"I looked at Google and didn't get anything blindingly obvious. What are you referring to?"
One of the mid east cosmologies (see Zohar) estates that the world is sustained upon seven pillars (you know the number seven holds a mistic meaning); Jews took this and expanded it stating that at any given moment there will be at least seven (hidden) just men alive that will "sustain the world" at the eyes of God in that He won't destroy it because of their mere existance (they alone compensate for all the rest of us, greedy bastards). That way, "the seven just men" becomes an idiom about an "entity" of honesty beyond doubt... if only we could recognice them.
But I must admit I probably overstated how easy is to find is the original reference. I did a search on Google and while there are some entries about "seven just men" they implicitly expect you already know what this does mean.
Somehow this still makes for my original point in that it's impossible to reach a consensus about what's needed in order to consider somebody "good enough" to vote and who is the one to tell it: you are not cult enough in my account since you didn't know about "the seven just men" (and it's a relevant knowledge since the fact they are hidden is what makes them useless for any practical meaning -it's, of course, a poetic license, I don't literally mean we should look for them, but that there's no way we can be sure beyond doubt about the honesty and good sense of anybody to allow him to become our benevolent tirant, even if such a person exists to begin with -Ancient Greece and Rome tried this with less than perfect results). On the other hand, it's me the one unfitted on your regard... who is to judge who is right and who is wrong? The only solution it's admitting there's no one, so the "one person one vote" should stand.
"1 - did i say what the salary was? No, i did not. Perhaps the salary is high enough to compensate for the after hours work. that thought ever cross your mind? But i refuse to pay a person extra to get them to do their assigned duties."
Did you missed the point where I said "my working hours are 8 to 5, it's right there, on my contract"? The negotiated salary no matter how high or low is based on the written down expectation that it means working 8 to 5. I'm open to negotiate almost any contract conditions and find the agreament point where both parties find it fair enough, but I don't ever found fair sign for something and then being asked to do something else. On our hipothical case, it certainly could be exposed by the time of hiring that working after hours was expected, how much of it was expected and how would that job be compensated. In fact, since it's quite a usual expectation I wouldn't (and don't) sign a contract without such issues being risen to the table.
"Ever hear of time management? If you are going to spend 2 hours this weekend on upgrades, leave early an hour 2 days during the week, or add it to your vacation time.."
Did you missed the part where I already stated being open for a negotiation to find an agreement point? On that regard if you find the value of an hour on the 8-5 window to be the same than off bussiness hours, good for your manager, stupid for you and bad for me since as I already told on a different message, it's people like you the ones that rise stupid expectations on managers. Remember it next time you have to call the plumber or the locksmith off-hours and find they stubbornly insist on asking for a premium rate for their services.
"But no, you are too wrapped up in the 'me me me' attitude."
What I won't do is accepting being abused by a manager that thinks to be "so clever" that manages to get 50 hours out of a 40 hours resource (me).
""No responsible admin would arbitrarily inconvenience their users because they are too damned lazy to perform normal maintenance after hours" is that better?"
Yes, that's better, but that's not my position. On one hand, yes, I'm "too lazy" to perform normal maintenance after hours *for free* -indeed, I'm "too lazy" to perform *anything* for free for my employer -curiously enough, my employer seems to be "too lazy" too to give me nothing for free either: I don't get non-negotiated bonuses just because, nor I'm payed if one morning I tell them "I won't go to work today: look, it's so pretty a day I just feel better going to a park and lay down on the grass; on the other, I never said I would "arbitrarily inconvenience" our systems users: I already offered two examples of negotation about the best moment for such maintenances so the one with authority and responsibility about it would be able to take the decision that makes the best bussiness sense: it can be by night, by day, at dinner time, as soon as I see fit, on fixed windows, notizing by triplicate one month before... anyone of them has their own advantages, inconveniencies and distinct costs, it's only I won't artificially make some options cheaper by accepting to do them for free.
"With that attitude of ' you will pay me extra to do my job ' you would never work for me"
Good to know. With that attitude of 'I won't pay you extra to do extra work' I'll never work for you either.
"People like you give the rest of us a bad name."
So, let's see: are you one that will hire people to do extra work for nothing or are you the one hired to do extra work for nothing? If the former, I already stated my position, if the latter, it's people like you the one I meant when I said "there are too many stupid admins that think it's so fun being abused and working for free" so you are directly hurting me by rasing among managers the expectation that they can abuse IT people since they are willing to do extra work for free.
"Your right to vote has been revoked for failing the civics test. There are NINE supreme court justices."
Your right to vote has been revoked for failing the minimal cultural test: "the seven just men" is an idiom not the number of judges on USA Supreme Court (did you notice I put it in quotes?). Search on Google for further information.
"No kidding, what responsible admin would take systems down during the day except for emergencies?"
