You're right, the OP misses an important component: culture. The practices and traditions that shape a culture will affect how those people respond to a particular governmental framework.
For instance, compare the 18th century American revolution with the French one. Both produced roughly equivalent forms of government, but the French revolution was far more brutal, probably the result of people having long been oppressed and ruled by sometimes-incompetent monarchs. The American colonists, by contrast, had several generations to forget about why they left England, and the American revolution was much more subdued. The rhetoric used to justify each conflict was far more philosophical on the American side: it was an attempt to correct the failings of foreign governance. French rhetoric saw their fight as nothing less than emancipation of the people.
As a result, French society traditionally put more of an emphasis on government-in-service-of-the-people than American society, who see government as more of a protectorate of the state. The United States has had invasive intelligence-gathering for a long time, and I think the reason why most Americans don't care is because they still see the state as a benevolent entity. I suspect that the French do not share this opinion, especially in the case of Sarcozy.
It still doesn't matter. It's speculated that the finest resolution a spy satellite can get is in the 5-10cm range, and that's probably using many digital imaging tricks. TFA doesn't say what kind of resolution the software requires, but I doubt that 5-10cm is fine enough-- when I walk, there's probably, what, 2-3cm of bounce in my step?
Still, this is clever idea. Attention conspiracy theorists: make sure to walk outside only at noontime. At the equator.
I consider myself a pretty typical geek, but I have no idea what that means. It's not ringing any bells with the Dijkstra that I'm familiar with, and in recursive-geek-irony, googling "Dijkstra's Scorpio" yields your very post. Someone care to enlighten me?
It depends on your requirements. If you're going to be at full cooling load all of the time, it's better to go with a traditional unit. That's what we have, as our unit was chosen based on the expected heat load for the room (6 tons). But if your expected heat load is not constant, you're right, it would probably be better to go with an inverter-based unit. For our heat load, an inverter unit would have been cost-prohibitive and also unnecessary.
Not to mention-- adding another AC unit without that unit being tied into the same controller will cause problems for your AC units. They'll be turning on and shutting off more frequently. This will greatly affect the life of your AC units. If you're a hardware hacker, you can probably add capacity on the cheap by hacking your thermostat to coordinate multiple units. Otherwise, you really do just need to pony up and pay for the AC upgrade. Cool air is a basic business expense nowadays. If the people writing your budget don't see it that way, then your company is in trouble.
Why, are China and India doing basic science research? My impression that pretty much *everyone* is getting out of the game. Deregulating telecom and breaking up AT&T did wonders for telephone customers, but it did not do good things for smart people with big budgets. Consider the fact that UNIX started as an excuse to hack on computer games.
But there's loads of software that runs on x86. Developer time costs magnitudes more than equipment, and for many places, my workplace included, these things are just great. If you're lucky enough to have software that's easily portable, and you don't mind dealing with the inevitable byte-order issues, character encoding issues, and other problems, by all means, use specialized machines. x86 is just so damned flexible that it's worth the extra cost.
We use Soekris [soekris.com] boxes all over the place because their power requirements are minimal, they can withstand harsh environments, they don't need cooling, they run x86 software, they're designed to run headless, and they don't take up much space. So even though you could buy a much more capable (speed- and expansion-wise) Wal-Mart PC for about the same price, the Soekris boxes still win out.
Yeah, that's why I speculated that SQL might be done so easily-- Oracle really is a rabbit hole. I've done some relational algebra in a database course (and also was exposed to set theory in my discrete maths course), but it was unclear to me whether query optimizers actually broke a query down into relational algebra or not. In fact, I remember that despite having had prior experience with SQL, relational algebra was much easier for me to wrap my head around than SQL. My professor was hesitant to go too much into optimizers, though, since most of that is implementation-dependent. He thought it was much more important to talk about ACID, so that's what we spent a good deal of time on.
