is an excellent platform on which to learn programming, which opens up yet another career.
Augh...wait a minute; what am I saying? Train no one in Linux! The fewer people there are who know how to use it, the more in-demand those of who do know will be. (Now you know the real reason why people report that Linux is hard to use
For one thing, it's really shameful that lines like these two from acct.c:
ac.ac_io = encode_comp_t(0
ac.ac_rw = encode_comp_t(ac.ac_io / 1024);
have been hanging around for so long (and in linux/kernel/, at that!). As I recall, a patch for this was submitted for 2.3.x, but I guess it went nowhere.
Process accounting may not seem like a big deal to the typical desktop user, but for sysadmins (especially on big systems), it can be really important.
This is not to say that I think keeping silent is right (taking responsibility for your mistakes is the Right Thing To Do), but it is certainly understandable.
Very well, you're right. You happen to have caught me in a grumpy mood today, and I admit I was probably looking for a fight. I certainly thought that you were trying to offer one, but I see that I was mistaken.
Whether or not Jupiter has a solid core is still being debated. Most notable astrophysicists agree that there is no way of knowing if Jupiter has a core mass or not. The last time I heard anyone argue FOR the solid core was 1989.
Of course it's impossible to know for certain if Jupiter has a rocky core without direct measurement, and some people may not believe until they reach out and touch it with their hands, but that's not happening anytime soon, and is irrelevant. As for the '1989' date, you seem to be implying that no one believes in the rocky core anymore. Perhaps, then, you should try the following experiment. Go to Google. Enter "jupiter rocky core". Press "I'm feeling lucky." See JPL mention rocky core. Or, just press "Google Search" and read links that deal with the concept more directly. This, of course, does not prove that the core exists, but rather goes to refute the idea that you present above, which is that scientists no longer believe that it exists.
Maybe that's just me. You're right, though, it is believed to have a rocky core. But it is also believed to not have one. You tell me to disregard Galileo's data but you don't have a problem quoting implied data from Pioneer 10 and 11. If you meant that YOU believe it has a rocky core, I would like some more justification.
You're right. While you (I assume you wrote the post to which I replied, AC) can come out and state that 'there is no "solid" mass on Jupiter' and give no justification whatsoever, mere mortals such as I must give it. And I didn't say that one should disregard Galileo's data, merely that not finding the expected abundance of water along the probe's entry path (which infrared images revealed to be a relatively unclouded area) does not imply a total absence of water on Jupiter. I never argued that there was as much water as we originally thought, merely that there is some water.
What "rocky core" do you propose can sustain it's solid phase at the awesome pressure and temperature at the core of Jupiter, after being immersed in liquid metallic hydrogen?
Sure, the temperature is an issue, but the "awesome pressure" would encourage most materials to go into solid phase (except for weird ones like water, of course). And what the heck does the liquid metallic hydrogen sitting above it have to do with anything? Are you proposing that it would dissolve the core or do you just think the words "liquid metallic hydrogen" are going to convince me that you know what you're talking about?
Impact events would make no difference one way or another, unless you are implying that Jupiter collided with another planet. That could give it a core mass.
Oh, sure, that's scientific. So, an asteroid couldn't do it, but a planet could? What about a really big asteroid, or a really small planet? My point is: what magic line determines that a body is big enough for you to dub it a "planet" and simltaneously allow your imagination to grant it the power to give Jupiter a rocky core, where all other impacts "would make no difference"? Give me a break.
Where did this theoretical planet come from? Extrasolar?
Oh, sure... the only bodies around while the solar system was forming are the ones we see today, is that what you are saying? Your argument keeps getting better and better.
Most of the extrasolar planets we have ever discovered are gas giants themselves, and larger than Jupiter to boot.
Right, this is irrelevant, but I suppose you believe that those numbers are in no way skewed by the detection methods currently in use? Obviously if we look for planets by looking at the gravitational wobble of stars, we're going to detect massive ones that orbit close to their stars. Heck, that much is even in the article to which this story referred, if you were unable to figure it out for yourself using basic physics. If there were any previous indication that you were scientifically qualified, this last paragraph is squashing that.
Which would mean Jupiter used to BE an extrasolar planet, and collided with a smaller planet with the orbit Jupiter now holds. If that is your theory, I advise you to examine the feasibility of it.