Why not? Is it that day temperatures make equipment unbootable while night freshness makes everything easier or what?
I can shutdown a server by day or by night, I find no operative differences in doing it one way or the other. But my working hours are 8 to 5, it's right there, on my contract, so that's when I'll do my job including maintenance shutdowns. If they want it off bussiness hours, well, no problem, just pay me extra for that extra hours. And if they don't want neither maintenance shutdowns on bussiness hours nor pay extra for off-hours, well, no problem either: it's just that there won't be maintenance shutdowns. Risk mitigation and bussiness value calculations is my manager's problem, not mine.
Now, what probably you meant to ask was something like this: "No kidding, what responsible manager wouldn't pay after hour rates to a single person to shut down a system by night instead of having fifty people circling thumbs if the server is shut down during bussiness hours either for maintenance or due to a failure for lack of maintenance?"
Well, the answer to that question is "Just too many, d'you know why? Because there are too many stupid admins that think it's so fun being abused and working for free".
"My idea from my previous post, which is that if even one citizen can objectively rigorously prove that a law has failed, that should cause the repeal of the law. I like this idea because it makes it possible to remove the politics from the equation. If the facts are clearly and demonstrably on your side..."
Well, good luck with it. Who will be the one that will tell when "the facts are clearly and demonstrably on someone's side"? Even for non-political somehow "clearcut" issues like penal justice (is this guy a murderer or is he not?) "the system" needs for a complex hierarchy of people that even on its top (the Supreme Court) can't be reduced to a single man. And even those "seven just men" opinions' manage usually to diverge.
"My other solution is that no one should be allowed to vote unless they can pass an extremely tough civics test."
So your other idea is abolishing universal sufrage so only "good citizens" can vote. You already stated why you don't vote: who do you think will be the ones to say who the good citizens are? You failed twice -with unexcusable naivety, the "first mandament for a civic society": never ever think about a law based on the good that could be achieved from it but based on how it can be abused, for certainly it will be.
Luckily your ideas are far from even being far of being deployed but if they were to, by your own account, you would never be allowed to vote in my book and that's the very problem of any society not based on universal sufrage: the valuable ones of somebody are always the morons of another. Would you really accept your incompetence and would renounce your right to vote based on my opinion or would you say that I'm no one to tell who has the right to vote, specially when I put you in the list of those who have it not?
"If you melted it into liquid and stirred it around you wouldn't have to do any over-writing."
That's security by overlying.
Each of those steps (specially the last ones) are by themselves more than enough for the data being unrecoverable... IF properly done. Then you have deep security by overlying them. See:
Overwriting the disks with random data is probably good enough... unless the software somehow fails in an unnoticed manner (it doesn't write anything, for instance).
But then, you degauss the platters wich, again, is enough by itself, in case the software overwriting did in fact failed. But what if GI Joe took a door stopper instead of the degausser? You know, nowadays, GI Joe troopers are not PhD's all of them.
But then, you take the still entire HD and take *off the premises* to physical destroy it. What if the HD gets lost on its trip (I knew I shoulda taken that left turn at Albuquerque)? You'd better have the disks free of secret data by that time.
And then, what if the unforeseen happens and the people in charge of the physical destruction is bought by the enemy? You'd better chain at least two unrelationed means to destroy it (like, first pass a tank over it and then send it to a smelter) and attach them by a paper track so if step 1 tries to send it to the enemy (on the slim chance that rewriting and degaussing failed), then step 2 will ask were the advertised HD were.
You know, plan as a pesimist, execute as an optimist.
"This process is very common all across Southern Europe, and works quite well to disinfect the product."
Nope.
Trichinosis is a parasitic illness produced by eating "eggs" (I don't know the proper English word) from 'Trichinella spiralis'. Only proper cooking will kill them. Cured ham (Spanish jamón, Itailan prosciuto, etc.) is a known vector for this illness.
Curing meat is *not* a disinfectation process but a conservation one... and it conserves Trichinium "eggs" just as good as it preserves pig's flesh.
OR you will get net jobs. On this regard there's no magic on government "borrowing" from taxes vs a big corporation borrowing from private investors via saving funds. What makes the difference is the way such money is used and it can be badly used both ways (see the current financial crisis for a glaring example of badly invested money).
On a side note, while I share the concerns about government managing money (they are unefficient because they are big and they are not risking their own money, etc.), I find a bit naive not seeing this is not a problem with the government but a problem with each and every big organization, either public or private: a transnational CEO's doesn't risk his money any more than a public employee; in fact, the CEO has now much more weapons on his toolset than the government employee to turn his malpractices against his company and the society as a whole into personal benefits, so all things being equal we should expect such malpractices coming more usually from private companies than government and, in fact, I'd say that's exactly the case and it makes no argument saying that, well, that would mean the company will go bankrupt and it'll give space for new opportunities: we've seen not only that those malpracticers got their bonuses but the companies are not being displaced by "better" ones; the companies will be "saved" anyway or companies of the same kind, with even the same executives on board will take the place reducing even more overall rivalry.