Actually, the two are paired: programming model and implementation. The reason there's a programming model is that functional methods allow Google's implementation to automatically parallelize the input data for feeding to the cluster. So the implementation is very important, because that's actually how the data is processed and returned.
In that sense, Oracle's clustering optimizations are also a paired programming model and implementation, since, presumably, you need to know Oracle's SQL language extensions in order to take advantage of them (disclaimer: I don't use Oracle). From what I understand about functional programming, SQL should be ideally positioned to take advantage of these kinds of optimizations, since the actual implementation details of any SQL query are always left to the query optimizer, SQL being a declarative language. I'm going to speculate wildly and say that you could probably write a SQL interpreter using a functional style as well, and that good ones probably already do.
He's not saying they're the same. He's saying that they have their roots in the same idea. That idea is, as far as is practical, the elimination of a ruling class and the fullest actualization of human freedom. How they differ is in what extent they consider to be practical. Anarchists are the ultimate capitalists: let natural human behavior self-regulate, i.e., there is no government (or, at least, no compulsory government). Communists believe that compulsory government is still necessary, but should only exist insofar as it helps to bring about the communist ideal of freedom.
You're focusing too much on labels here: a "political philosophy" thus implemented becomes a socioeconomic structure. China calls its political philosophy "communism", but anyone can see that China's socioeconomic structure is anything but. Calling itself "communist" is simply political theatre.
Hmm, interesting. It looks like MSIE 7 does do privilege separation, along with a couple other protections. I stand corrected. In my defense, I do not have Vista, so I was unaware of this.
All in all I think the web stack is pretty secure by default, when comparing it to the alternatives.
Really? My opinion is that the "web stack" (not sure which stack you mean here; MSIE-Windows, FF-Windows, Safari-MacOSX, Konq-Linux, etc) has by far the worst record so far. MSIE-Windows has to be the #1 vector for infection now, and has been for at least the last 6-7 years. Which alternative are you thinking of? Because the "web stack" is, in my opinion, the premier virus runtime environment.
My opinion is that web designers made a HUGE mistake in not treating network input cautiously. The emphasis has been on "rich APIs", "data structure passing", extensibility, desktop integration, and so on. These are undoubtably good things in the absence of malicious input, but the fact is, there is a lot of malicious input out there. Web browsers would benefit greatly from some simple privilege separation; the Mozilla camp could do this with some effort, but MSIE is pretty much dead in the water here due to the level of integration with the base system. I understand the HTML5 camp's worry that Flash/Flex will become a de facto standard, but in my opinion, web security has not been taken seriously enough. These kinds of vulnerabilities have become a major source of income for organized crime in the East, and still people like you are saying that security is not the most important issue? Gimme a break.
I use something similar: OpenBSD's spamd. The caveat is that, rarely, legitimate mail is rejected. At work, typically the person on the other end will contact us some other way, and we'll whitelist them. Yahoo is the one egregious offender here; their relays never retry from the same box, so spamd thinks that it's a new connection and starts the connection count over again. But I've found, in general, that the false positive rate with spamd alone is lower than the false positive rate with SpamAssassin alone, and that OpenBSD's spamd is far more effective in general. The only "spam" we see now is the stuff that users signed up for; mailing lists, mostly. Stuff that-- back when we were using Bayesian filters-- users would classify as spam and foul up the the spam pool for the Bayes engine.
I hate to break it to you, but LDAP is not a directory system. It is a directory protocol. AD provides an LDAP interface. So your directory system can be structured and provide storage in the backend pretty much any way you want. Microsoft, for instance, uses Jet for storing their data, and X.500 for structuring it. But if you wanted to build your directory using post-it notes and robot, then fine, as long as you provide an LDAP interface, you're an LDAP directory.
AD *can* store any arbitrary information with schema additions. So if you can query LDAP on the Linux side for window manager policy, and you can come up with a schema that represents that policy, go ahead, store it in AD. Mac people have been doing this for years, although Apple would prefer that you use their Open Directory system.