Wow. Where are you getting these ideas and whatever it is that you are smoking? That's quite a ridiculous theory that you put in my mouth, and I will shortly spit it back at you.
Well that was fun. Let me take some time out at the end here, leaving all the problems with what you just said behind, to explain the problems with your initial argument. The bottom line is that you have made statements that cannot be reasonably defended.
You say that "there is no water on Jupiter" [emphasis mine]. Now, I could talk about solar oxygen levels and the presumed abundance of oxygen in the early solar system and so on, but I don't have to. I'm even willing to believe that Jupiter has a relatively low oxygen abundance in its atmosphere. Nevertheless, we all saw a bunch of chunks of frozen water slam into Jupiter a few years back. Discarding for the sake of argument the idea that there was any water present on Jupiter before the impacts, if one single molecule of water from the impacts survived, then there is water and it's on Jupiter, so you lose; end of argument.
Similarly, you say "there is no 'solid' mass on Jupiter". Again, if there is a single speck of dust on Jupiter, you lose; end of argument.
The smart thing to do would have been to admit that you were exaggerating (that's why I gave you that "out" at the top of my post), and then we could have settled into a more reasonable discussion of planetary formation and elemental abundances. Instead you went off on these weird tangents, became insulting, and capped it off with ridiculous statements like the one about extrasolar planets mostly being giants, which thoroughly discredits you, made it a joy to write this reply, and makes it certain that neither I nor anyone else will pay any mind to anything more you have to say.
Alright that's it. It's time to start working on a poor man's massive IDE RAID solution. I just got back from talking with someone else about this, and now I'm psyched.
There is no water on Jupiter.
Ummm, that's just plain wrong. Sure, our probe found less water than expected, but it dropped into a hot spot. Besides, do you think Shoemaker-Levy contained no water at all?
If there is any oxygen at all there, it's in very trace amounts
Well sure, everything that's not hydrogen or helium on Jupiter you might call "trace," since that's not really a quantitative term. Still, Jupiter is big enough that there is plenty of H2O and CO. I would advise against treating the Galileo probe's measurements as authoritative for Jupiter's entire atmosphere. There very well may be concetrations of oxygen in certain portions of that vast ocean of gas.
There is no "solid" mass on Jupiter, unless you consider pressurized hydrogen a solid mass.
Also incorrect. Jupiter is believed to have a rocky core. Do you really think that in the enitre history of Jupiter, nothing rocky ever hit it and sank through all that gas?
RAID by itself is no substitute for backups. Sure, it protects against the failure of a disk, but what about other sorts of disasters, like bad RAID controller firmware, fire, flood, filesystem corruption, hostile intrusion, and so on?
Of course, there are still the biggest use of backups: user carelessness.
Of course, with the falling price of IDE drives, I have been thinking that it might be possible to build some IDE-RAID-based storage system that would rival tape libraries in cost/terabyte, but be much faster. (Maybe this is what you meant.) You could still make tapes for offsites and archives and such. Unfortunately, all the commercial products out there that do something like IDE RAID -> SCSI (which one might need to keep things simple) are way more expensive than their components would seem to justify, and I have no money, so it remains a pipe dream.
I suppose you could just get tons of pentium systems doing software RAID and NFS.... hmmmm.... this sounds like a job for those folks at NASA who brought us Beowulf (and all the "imagine a Beowulf cluster of [ ]" on Slashdot.
Sadly (and I don't mean to be cruel here), it looks like Anand could really benefit from a copy editor or secretary with a good understanding of english grammar and punctuation rules. Some of the sentences in the article are downright obfuscated.
What we really need is an english compiler (gec?):
Warning: Syntax error on line 789 (missing comma?)
Warning: ANSI english forbids implicit verb on line 1031.
Warning: What the hell are you trying to say on line 1532?
It would have to be a lot better than Word's stupid "I don't understand the subjuctive mood" green squiggles, though.
Is it really? Let's look at the three statements that the article identifies as being most common in an EULA.
1) "First, there is always an assertion of the company's copyright and ultimate ownership of the intellectual property represented by, and associated with the game"
Is this really necessary? Either the company owns the code or it doesn't. What the end user happens to think about the matter is of no effect. If someone steals my car, will I not be able to press charges unless I had previously made the thief agree that the car was mine? Proof of ownership is not a matter for EULAs, but for property law, unless you think I should put a sign on the car door saying "By touching this car, you agree that...."