"May be IT investment creates more jobs than physical infrastructure investment. But if the jobs are created in India, are we (Americans) really better off?"
Yes, even then, provided they are cheaper and get the job done: they allow for new opportunities backed up from the saved costs. It's up to you to take advantage of them or not.
I'm aware of the differences (or lack of them) between inertial and gravitational mass and yes, you are right, to my knowledge we still don't know why they (seem) to be the same. But this was "just" about mass vs weight. Weight is related to the gravitational mass but it is *not* neither gravitational mass nor a measure of it. For all I know, a massive body will have by itself both gravitational and inertial mass, but it won't have weight by itself. When I say this object weights 70 pounds, what I'm really saying is that the gravitational interaction between that object and Earth is measured as 70 pounds. Both, one against the other is what values 70 pounds. With just one of them, you won't have weight; you still will have mass though. While your question is a very interesting one (probably it is *the* interesting one, since the original was mostly semantic) is a different one (though related).
"In other words, can there exist a massive object having no gravitational effect"
But the point is that that's not weight! Weight is not the gravitational field of an object (or the space-time deformation it induces for that matter) but how *a second object* interacts with it due to the fact they both are massive. For a massive object to be to all known effects weightless (but not massless) it suffices being the only object in Universe.
"So one can still ask the question, and many do, as to whether weight and mass are really measures of two different things, or simply two related variables which measure the same underlying property of a piece of matter."
Unless Match got it right and Einstein wrong, mass and weight can't be "different measures of a single underlying propiety". Just pay attention to this "mental experiment": please devise a (theoretical) mean to weigh the whole Universe. You can't. Mass is an intrinsic propiety of matter while weight is an observable of matter interaction. You need at least two masses attracting each other (by means of their mass) to gain weight so one "weighs" against the other (you against the Earth, for instance).
"I've got a couple problems: 1) How do you deal with an HTML document that links to something outside the filesystem? As in, what if I'm offline?"
No worse than currently (you see, man pages do point to external resources too, didn't you know?). But what if (gasp) you happen to be online? Much better than now.
"2) You can do a hell of a lot in HTML that you cannot render on a console. What if the document you wrote uses an image or some javascript?"
Don't do that. The fact that you *can* use this or that doesn't mean you *must* use it for man-like documentation.
"Whatever you propose must exist when I type "man joeblow"."
I think you have a case here. It doesn't seem too difficult to do though, and even probably it would be better the way you describe, since man-like html pages would still not be "just any HTML page you happen to write" but it would expect just plain HTML (maybe using just a subset of tags, or/and maybe with standard CSS overload for semantics), on standard locations, etc. As a first approach, man could be a wrapper that would look at the man directories, much like now, but if a whatever.5.html does exists it would launch a links-like tool instead of the default man pager.
All in all, in my opinion man pages are still roff instead of html because man pages are "just good enough" and a) info came before the HTML explosion and b) GNU people are a bit stubborn sometimes (given the less than massive acceptance of info they probably would better start anew with an html-based version or resign and return to man).
It is not. You can repeat it one thousand times and it won't become true. Some "things" depending on other "things" means not "broken by design", it is called "coperativeness".
"Reliable systems are built to withstand problems."
Yes. But a database "system" is not a "system" (the system is the computer and network hardware, the operative system, the user space tools, the database manager, the data *and* the software accessing such data). Unless, of course, you want to consider the database a "system" by itself in which case, it should be built to withstand *database* problems, not any other kind of problems. Both the whole system and the database system can be built to withstand their scope-related problems so your point is... you don't have a point.
"A database which crashes hard because of a kernel bug or a hardware failure, or (gasp!) an inopportune snapshot is a broken database."
It is a broken database no more than a computer is broken if it stops working whenever electricity stops passing through: electric-dependability is an out-of-scope problem regarding stand-alone hardware systems; that's why computers do work *provided* electricity is there (note you still can build *whole* systems that can withstand increasingly severe restrictions regarding electric disposibility); operative-system stability is an out-of-scope problem regarding *any* user space application, not only RDBMs: if the underlying OS fails you can't and shouldn't expect *any* user-space executable to be reliable; you still can build *whole* systems under given severity constraints and -as I already said, one of you most basic considerations will be knowing what you can and should expect from any given subsystem so you don't ask "A" to accomplish tasks under "B's" responsibilities.
"I accept that most or all databases are broken in this fashion"
Taking into account the ammount of talent and money devoted to the database management bussiness, I'll ask you to consider the slim chance that it might be you, not them, the one in the wrong position.