Also-- AD uses Kerberos. How do I know? Because I have Linux machines (MIT Kerb), OpenBSD machines (Heimdal), and Macs (MIT/Apple Kerb) all authenticating against our AD. There are some little oddities here are there (your machines have to support Microsoft's cipher-- which I believe is now installed by default on all recent Kerberos distributions), but in general, it works surprisingly well. For me, on Linux machines, the trick was learning the ins and outs of PAM and winbind. After that, it was easy.
Anyway, if you're expecting LDAP to provide authentication, you're mistaken about the purpose of LDAP. Think of it as a fancy phone book. What you need are a lock and key. Also-- accounting? For that, you want a piece of logging software. Microsoft supplies all of these things neatly packaged together, and if you don't want to bother with the details, then it's a decent choice. But don't confuse the two, because LDAP only provides a subset of the services that AD does. Complaining that LDAP does a "shit job" at authentication and accounting is like complaining that your tires do a "shit job" of steering. Well, duh.
There's a glyph. You don't know what it means. You might not want to activate the control it is connected to-- what if it means "delete my document?" So how do you find out what it means? You can't look it up, because it's a picture. The best you can do is describe it, in words, to another person, who will give you an answer, in words. So you see, language is the lowest common denominator here. Transcription of language, therefore, should be the thing we strive for-- oh, wait, we have that: the written language. Why use a computer if you can't read? Because, come on, most of the Internet is useless unless you are literate. It's a machine for processing the semantic and logical relationships of text.
You can't tell me that this symbol is in any way intuitive. I remember the first time when I saw it, sometime in the 90's. I remember thinking "what the fuck does that mean?" Sure, have some fancy symbol, but then, maybe would it kill you to print POWER underneath it? I think language is FAR more expressive than an abstract picture. Once that picture has been internalized by everyone, it is language. Why make it harder for everybody when we already have symbols that describe what you want to say?
As of today, we've extricated ourselves from the hell that was Xserves. We purchased a number of these machines because it seemed like an easy and cheap way to get a fileserver going that did both AFP and SMB, was AD-integrated, and could have its file store on a SAN. Well, after much money and a year later, the answer is that Apple very much oversold their ability to integrate into a Windows environment. Here are my gripes:
AD-binding is not straightforward. Apple really wants you to run an OpenDirectory, as this allows you to both manage Apple desktops and do single-sign on. If you just want to allow AD authentication on your MacOS X servers, good luck. You're in for a bugfest, with partially-working GUIs and many, many quirks.
#1 quirk being: you can't do cross-domain authentication, even if those domains are trusted. This was a showstopper for us.
There is only ONE backup application for Xsan that is both a) reliable, and b) has a reasonable support contract. We tried Retrospect (total POS), Veritas (ridiculous wait times for support), and finally, BRU. BRU has a decent product, but the number of MacOS bugs that plague this application make it unreliable and frustrating to use. OSS applications don't handle the numerous HFS+ corner cases. Rsync, which we used for snapshots, routinely hemorrhaged itself on files with extended attributes, despite the fact that this was APPLE'S OWN VERSION.
Ever try running a shared AFP/SMB volume on an Xsan? You can't. Surprise, surprise: Xsan is not HFS+ formatted. It uses CVFS, which is a Quantum/ADIC filesystem. Why? Because Xsan is simply a rebadged version of StorNext! So your AFP daemon will spew Mac metadata everywhere which your SMB daemon will not honor, thus totally corrupting your data. Fuck you, Apple. Seriously.
You can't modify MacOS X Server files on the command line. Oh, well, you could on 10.4 server; then lock the file and hope you never had to use the GUI again. But on 10.5, even that does not work-- it still overwrites your file; smb.conf is a perfect example. I figured, OK, maybe I should set the immutable flag, but then I started thinking... WHY am I using Apple products again?