2) "Second, there is always a promise that the game (or its associated documentation) will not be copied or shared" Again, this is not a matter for EULAs, but rather for copyright law. I am not prevented from photocopying a book or the work of a professional photographer by an EULA or even a copyright symbol, but rather by copyright law.
3) "The third promise every game obtains from the user is that they will not try to reverse engineer or modify the product in any way using the program they've received." This is a more debatable matter, and one I'm not going to get into in this post, except to bring up the "car with the hood welded shut" example and to note that while such provisions are of course beneficial to the software company, they have questionable value to society as a whole.
Why is it that the software market enjoys this unique priviledge of a prohibition on reverse engineering? (Yes, I know that computer make it a little easier, but that is beside the point.) I understand that software developers don't want people ripping off their ideas, but is this really good for society? What if, for example, Gillette had sued Shick for ripping off the idea of the disposable razor and thereby locked up the disposable razor market? "Your honor, we submit that it would have been impossible for Schick to develop a disposable razor without reverse-engineering ours. We even have evidence that a janitor at Schick corporation once saw a Gillette razor!"
Instead, Gillette has to rely on making better and better razors, which I would argue is good for the consumer and society as a whole. (Just think of all the extra trees we'd have to cut down to make toilet paper to dab up the blood from the nicks and cuts from straight razors.)
Anyway, this is all a tangent (it looks like I "got into it" anyway), and I suddenly feel the need to shave.
To the law, clicking "I agree" is different than buying a toaster, even if you wait to flip through the warranty papers back in your easy chair at home: you've still bought that toaster. (The difference has been that most appliance stores would take your toaster back if it turned out you disliked the warranty for some reason: computer retailers have often been less understanding.)
The other (real) difference is that the warranty on the toaster amounts to saying "If you do certain things, we don't have to replace this toaster", whereas a toaster with an EULA might say "You may not allow your friends to make toast with this toaster. You may not toast anything but white bread in this toaster. Once you plug this toaster into one outlet, you may not move it to another outlet. By opening the box to this toaster, you agree that if this toaster explodes and burns down your house, we are not liable, even if we knew about the exploding-toaster bug. This toaster and any toast you make with it remain the property of Hyper-mega-toaser-co, Inc."
The typical EULA is nothing like the typical Warranty. The former attempts to retrict the ways in which you may use a product, while the latter just says that if you do stupid things with the product, the company is not liable for any damage you might cause.
Are you replying to my post specifically or just the idea of circumventing RFID chips in general? It seems the latter.
;) How about this: since the tag is right there, why not deliver a high-power pulse directly to the chip or its ink antenna? Here's something simpler. Locate the little piece of paper with the chip on it. Unbend a paper clip, poke it into the tag and pop the chip off its paper antenna. There are a thousand other ways to accomplish the same thing. The idea was to disable the tag in some way (and it is impossible to prove that it can not be disabled, so don't pick nits) and to take advantage of the automated checkout to look to the cameras like you're buying stuff but to cause selected high-ticket items not to be rung up.
;)
You'll nead to carry a portable radar amplifier with you to zap these chips - ther're pretty tough.
Presumably a "portable EMP generator" might incorporate this (and not be just a 9V battery and a piece of wire)? Since I'm just talking hot air on this one, I'm going to say my imaginary EMP generator would.
Two - they can be used in conjuction with proven anti-theft technologies.
There is no "proven anti-theft technology" that cannot be disabled by a skilled and knowledgable person. Moreover, the problem is in using these with an automated cart-at-once checkout as seen in that (AT&T I think) commercial (this being the idea behind my original hastily-written post). That is, where you just roll your cart through the exit and it bills your credit card and prints out a receipt. No holding items up one by one to a camera or receiver or tag remover or deactivator.
Take a look at Sensormatics smartEAS - based on Intermecs chip.