"In that case, please allow me to submit that the database is broken."
No, it isn't.
"Snapshots are a frozen moment in time."
Sorry but no, they aren't. Snapshots are *filesystems* frozen in a moment of time. Nothing more, nothing less. They explicitly are not RAM frozen in time and they are explicitly not done at a database-safe moment of time.
"Failing, under any circumstances, to be able to recover a working database from a snapshot, is the same as failing to recover a working database after hardware or power failure or an obscure kernel bug crashes the system."
That's right, so what? Nobody is forcing you to have database snapshots if not needed and there's no way that your database backlog wil be sincronized to your snapshot policies and procedures unless a sysadmin has worked for them to be that way.
"Things should not be so fragile."
They are, and you better work under reality checks, not based on what they "should" be.
Of course, you can have whole *systems* to be such and such reliable but in order to achieve that you need to "tell" each subsystem to do their own homework: the database-thing to take care of their own "things" (i.e.: transaction-based checkpoints); the RAM-thing to take care of their own "things" (like ECC parity); the persistance-thing to take care of their own "things" (like LVM, RAID, system-level backups, etc.)...
Once you learn which "things" are part of what "things", you can design and deploy a reliable system; failing that you will have people actually thinking that RAID is a form of backup; that filesystem-level copies can allow you to forget about how to properly secure your database, etc.
"actually yes, you can do just that - tell the database that the filesystem level 'thing' needs to get a coherent snapshot of the database level 'thing'"
Thank you for actually reinforce my point: in order to do "a database thing" (a backup) you "tell the database..." (your own words).
"Can't these problems generally solved with snapshots (in FreeBSD and Linux LVM lingo) or shadow copies (in Windows parlance)?"
Short answer: no. Long answer: filesystem level "things" don't know about database "things" so there may be "things" at the database level that lead to a non-functional database when coped with at the filesystem level.
This assumption is valid not only when talking about filesystems and databases but on every two unrelated items. Learn this and you probably know half you need to be a proper IT professional.
"The $hardware$ $programmer-time$ equation is always based on the assumption that the programmer is always worth their qualifications."
The $hardware$ $programmer-time$ equation is always based on the *stupid* assumption that the programmer is always worth their qualifications. There, corrected for you.
As I read the article, this all about a programming problem, not a running environment. While your post is quite sensible, real environments are not. My reality check is that when on a dev mill the question is risen about throwing iron or minds to a problem the answer is *always* "minds". Think about it for a moment: it is either stated as "this is CPU|I/O|whatever problem we have, our experts can|can't cope with it and it would cost $$$ deal with the problem with new hardware for so much time at current growth rate, or we can|hope to cope with it if Joe and Anne work on it somewhere between 100 and 200 hours each" and then there's no margin for even begging the question; the answer is an almost immediate output from the analysis... Or it is risen like "we don't have the slightest idea why this runs so slow, but we *hope* we can deal with it with bigger iron, what should we do?"
Even when a sensible manager faced at the second situation thinks like "this is obviously a 'more mind' -well, more like 'better minds' situation, but we don't have the expertise here, nor the time and/or the ability to find it elsewhere, so let's go for the more iron path and hope for the better", things go no better. The problem is the development team tried to bite bigger than their mounths and the more iron you throw to the problem, the more stressed the lacking codebase becomes and the more it falls appart.
I am not telling that "all things being equal", throwing more iron is not a solution -in fact it is, but that most of the time, the "more iron or more people" dichotomy it's only "we don't know any better" in disguise, which is, in fact, "we don't know good enough to make it work". You can find a "more iron" situation rarely and on short periods if your app changes niche (our nice app designed to work on a single server does indeed so nice that some fortune-500 want it so it's going from 10tps to 300tps. What can we do to fullfill the situation for some months while we rethink it?), but I'd say 99% when an app is not outspec'd but still seems to require more iron than initially planned it is because a design fault, and design faults on anything more complex than a "hello world" are never so easy to fix as changing just one environment variable: they tend to fail on a lot of fronts and you are not going to cope with the situation just "throwing more iron to it".
Slashdot Mirror
"I looked at Google and didn't get anything blindingly obvious. What are you referring to?"
One of the mid east cosmologies (see Zohar) estates that the world is sustained upon seven pillars (you know the number seven holds a mistic meaning); Jews took this and expanded it stating that at any given moment there will be at least seven (hidden) just men alive that will "sustain the world" at the eyes of God in that He won't destroy it because of their mere existance (they alone compensate for all the rest of us, greedy bastards). That way, "the seven just men" becomes an idiom about an "entity" of honesty beyond doubt... if only we could recognice them.
But I must admit I probably overstated how easy is to find is the original reference. I did a search on Google and while there are some entries about "seven just men" they implicitly expect you already know what this does mean.