Apple's enterprise support blows. Sometimes you get an answer, but no matter what, expect a long wait while people on the other end decide whether they want to bother answering your question or not. Want to follow-up on a bug that someone else reported? Good luck. Their bug reporter is terrible. Would it be so hard to run Bugzilla?
Apple needs to get their shit together. Unless your needs are VERY straightforward, even 10.5 does not solve them. I'll admit that 10.5 has a much nicer server admin GUI, but it does not overcome the problems with the platform.
We've moved all of these services to CentOS machines. By contrast, getting them working reliably was a walk in the park. Equivalent hardware (hotswap RAID (SCSI, I should add), redundant PSU, fiber channel card, GigE, dual processor machines in a 3U form factor (SuperMicro chassis) come out to about $1k less than an Xserve, on average. And when a part dies, like a backplane, I can BUY THAT PART. With Apple, you have to buy an entire parts kit, which comes with stuff you may not want.
We now run Samba and Netatalk on CentOS on generic server hardware, connected to our StorNext network. There may be better SAN stuff out there than StorNext (in fact, their licensing department leaves much to be desired-- do they even know how to use their own product?), but we already had a lot invested (three Xserve RAID cabinets). Things run great now, and with the Linux version of BRU, our full tape backup [inexplicably] finishes 9 hours earlier (used to take 60 hours, now takes 51).
My advice: Apple makes some nice desktops, but their server stuff is only for novices. I went into the experience very optimistic about Apple's stuff, but now I have a very bitter taste in my mouth.
Maybe because you are sick to death of dealing with version skew? This was a major reason why we chose RHEL for our servers at work, despite the fact that many of us here have lots of Debian, Gentoo and other UNIX experience (BSD, DECUNIX, etc)-- and arguably, we'd save money by going with them. But we don't save time or frustration. I run Ubuntu on my home computer because I just couldn't handle one more libc version bump without losing my mind. Ubuntu's Add/Remove isn't quite fire-and-forget, but it's pretty good.
I am capable of picking through system internals. Point is, I don't want to anymore.
It is also important to note that on {many | most} SAN-connected RAID enclosures, there's actually a battery backup unit that writes pending transactions to disk before the unit switches itself off due to power loss. Now, this doesn't help you when one of the SAN clients starts blatting out garbage, but assuming your clients are connected to UPS devices, that shouldn't happen.
Actually, UPS devices are useful for other kinds of things as well. Need to distribute load more evenly across your circuits? If you have the machine plugged into a UPS, you simply unplug the UPS and plug it into the other circuit. Heck, you could even do something really dumb like physically move the machine while it's running if you had it connected to a UPS.
My thoughts exactly-- make it fun! This is what got me into programming when I was a kid. HTML, being purely declarative, is a nice introduction to the idea of putting together a structured document, and the results are immediate. You can have him work with the W3C validator open, and it will help him find his mistakes. Then, when he's comfortable, add a real programming language. I recommend something like PHP, which is simple and relatively uncomplicated. Javascript is rather unforgiving when it comes to things like infinite loops, and (no offense to hardcore Javascript people out there) the language is pretty weird (my curse-to-code ratio is pretty high here).
When I was a kid, I learned BASIC and then moved on to C, but it was a HUGE leap to doing GUI programming for me, and I didn't really grasp that part until after college. Web technologies are, for the most part, quite simple, and people who know them well are in high demand. So unlike learning a minilanguage like the kind that comes with RoboWar, which is fun, but not very useful, web languages are a lifelong skill.
The cool part is that you're a programmer, so if he has questions, he can ask an expert. I did not have access to this resource as a kid-- it made learning the stuff much harder. As a physicist, my father had some exposure to APL and Fortran, but he wasn't very good at translating that experience to C.
Also-- I don't know if he's had exposure to algebra and geometry yet, but make sure he learns these-- they are essential to understanding the operation of modern applications. I haven't come across any higher math yet that hasn't benefitted my programming skills in some way (set theory probably being the #1 most useful subject to study).