I have. Those things are as big and bulky as current anti-theft clothing tags, meaning they would have to be removed at checkout. (The smaller versions look to be attached with adhesive, but that just makes them easy to remove.) I'm not sure how they can claim (as they do) that this will enable "automated inventory returns, active merchandising, self-checkout." I certainly wouldn't want one of those things in my undies poking me you-know where. Therefore, at some point (probably POS), it must be removed. (Sensormatic explicitly says this.) If they are going to sell "readers (fixed and hand-held), detachers and deactivators," what is to stop a disgruntled Wal-Mart employee from lifting a detacher/deactivator and turning it over for reverse-engineering and duplication? (I have, ahem, heard that this has already been done.) Anyway, this still prohibits the automated cart-at-once checkout idea, which was the topic of my original post, so these devices, others like them, and this entire paragraph are not really relevant.
Lastly - jamming won't work easily. The chips work woth frequency hopping interrogatorators.
Who said anything about jamming? Presumably Jamming would just be silly--it's a good way to get busted for shoplifting and get in trouble with the FCC. It's about as smart to disable the interrogators by kicking them over and ripping them out of the floor.
Anyway, the point of my original post was that attempts to use RFID tech to allow self-checkout leaves one open to abuse. Retailers know this, and that's why these RFID tags are probably just going to end up doing three things in the short term:
1) reducing the amount of time the clerk spends looking for the barcode and repeatedly trying to scan the same barcode-less side of an item,
2) providing better inventory control for retailers, though that is debateable (Why do they need to know which box of cereal sold? Because management says so.), and
3) creating more jobs for programmers to re-write inventory systems to handle the new data.
There is, of course, the possibility that retailers will find the level of shoplifting allowed by self-checkout and imperfect security acceptable in light of the ability to lay off a few clerks and shorten lines. Which, as I said, I think is a great idea!
Corgha
I note two things about this article that might make shopping with automated checkouts fun.
;)
First, they say that this BiStatix chip is programmable. It sure would be nice to reprogram the chip and get an Armani suit for the price of a pair of Armani socks. (Maybe even switch the 96-bit ID between the two and leave a nice surprise for somebody else.) Presumably there would have to be someone watching as there is at automated (bar-code) grocery checkouts now.
Second, it's a chip connected to an antenna. What happens when I walk around the store with a portable EMP generator and zap the chips on some portion of the goods I pick up? My cart is full of stuff--how are they going to notice if a few of the more expensive items in the middle of the cart don't register?
Still, very few people are likely to do this, and it may be like credit cards--VISA and MC accept a certain level of fraud in exchange for the business they get.
That said, I think these chips are a great idea!
What about rants is supposed to be fair? I thought the point was to blow off steam, not to present a balanced review. ;)
Anyway, of course one can just tar up the package and untar it (I'm not an idiot). (You could also use a shared /usr/local, as is the case with our workstations, not that I'm even planning on doing that -- it was just an example to show why GUI installs are not always good, but I digress...) One can also write a script to automate the user end of setup, as I have done. Still, those both amount to bypassing and re-writing the installation program, and the point of the post was that the installation program sucks from a sysadmin point of view. The fact that one is tempted to replace this lauded java installer with a home-grown solution only proves the point. How, then, is this "unfair"?
As for the wrapper, the truth is that I don't want the setup program to be run at all. Hence the aforementioned script.
You want to see something else great and in the same vein as the "/net"? Take a look at the Xpdefaults file ($HOME/.Xpdefaults or $DEST/share/xp3/Xpdefaults):
...and it goes on like that. Looks just like win.ini. Sigh.
[windows]
; This file contains the settings for the printer installation, which you
; edit with the program "spadmin".
; The following line specifies the standard printer. It must be equal to
; one of the lines in the section [devices].
device=fooprinter,SGENPRT PostScript,fooprinter
[devices]
Corgha
I suppose it depends on your perspective. For a typical home single-user environment, I suppose Star Office's Previous/Next/Finish wizard installation routine might seem comfortingly Windows-ish.
/net" (which I suppose has something to do with an expectation that you will be using a shared network filesystem among single-user workstations), except that it's not "/net", it's "-net" (Apparently StarOffice was developed under DOS). That is if you can find the setup program, because it's not where the documentation says it is either (looks like they added multiple language support at the last minute -- it wasn't that hard to figure out, just sort of JV).