Somehow this still makes for my original point in that it's impossible to reach a consensus about what's needed in order to consider somebody "good enough" to vote and who is the one to tell it: you are not cult enough in my account since you didn't know about "the seven just men" (and it's a relevant knowledge since the fact they are hidden is what makes them useless for any practical meaning -it's, of course, a poetic license, I don't literally mean we should look for them, but that there's no way we can be sure beyond doubt about the honesty and good sense of anybody to allow him to become our benevolent tirant, even if such a person exists to begin with -Ancient Greece and Rome tried this with less than perfect results). On the other hand, it's me the one unfitted on your regard... who is to judge who is right and who is wrong? The only solution it's admitting there's no one, so the "one person one vote" should stand.
"1 - did i say what the salary was? No, i did not. Perhaps the salary is high enough to compensate for the after hours work. that thought ever cross your mind? But i refuse to pay a person extra to get them to do their assigned duties."
Did you missed the point where I said "my working hours are 8 to 5, it's right there, on my contract"? The negotiated salary no matter how high or low is based on the written down expectation that it means working 8 to 5. I'm open to negotiate almost any contract conditions and find the agreament point where both parties find it fair enough, but I don't ever found fair sign for something and then being asked to do something else. On our hipothical case, it certainly could be exposed by the time of hiring that working after hours was expected, how much of it was expected and how would that job be compensated. In fact, since it's quite a usual expectation I wouldn't (and don't) sign a contract without such issues being risen to the table.
"Ever hear of time management? If you are going to spend 2 hours this weekend on upgrades, leave early an hour 2 days during the week, or add it to your vacation time.."
Did you missed the part where I already stated being open for a negotiation to find an agreement point? On that regard if you find the value of an hour on the 8-5 window to be the same than off bussiness hours, good for your manager, stupid for you and bad for me since as I already told on a different message, it's people like you the ones that rise stupid expectations on managers. Remember it next time you have to call the plumber or the locksmith off-hours and find they stubbornly insist on asking for a premium rate for their services.
"But no, you are too wrapped up in the 'me me me' attitude."
What I won't do is accepting being abused by a manager that thinks to be "so clever" that manages to get 50 hours out of a 40 hours resource (me).
""No responsible admin would arbitrarily inconvenience their users because they are too damned lazy to perform normal maintenance after hours"
is that better?"
Yes, that's better, but that's not my position. On one hand, yes, I'm "too lazy" to perform normal maintenance after hours *for free* -indeed, I'm "too lazy" to perform *anything* for free for my employer -curiously enough, my employer seems to be "too lazy" too to give me nothing for free either: I don't get non-negotiated bonuses just because, nor I'm payed if one morning I tell them "I won't go to work today: look, it's so pretty a day I just feel better going to a park and lay down on the grass; on the other, I never said I would "arbitrarily inconvenience" our systems users: I already offered two examples of negotation about the best moment for such maintenances so the one with authority and responsibility about it would be able to take the decision that makes the best bussiness sense: it can be by night, by day, at dinner time, as soon as I see fit, on fixed windows, notizing by triplicate one month before... anyone of them has their own advantages, inconveniencies and distinct costs, it's only I won't artificially make some options cheaper by accepting to do them for free.
"With that attitude of ' you will pay me extra to do my job ' you would never work for me"
Good to know. With that attitude of 'I won't pay you extra to do extra work' I'll never work for you either.
"People like you give the rest of us a bad name."
So, let's see: are you one that will hire people to do extra work for nothing or are you the one hired to do extra work for nothing? If the former, I already stated my position, if the latter, it's people like you the one I meant when I said "there are too many stupid admins that think it's so fun being abused and working for free" so you are directly hurting me by rasing among managers the expectation that they can abuse IT people since they are willing to do extra work for free.
"Your right to vote has been revoked for failing the civics test. There are NINE supreme court justices."
Your right to vote has been revoked for failing the minimal cultural test: "the seven just men" is an idiom not the number of judges on USA Supreme Court (did you notice I put it in quotes?). Search on Google for further information.
"No kidding, what responsible admin would take systems down during the day except for emergencies?"
Why not? Is it that day temperatures make equipment unbootable while night freshness makes everything easier or what?
I can shutdown a server by day or by night, I find no operative differences in doing it one way or the other. But my working hours are 8 to 5, it's right there, on my contract, so that's when I'll do my job including maintenance shutdowns. If they want it off bussiness hours, well, no problem, just pay me extra for that extra hours. And if they don't want neither maintenance shutdowns on bussiness hours nor pay extra for off-hours, well, no problem either: it's just that there won't be maintenance shutdowns. Risk mitigation and bussiness value calculations is my manager's problem, not mine.