Most importantly, make sure you are excited. If he sees how cool you think it is, he'll be extra motivated to push through the hard parts, in part because he'll want to please you, but also, because he'll want to see what all the fuss is about. Programming is one of the more fulfilling aspects of my life, so obviously, conjuring up the enthusiasm for me is very easy.
IANAL, but my impression was that, in order to be a criminal, you have to commit a crime. You're allowed to say no to your employer. And your employer is allowed to fire you for refusing to do the job they pay you for. If that's all that's going on here, well, then shame on the city for turning this into a spectacle.
Now, it's a different story if he tampered with the city's computer system to cause harm. But it's not at all clear from the stories if that's the case. My issues with the Reiser discussion were the same: we weren't getting the whole picture from the press, even though in the end there it turned out that Reiser really was a bad guy.
Slashdot Mirror
Well I didn't expect a kind of French Inquisition.
You're right, the OP misses an important component: culture. The practices and traditions that shape a culture will affect how those people respond to a particular governmental framework.
For instance, compare the 18th century American revolution with the French one. Both produced roughly equivalent forms of government, but the French revolution was far more brutal, probably the result of people having long been oppressed and ruled by sometimes-incompetent monarchs. The American colonists, by contrast, had several generations to forget about why they left England, and the American revolution was much more subdued. The rhetoric used to justify each conflict was far more philosophical on the American side: it was an attempt to correct the failings of foreign governance. French rhetoric saw their fight as nothing less than emancipation of the people.
As a result, French society traditionally put more of an emphasis on government-in-service-of-the-people than American society, who see government as more of a protectorate of the state. The United States has had invasive intelligence-gathering for a long time, and I think the reason why most Americans don't care is because they still see the state as a benevolent entity. I suspect that the French do not share this opinion, especially in the case of Sarcozy.
It still doesn't matter. It's speculated that the finest resolution a spy satellite can get is in the 5-10cm range, and that's probably using many digital imaging tricks. TFA doesn't say what kind of resolution the software requires, but I doubt that 5-10cm is fine enough-- when I walk, there's probably, what, 2-3cm of bounce in my step?
Still, this is clever idea. Attention conspiracy theorists: make sure to walk outside only at noontime. At the equator.
I consider myself a pretty typical geek, but I have no idea what that means. It's not ringing any bells with the Dijkstra that I'm familiar with, and in recursive-geek-irony, googling "Dijkstra's Scorpio" yields your very post. Someone care to enlighten me?
It depends on your requirements. If you're going to be at full cooling load all of the time, it's better to go with a traditional unit. That's what we have, as our unit was chosen based on the expected heat load for the room (6 tons). But if your expected heat load is not constant, you're right, it would probably be better to go with an inverter-based unit. For our heat load, an inverter unit would have been cost-prohibitive and also unnecessary.
Not to mention-- adding another AC unit without that unit being tied into the same controller will cause problems for your AC units. They'll be turning on and shutting off more frequently. This will greatly affect the life of your AC units. If you're a hardware hacker, you can probably add capacity on the cheap by hacking your thermostat to coordinate multiple units. Otherwise, you really do just need to pony up and pay for the AC upgrade. Cool air is a basic business expense nowadays. If the people writing your budget don't see it that way, then your company is in trouble.
I am fairly sure it was a joke.
Well, as long as you're laughing, does it matter?
Why, are China and India doing basic science research? My impression that pretty much *everyone* is getting out of the game. Deregulating telecom and breaking up AT&T did wonders for telephone customers, but it did not do good things for smart people with big budgets. Consider the fact that UNIX started as an excuse to hack on computer games.
But there's loads of software that runs on x86. Developer time costs magnitudes more than equipment, and for many places, my workplace included, these things are just great. If you're lucky enough to have software that's easily portable, and you don't mind dealing with the inevitable byte-order issues, character encoding issues, and other problems, by all means, use specialized machines. x86 is just so damned flexible that it's worth the extra cost.