:)
From a system administrator's point of view, however, StarOffice has the most bone-headedly stupid install that I have ever seen. This evening, already tired, grumpy, and low on caffeine, I installed Star Office 5.2 on a machine running Solaris 7 (or whatever it is in Sun's current numbering scheme). Let me elaborate (and sorry about the length, but I'm still grumpy and ranty):
print <<"EndOfRant";
My first complaint is the stupid java-based and pointlessly-GUIed installation program. If I were to install this on our lab full of workstations, I would have to spend an entire day walking from workstation to workstation clicking "Next." (Maybe we're going to have to write an X version of Expect.) You call that easy to install? I call running a script to ssh into each workstation and run "make install" while I read Slashdot easy to install. This is just a waste of my time. Anyway, I digress.
The second thing that I noticed about the installation was that it defaults to being installed in a user's home directory. What? That's just insane! In order to get it to install itself in some reasonable place, the directions seem to indicate that I have to run "./setup
The third crazy thing is that the default "-net" install directory is "/office52". What a great idea! I'm just going to stop installing things in "/usr" and "/usr/local", and I'll put them all in one big "/" partition. Anyway, I changed the directory to "/usr/local/staroffice-5.2" because I like descriptive names and don't feel the need to limit myself to 8 characters (there's that DOS thing again).
At this point it gets a little blurry. I think I remember waiting for some little blue bar to move to the right while some sort of slide show went on. I think the slides were trying to sell me on this something called "StarOffice". It looked pretty neat, and I thought about getting it, but then I realized I was already installing it! What is that all about? Sigh.
After the little "complete" dialog came up, I thought I was done. "Not so," said StarOffice. Turns out that each of our 20,000 users was going to have to run a setup script for him- or herself that will install over 2.5 MB of files into some user-defined directory. What? I was in shock. Programs that blindly create little ".program" files and directories all over the place are annoying enough, but even Netscape is smart enough to make ".netscape" all by itself (unless you wrap it in a script that makes the directory and preferences files for Netscape).
This presents new problems:
First, I am obviously going to have to write some sort of wrapper that makes sure setup is run before Star Office is run. Since the destination directory is user-defined, I would have to create some separate tracking mechanism ("touch $HOME/.so52-setup-done" or some crap like that).
Second, the per-user install starts with a scary warning that the machine needs to be patched, as well as other messages which are obviously intended for the system administrator. (Yes, the box is already patched.) I can hear the phone ringing already.
Third, 2.5 MB? Many of our users are already bumping up against their quotas. Another 2.5 MB might just push them and their 99MB inboxes over the edge, and then they'd have to learn how to press "d" in pine or to use gzip, which would surely break their minds.
What right does StarOffice have to take up 2.5MB with user config files, anyway? (And what else is it putting there besides config files?) I can understand the directory growing after use, but 2.5MB right off the bat seems a little weird.
Fourth, in the process of playing with and configuring StarOffice my test account, I repeatedly came across messages that
"/usr/local/staroffice-5.2/foo/bar" is not writable. Well of course it isn't! Isn't that what the stuff in the home directory is for? I guess not. Odd. (Clearly this is more hold-over from StarOffice's single-user roots.)
EndOfRant
What makes this all so crazy is that this software is (well is now, anyway) from a division of Sun "The Network is the Computer", Inc. and was being installed on a machine and operating system designed and built by Sun, Inc. Why, then, is this software so poorly adapted to a multi-user/shared/distributed environment? It's just weird.
Anyway, it's not that big of a deal. I just wanted to point out that one person's "easy graphical wizard" is another person's "child of satan that makes me take my hands away from the keyboard to use a rodent."
Whew. That's a long post. Tomorrow morning I'm going to write some little perl-based installation routine to bypass all of the GUI crap for each user. (I'll be nice and pop up the license, but the rest can be skipped.) By then, I will have probably discovered some "--skip-java-gui" and "--dont-hog-my-home-directory" flags for the install, but for now the "GUI rage" is still fresh on my CLI fingertips.
Corgha
Alright, before I hear any more complaints about the data rate, keep in mind that Galileo is 50 light minutes from Earth. It's one thing to get 100Mbps over twisted pair copper between machines a kilometer apart, but quite another to do it by radio over about 900,000,000 times that distance (by some quick back of the envelope calculations) with a broken antenna.
Bottom line: don't mess with NASA -- chances are one of their engineers wrote your network driver.
Corgha