Now, what probably you meant to ask was something like this: "No kidding, what responsible manager wouldn't pay after hour rates to a single person to shut down a system by night instead of having fifty people circling thumbs if the server is shut down during bussiness hours either for maintenance or due to a failure for lack of maintenance?"
Well, the answer to that question is "Just too many, d'you know why? Because there are too many stupid admins that think it's so fun being abused and working for free".
"Or is it that they are they blocking ALL web access from the development environment which makes more sense than you Windows bashing post."
I'd bet is more that they are using a "windows only" proxy to get to the Internet.
"My idea from my previous post, which is that if even one citizen can objectively rigorously prove that a law has failed, that should cause the repeal of the law. I like this idea because it makes it possible to remove the politics from the equation. If the facts are clearly and demonstrably on your side..."
Well, good luck with it. Who will be the one that will tell when "the facts are clearly and demonstrably on someone's side"? Even for non-political somehow "clearcut" issues like penal justice (is this guy a murderer or is he not?) "the system" needs for a complex hierarchy of people that even on its top (the Supreme Court) can't be reduced to a single man. And even those "seven just men" opinions' manage usually to diverge.
"My other solution is that no one should be allowed to vote unless they can pass an extremely tough civics test."
So your other idea is abolishing universal sufrage so only "good citizens" can vote. You already stated why you don't vote: who do you think will be the ones to say who the good citizens are? You failed twice -with unexcusable naivety, the "first mandament for a civic society": never ever think about a law based on the good that could be achieved from it but based on how it can be abused, for certainly it will be.
Luckily your ideas are far from even being far of being deployed but if they were to, by your own account, you would never be allowed to vote in my book and that's the very problem of any society not based on universal sufrage: the valuable ones of somebody are always the morons of another. Would you really accept your incompetence and would renounce your right to vote based on my opinion or would you say that I'm no one to tell who has the right to vote, specially when I put you in the list of those who have it not?
"If you melted it into liquid and stirred it around you wouldn't have to do any over-writing."
That's security by overlying.
Each of those steps (specially the last ones) are by themselves more than enough for the data being unrecoverable... IF properly done. Then you have deep security by overlying them. See:
Overwriting the disks with random data is probably good enough... unless the software somehow fails in an unnoticed manner (it doesn't write anything, for instance).
But then, you degauss the platters wich, again, is enough by itself, in case the software overwriting did in fact failed. But what if GI Joe took a door stopper instead of the degausser? You know, nowadays, GI Joe troopers are not PhD's all of them.
But then, you take the still entire HD and take *off the premises* to physical destroy it. What if the HD gets lost on its trip (I knew I shoulda taken that left turn at Albuquerque)? You'd better have the disks free of secret data by that time.
And then, what if the unforeseen happens and the people in charge of the physical destruction is bought by the enemy? You'd better chain at least two unrelationed means to destroy it (like, first pass a tank over it and then send it to a smelter) and attach them by a paper track so if step 1 tries to send it to the enemy (on the slim chance that rewriting and degaussing failed), then step 2 will ask were the advertised HD were.
You know, plan as a pesimist, execute as an optimist.
"This is why you cure ham."
Nope.
"This process is very common all across Southern Europe, and works quite well to disinfect the product."
Nope.
Trichinosis is a parasitic illness produced by eating "eggs" (I don't know the proper English word) from 'Trichinella spiralis'. Only proper cooking will kill them. Cured ham (Spanish jamón, Itailan prosciuto, etc.) is a known vector for this illness.
Curing meat is *not* a disinfectation process but a conservation one... and it conserves Trichinium "eggs" just as good as it preserves pig's flesh.
"If you use Linux, you can install R with
yum install R*"
I do use Linux, let's see:
mybox:~# yum install R*
bash: yum: command not found
mybox:~#
Let's rewrite your sentence:
"On *some* variants of Linux (notably the SUSE family) you can install R with `yum install R*`"
There.
" it is akin to worrying about flying sharks into the middle of the sahara desert and worrying they will wipe out all of the local lifeforms."
Or rabbits (or toads) in Australia. Is that your point?
"You will get NO net jobs"
OR you will get net jobs. On this regard there's no magic on government "borrowing" from taxes vs a big corporation borrowing from private investors via saving funds. What makes the difference is the way such money is used and it can be badly used both ways (see the current financial crisis for a glaring example of badly invested money).
On a side note, while I share the concerns about government managing money (they are unefficient because they are big and they are not risking their own money, etc.), I find a bit naive not seeing this is not a problem with the government but a problem with each and every big organization, either public or private: a transnational CEO's doesn't risk his money any more than a public employee; in fact, the CEO has now much more weapons on his toolset than the government employee to turn his malpractices against his company and the society as a whole into personal benefits, so all things being equal we should expect such malpractices coming more usually from private companies than government and, in fact, I'd say that's exactly the case and it makes no argument saying that, well, that would mean the company will go bankrupt and it'll give space for new opportunities: we've seen not only that those malpracticers got their bonuses but the companies are not being displaced by "better" ones; the companies will be "saved" anyway or companies of the same kind, with even the same executives on board will take the place reducing even more overall rivalry.