We use Soekris [soekris.com] boxes all over the place because their power requirements are minimal, they can withstand harsh environments, they don't need cooling, they run x86 software, they're designed to run headless, and they don't take up much space. So even though you could buy a much more capable (speed- and expansion-wise) Wal-Mart PC for about the same price, the Soekris boxes still win out.
Plus, they're just neat.
Yeah, that's why I speculated that SQL might be done so easily-- Oracle really is a rabbit hole. I've done some relational algebra in a database course (and also was exposed to set theory in my discrete maths course), but it was unclear to me whether query optimizers actually broke a query down into relational algebra or not. In fact, I remember that despite having had prior experience with SQL, relational algebra was much easier for me to wrap my head around than SQL. My professor was hesitant to go too much into optimizers, though, since most of that is implementation-dependent. He thought it was much more important to talk about ACID, so that's what we spent a good deal of time on.
Actually, the two are paired: programming model and implementation. The reason there's a programming model is that functional methods allow Google's implementation to automatically parallelize the input data for feeding to the cluster. So the implementation is very important, because that's actually how the data is processed and returned.
In that sense, Oracle's clustering optimizations are also a paired programming model and implementation, since, presumably, you need to know Oracle's SQL language extensions in order to take advantage of them (disclaimer: I don't use Oracle). From what I understand about functional programming, SQL should be ideally positioned to take advantage of these kinds of optimizations, since the actual implementation details of any SQL query are always left to the query optimizer, SQL being a declarative language. I'm going to speculate wildly and say that you could probably write a SQL interpreter using a functional style as well, and that good ones probably already do.
He's not saying they're the same. He's saying that they have their roots in the same idea. That idea is, as far as is practical, the elimination of a ruling class and the fullest actualization of human freedom. How they differ is in what extent they consider to be practical. Anarchists are the ultimate capitalists: let natural human behavior self-regulate, i.e., there is no government (or, at least, no compulsory government). Communists believe that compulsory government is still necessary, but should only exist insofar as it helps to bring about the communist ideal of freedom.
You're focusing too much on labels here: a "political philosophy" thus implemented becomes a socioeconomic structure. China calls its political philosophy "communism", but anyone can see that China's socioeconomic structure is anything but. Calling itself "communist" is simply political theatre.
Hmm, interesting. It looks like MSIE 7 does do privilege separation, along with a couple other protections. I stand corrected. In my defense, I do not have Vista, so I was unaware of this.
All in all I think the web stack is pretty secure by default, when comparing it to the alternatives.
Really? My opinion is that the "web stack" (not sure which stack you mean here; MSIE-Windows, FF-Windows, Safari-MacOSX, Konq-Linux, etc) has by far the worst record so far. MSIE-Windows has to be the #1 vector for infection now, and has been for at least the last 6-7 years. Which alternative are you thinking of? Because the "web stack" is, in my opinion, the premier virus runtime environment.
My opinion is that web designers made a HUGE mistake in not treating network input cautiously. The emphasis has been on "rich APIs", "data structure passing", extensibility, desktop integration, and so on. These are undoubtably good things in the absence of malicious input, but the fact is, there is a lot of malicious input out there. Web browsers would benefit greatly from some simple privilege separation; the Mozilla camp could do this with some effort, but MSIE is pretty much dead in the water here due to the level of integration with the base system. I understand the HTML5 camp's worry that Flash/Flex will become a de facto standard, but in my opinion, web security has not been taken seriously enough. These kinds of vulnerabilities have become a major source of income for organized crime in the East, and still people like you are saying that security is not the most important issue? Gimme a break.