"May be IT investment creates more jobs than physical infrastructure investment. But if the jobs are created in India, are we (Americans) really better off?"
Yes, even then, provided they are cheaper and get the job done: they allow for new opportunities backed up from the saved costs. It's up to you to take advantage of them or not.
"No, you're confusing the issue."
I'm aware of the differences (or lack of them) between inertial and gravitational mass and yes, you are right, to my knowledge we still don't know why they (seem) to be the same. But this was "just" about mass vs weight. Weight is related to the gravitational mass but it is *not* neither gravitational mass nor a measure of it. For all I know, a massive body will have by itself both gravitational and inertial mass, but it won't have weight by itself. When I say this object weights 70 pounds, what I'm really saying is that the gravitational interaction between that object and Earth is measured as 70 pounds. Both, one against the other is what values 70 pounds. With just one of them, you won't have weight; you still will have mass though. While your question is a very interesting one (probably it is *the* interesting one, since the original was mostly semantic) is a different one (though related).
"In other words, can there exist a massive object having no gravitational effect"
But the point is that that's not weight! Weight is not the gravitational field of an object (or the space-time deformation it induces for that matter) but how *a second object* interacts with it due to the fact they both are massive. For a massive object to be to all known effects weightless (but not massless) it suffices being the only object in Universe.
"So one can still ask the question, and many do, as to whether weight and mass are really measures of two different things, or simply two related variables which measure the same underlying property of a piece of matter."
Unless Match got it right and Einstein wrong, mass and weight can't be "different measures of a single underlying propiety". Just pay attention to this "mental experiment": please devise a (theoretical) mean to weigh the whole Universe. You can't. Mass is an intrinsic propiety of matter while weight is an observable of matter interaction. You need at least two masses attracting each other (by means of their mass) to gain weight so one "weighs" against the other (you against the Earth, for instance).
"The selfsacrifice is a known mechanism in cells. The tiny cell wont be able to conceptualize or understand its purpose."
Then, it is not "selfsacrifice" since there's not "self" to sacrifice.
"Yet we don't know ourselves, for certain, what we are part of in a larger scale."
We know we are not part of nothing large scale. As a matter of fact we know, full stop.
"Statistically speaking, you will die. "
Statistically speaking, no you won't: from Adam to-date, less than half of Humankind has ever died, and even those that did, just died once!
You know, there're lies, damn lies and statistics.
"I've got a couple problems:
1) How do you deal with an HTML document that links to something outside the filesystem? As in, what if I'm offline?"
No worse than currently (you see, man pages do point to external resources too, didn't you know?). But what if (gasp) you happen to be online? Much better than now.
"2) You can do a hell of a lot in HTML that you cannot render on a console. What if the document you wrote uses an image or some javascript?"
Don't do that. The fact that you *can* use this or that doesn't mean you *must* use it for man-like documentation.
"Whatever you propose must exist when I type "man joeblow"."
I think you have a case here. It doesn't seem too difficult to do though, and even probably it would be better the way you describe, since man-like html pages would still not be "just any HTML page you happen to write" but it would expect just plain HTML (maybe using just a subset of tags, or/and maybe with standard CSS overload for semantics), on standard locations, etc. As a first approach, man could be a wrapper that would look at the man directories, much like now, but if a whatever.5.html does exists it would launch a links-like tool instead of the default man pager.
All in all, in my opinion man pages are still roff instead of html because man pages are "just good enough" and a) info came before the HTML explosion and b) GNU people are a bit stubborn sometimes (given the less than massive acceptance of info they probably would better start anew with an html-based version or resign and return to man).
"It is broken."
It is not. You can repeat it one thousand times and it won't become true. Some "things" depending on other "things" means not "broken by design", it is called "coperativeness".
"Reliable systems are built to withstand problems."
Yes. But a database "system" is not a "system" (the system is the computer and network hardware, the operative system, the user space tools, the database manager, the data *and* the software accessing such data). Unless, of course, you want to consider the database a "system" by itself in which case, it should be built to withstand *database* problems, not any other kind of problems. Both the whole system and the database system can be built to withstand their scope-related problems so your point is... you don't have a point.
"A database which crashes hard because of a kernel bug or a hardware failure, or (gasp!) an inopportune snapshot is a broken database."
It is a broken database no more than a computer is broken if it stops working whenever electricity stops passing through: electric-dependability is an out-of-scope problem regarding stand-alone hardware systems; that's why computers do work *provided* electricity is there (note you still can build *whole* systems that can withstand increasingly severe restrictions regarding electric disposibility); operative-system stability is an out-of-scope problem regarding *any* user space application, not only RDBMs: if the underlying OS fails you can't and shouldn't expect *any* user-space executable to be reliable; you still can build *whole* systems under given severity constraints and -as I already said, one of you most basic considerations will be knowing what you can and should expect from any given subsystem so you don't ask "A" to accomplish tasks under "B's" responsibilities.
"I accept that most or all databases are broken in this fashion"
Taking into account the ammount of talent and money devoted to the database management bussiness, I'll ask you to consider the slim chance that it might be you, not them, the one in the wrong position.
"In that case, please allow me to submit that the database is broken."
No, it isn't.
"Snapshots are a frozen moment in time."
Sorry but no, they aren't. Snapshots are *filesystems* frozen in a moment of time. Nothing more, nothing less. They explicitly are not RAM frozen in time and they are explicitly not done at a database-safe moment of time.
"Failing, under any circumstances, to be able to recover a working database from a snapshot, is the same as failing to recover a working database after hardware or power failure or an obscure kernel bug crashes the system."
That's right, so what? Nobody is forcing you to have database snapshots if not needed and there's no way that your database backlog wil be sincronized to your snapshot policies and procedures unless a sysadmin has worked for them to be that way.
"Things should not be so fragile."
They are, and you better work under reality checks, not based on what they "should" be.
Of course, you can have whole *systems* to be such and such reliable but in order to achieve that you need to "tell" each subsystem to do their own homework: the database-thing to take care of their own "things" (i.e.: transaction-based checkpoints); the RAM-thing to take care of their own "things" (like ECC parity); the persistance-thing to take care of their own "things" (like LVM, RAID, system-level backups, etc.)...
Once you learn which "things" are part of what "things", you can design and deploy a reliable system; failing that you will have people actually thinking that RAID is a form of backup; that filesystem-level copies can allow you to forget about how to properly secure your database, etc.
"actually yes, you can do just that - tell the database that the filesystem level 'thing' needs to get a coherent snapshot of the database level 'thing'"
Thank you for actually reinforce my point: in order to do "a database thing" (a backup) you "tell the database..." (your own words).
"Can't these problems generally solved with snapshots (in FreeBSD and Linux LVM lingo) or shadow copies (in Windows parlance)?"
Short answer: no.
Long answer: filesystem level "things" don't know about database "things" so there may be "things" at the database level that lead to a non-functional database when coped with at the filesystem level.
This assumption is valid not only when talking about filesystems and databases but on every two unrelated items. Learn this and you probably know half you need to be a proper IT professional.
"The $hardware$ $programmer-time$ equation is always based on the assumption that the programmer is always worth their qualifications."
The $hardware$ $programmer-time$ equation is always based on the *stupid* assumption that the programmer is always worth their qualifications. There, corrected for you.
As I read the article, this all about a programming problem, not a running environment. While your post is quite sensible, real environments are not. My reality check is that when on a dev mill the question is risen about throwing iron or minds to a problem the answer is *always* "minds". Think about it for a moment: it is either stated as "this is CPU|I/O|whatever problem we have, our experts can|can't cope with it and it would cost $$$ deal with the problem with new hardware for so much time at current growth rate, or we can|hope to cope with it if Joe and Anne work on it somewhere between 100 and 200 hours each" and then there's no margin for even begging the question; the answer is an almost immediate output from the analysis... Or it is risen like "we don't have the slightest idea why this runs so slow, but we *hope* we can deal with it with bigger iron, what should we do?"
Even when a sensible manager faced at the second situation thinks like "this is obviously a 'more mind' -well, more like 'better minds' situation, but we don't have the expertise here, nor the time and/or the ability to find it elsewhere, so let's go for the more iron path and hope for the better", things go no better. The problem is the development team tried to bite bigger than their mounths and the more iron you throw to the problem, the more stressed the lacking codebase becomes and the more it falls appart.
I am not telling that "all things being equal", throwing more iron is not a solution -in fact it is, but that most of the time, the "more iron or more people" dichotomy it's only "we don't know any better" in disguise, which is, in fact, "we don't know good enough to make it work". You can find a "more iron" situation rarely and on short periods if your app changes niche (our nice app designed to work on a single server does indeed so nice that some fortune-500 want it so it's going from 10tps to 300tps. What can we do to fullfill the situation for some months while we rethink it?), but I'd say 99% when an app is not outspec'd but still seems to require more iron than initially planned it is because a design fault, and design faults on anything more complex than a "hello world" are never so easy to fix as changing just one environment variable: they tend to fail on a lot of fronts and you are not going to cope with the situation just "throwing more iron to it".
"When did it become SUSE instead of SuSE? DuDE!!"
On September 30 2003, dude!!