I use something similar: OpenBSD's spamd. The caveat is that, rarely, legitimate mail is rejected. At work, typically the person on the other end will contact us some other way, and we'll whitelist them. Yahoo is the one egregious offender here; their relays never retry from the same box, so spamd thinks that it's a new connection and starts the connection count over again. But I've found, in general, that the false positive rate with spamd alone is lower than the false positive rate with SpamAssassin alone, and that OpenBSD's spamd is far more effective in general. The only "spam" we see now is the stuff that users signed up for; mailing lists, mostly. Stuff that-- back when we were using Bayesian filters-- users would classify as spam and foul up the the spam pool for the Bayes engine.
I hate to break it to you, but LDAP is not a directory system. It is a directory protocol. AD provides an LDAP interface. So your directory system can be structured and provide storage in the backend pretty much any way you want. Microsoft, for instance, uses Jet for storing their data, and X.500 for structuring it. But if you wanted to build your directory using post-it notes and robot, then fine, as long as you provide an LDAP interface, you're an LDAP directory.
AD *can* store any arbitrary information with schema additions. So if you can query LDAP on the Linux side for window manager policy, and you can come up with a schema that represents that policy, go ahead, store it in AD. Mac people have been doing this for years, although Apple would prefer that you use their Open Directory system.
Also-- AD uses Kerberos. How do I know? Because I have Linux machines (MIT Kerb), OpenBSD machines (Heimdal), and Macs (MIT/Apple Kerb) all authenticating against our AD. There are some little oddities here are there (your machines have to support Microsoft's cipher-- which I believe is now installed by default on all recent Kerberos distributions), but in general, it works surprisingly well. For me, on Linux machines, the trick was learning the ins and outs of PAM and winbind. After that, it was easy.
Anyway, if you're expecting LDAP to provide authentication, you're mistaken about the purpose of LDAP. Think of it as a fancy phone book. What you need are a lock and key. Also-- accounting? For that, you want a piece of logging software. Microsoft supplies all of these things neatly packaged together, and if you don't want to bother with the details, then it's a decent choice. But don't confuse the two, because LDAP only provides a subset of the services that AD does. Complaining that LDAP does a "shit job" at authentication and accounting is like complaining that your tires do a "shit job" of steering. Well, duh.
Ugh. Text-free UIs are a waste of time.
There's a glyph. You don't know what it means. You might not want to activate the control it is connected to-- what if it means "delete my document?" So how do you find out what it means? You can't look it up, because it's a picture. The best you can do is describe it, in words, to another person, who will give you an answer, in words. So you see, language is the lowest common denominator here. Transcription of language, therefore, should be the thing we strive for-- oh, wait, we have that: the written language. Why use a computer if you can't read? Because, come on, most of the Internet is useless unless you are literate. It's a machine for processing the semantic and logical relationships of text.
You can't tell me that this symbol is in any way intuitive. I remember the first time when I saw it, sometime in the 90's. I remember thinking "what the fuck does that mean?" Sure, have some fancy symbol, but then, maybe would it kill you to print POWER underneath it? I think language is FAR more expressive than an abstract picture. Once that picture has been internalized by everyone, it is language. Why make it harder for everybody when we already have symbols that describe what you want to say?
Wait, you guys invented the smashing guitars trick. As is our ilk, we just capitalized on it.
I've noticed this discrepancy on my Jeep as well. Out of curiosity, do you have a citation for this?
Apple needs to get their shit together. Unless your needs are VERY straightforward, even 10.5 does not solve them. I'll admit that 10.5 has a much nicer server admin GUI, but it does not overcome the problems with the platform.
We've moved all of these services to CentOS machines. By contrast, getting them working reliably was a walk in the park. Equivalent hardware (hotswap RAID (SCSI, I should add), redundant PSU, fiber channel card, GigE, dual processor machines in a 3U form factor (SuperMicro chassis) come out to about $1k less than an Xserve, on average. And when a part dies, like a backplane, I can BUY THAT PART. With Apple, you have to buy an entire parts kit, which comes with stuff you may not want.
We now run Samba and Netatalk on CentOS on generic server hardware, connected to our StorNext network. There may be better SAN stuff out there than StorNext (in fact, their licensing department leaves much to be desired-- do they even know how to use their own product?), but we already had a lot invested (three Xserve RAID cabinets). Things run great now, and with the Linux version of BRU, our full tape backup [inexplicably] finishes 9 hours earlier (used to take 60 hours, now takes 51).
My advice: Apple makes some nice desktops, but their server stuff is only for novices. I went into the experience very optimistic about Apple's stuff, but now I have a very bitter taste in my mouth.
Maybe because you are sick to death of dealing with version skew? This was a major reason why we chose RHEL for our servers at work, despite the fact that many of us here have lots of Debian, Gentoo and other UNIX experience (BSD, DECUNIX, etc)-- and arguably, we'd save money by going with them. But we don't save time or frustration. I run Ubuntu on my home computer because I just couldn't handle one more libc version bump without losing my mind. Ubuntu's Add/Remove isn't quite fire-and-forget, but it's pretty good.
I am capable of picking through system internals. Point is, I don't want to anymore.
It is also important to note that on {many | most} SAN-connected RAID enclosures, there's actually a battery backup unit that writes pending transactions to disk before the unit switches itself off due to power loss. Now, this doesn't help you when one of the SAN clients starts blatting out garbage, but assuming your clients are connected to UPS devices, that shouldn't happen.
Actually, UPS devices are useful for other kinds of things as well. Need to distribute load more evenly across your circuits? If you have the machine plugged into a UPS, you simply unplug the UPS and plug it into the other circuit. Heck, you could even do something really dumb like physically move the machine while it's running if you had it connected to a UPS.
My thoughts exactly-- make it fun! This is what got me into programming when I was a kid. HTML, being purely declarative, is a nice introduction to the idea of putting together a structured document, and the results are immediate. You can have him work with the W3C validator open, and it will help him find his mistakes. Then, when he's comfortable, add a real programming language. I recommend something like PHP, which is simple and relatively uncomplicated. Javascript is rather unforgiving when it comes to things like infinite loops, and (no offense to hardcore Javascript people out there) the language is pretty weird (my curse-to-code ratio is pretty high here).
When I was a kid, I learned BASIC and then moved on to C, but it was a HUGE leap to doing GUI programming for me, and I didn't really grasp that part until after college. Web technologies are, for the most part, quite simple, and people who know them well are in high demand. So unlike learning a minilanguage like the kind that comes with RoboWar, which is fun, but not very useful, web languages are a lifelong skill.
The cool part is that you're a programmer, so if he has questions, he can ask an expert. I did not have access to this resource as a kid-- it made learning the stuff much harder. As a physicist, my father had some exposure to APL and Fortran, but he wasn't very good at translating that experience to C.
Also-- I don't know if he's had exposure to algebra and geometry yet, but make sure he learns these-- they are essential to understanding the operation of modern applications. I haven't come across any higher math yet that hasn't benefitted my programming skills in some way (set theory probably being the #1 most useful subject to study).
Most importantly, make sure you are excited. If he sees how cool you think it is, he'll be extra motivated to push through the hard parts, in part because he'll want to please you, but also, because he'll want to see what all the fuss is about. Programming is one of the more fulfilling aspects of my life, so obviously, conjuring up the enthusiasm for me is very easy.
IANAL, but my impression was that, in order to be a criminal, you have to commit a crime. You're allowed to say no to your employer. And your employer is allowed to fire you for refusing to do the job they pay you for. If that's all that's going on here, well, then shame on the city for turning this into a spectacle.
Now, it's a different story if he tampered with the city's computer system to cause harm. But it's not at all clear from the stories if that's the case. My issues with the Reiser discussion were the same: we weren't getting the whole picture from the press, even though in the end there it turned out that Reiser really was a bad guy.