Right, that's because they decided licensing you wasn't worth the time and effort. To drive you must pass a test. To be able to vote, you have to not commit a felony before the age of 18 that is so bad you are tried as an adult. I know of nothing that will permanently recind your right to drink.
Me, I earned the right to stay out late when I came home on time regularly without causing any trouble.
I earned the right to decide what activities I wanted to participate in when my parents felt I was responsible and intelligent enough to decide what I thought was a constructive use my spare time.
I earned the right to watch cartoons on Saturday mornings when I got my chores done.
I earned the right to mow lawns for money after I'd done a good job on the maintence on the mower and lawn at home.
I earned the right to not eat with my family when I had the responibilities of working a job.
I earned all kinds of rights as a child. The state gives them to you at specific ages, because they can't deny you the rights given to by the Constitution forever, but it's clear your average 10 year old shouldn't be allowed to vote.
I wouldn't mind having a citizenship test for all American Citizens, that until you can pass the test, you aren't eligble to vote. That or the old Roman citizenship test where you had to serve the country or be sponsored and have a vote on your worthyness.
No children do NOT have all the rights adults have. Children do not have the right to vote. Children do not have the right to enter into a legally binding contract. Children do not have the right to drive a car. Children do not have the right to consume alchol. Children do not have the same rights in a court of law. Children do not have the right to make several determinations for themselves (which parent to live with in a divorce, weather or not they want to go to public school, what forms of medical treatment they will accept). Children do not have the right to own a gun, or get a carry concel permit for one.
Rights come with responsibilities. Children are inheriently irresponsible, precisely because they are children, thus they lack rights. Until they come of an age to take care of the associated repsonibilities they do NOT have the rights an adult has.
You live in fantasy land if you truely believe children have every right an adult does.
Even the Bill of Rights is limited in it's application to children.
It is the job of the child to earn those responsibilites, and the adults should nuture and enable the child to be able to handle responsibilities. However, should the parent not do so, the child is at fault when they come of majority age if they do not appropriately live withing the rights and responsibilities.
A child should learn to deal with those rights and responsibilites irrespective of the parents and the upbringing they receive. The fault lies with the child, not with the parent. While we may condemn the parent for the lack of parenting, when the child becomes an adult, it is the former child whom is punished, not the adult that failed to instruct the child.
I don't know, kids understand the conditioning we give them that drugs are bad, and ignore it. They understand and accept that they can't, drink, drive, vote, or join the military until a specific age. Trust me, kids will both understand and know that tracking them is a violation of the rights of an adult. However, it's very important that kids learn that kids aren't adults, and they don't have the rights adults have. They get them as they earn them.
I've got no problem with them using them. I've got no problem with somebody handing me one. Hell, I use one at work to get me in and out of buildings. It's an RFID in my wallet everywhere I go all day long. If they want to keep track of me, using my cellphone logs to tell where I was would be a good way.
My only issue would be if they are used to track physically where the children are in the building at a specific time. So if little Johnny was in the wrong place at the wrong time it can't be used as judge and jury when it comes time to discipline kids (well the RFID says you where there little Johnny). That'd be wrong. To take attendence with, and use to track what kids bought in the lunch line? Geez, do any of you people use a credit card? You realize that this isn't much different then the clock you punch into at work, and the credit card you use to pay for stuff.
I'd be more worried about them tying the proxy logs of where the kids visit on the web in the library then this.
The part I'm curious about is what do you have to do if the kid loses the card? What if they don't have it that day. Is there a way to override it? Is there a way to deal with it?
Please, your analogy is a at least as bad has the one you claim to be correcting for accuracy.
Nope, your ISP, didn't get root on your box. They gave you software, that connects you to their network, and asked you to log in as root to install it. The software does not self propogate. The software doesn't send secrets from computer A to computer B. Next you'll be telling me it's illegal for if Netscape just set Netscape to be your default browser without asking you. Netscape doesn't, and never has. IE I think at one point did, and the only reason that's illegal is because Microsoft was found to be a monopoly.
I suppose you think installing Service packs is hacking because they fool with settings that seem unrelated, poorly documented, and at various points break other perfectly fine software.
AOL didn't "hack" into your computer. You willingly ran the software they gave you. Now, I find what they are doing somewhat suspect, because they quietly changed a setting, that if I was a Windows user I would have found the knob for it long ago.
I'll point out that, recommending you comment out the telnet line, is completely different then when you install pppd it went into your/etc/inetd.conf and turned fiddled with it to turn it off for you.
I'd be pissed if pppd did that if it wasn't documented clearly (for a variety of reasons, upto and including the fact that I forgot to turn off telnet on a machine I ran). Mostly because the people who wrote pppd shouldn't be fiddling with my inetd.conf settings.
I didn't get the impression from the Slashdot story that they are doing it in software. However, that makes me think you are correct, it's FUD. Goodness, is it a crime to install software which enables IIS for you, because enabling IIS has security flaws? I'm pretty sure various pieces of software enable IIM for you when you install them. No 17 year old kid convinces you to install highly useful software, and pay them for a subscription service, and also happens to install BackOrifice on your computer. If it was documented to install BackOrifice, I don't think they'd even have a complaint until somebody actually logged into BackOrifice.
If they wanted to be on the up and up about it, they'd refuse to install AOL until the messagner service was turned off and give you instructions about how to do it. Possible have a dialog box that was set up for you to click okay to approve it, or uncheck this box to leave the service running.
Well, it's quite simple. Where is the money? Who has the broadcasting rights to where? How do you control that? Generally the business model of broadcasting television isn't well thought out. How do they advertise. See Television people really like advertising revenue. It's consistant. It's not a spur of the moment deal. It's known well in advance, and collections isn't an issue, as you are dealing with maybe 10 major corporations that are sending $100K's of money around, not rather then 100K people sending in 10 bucks.
Part of the problem, is that desipte what you say, the technology still isn't terrible good. It's surely not as good as it obviously could be. If it was as good as it could be, you'd request subscription from your gateway to multi-cast address X port Y. It would keep asking up the router stream until it got to the source. I've never seen a video feed that looked good live. I really don't want to know what it would be like if there might be 50,000 people who are interested in it. I know the Victoria's Secret special is always a fiasco (never seen it, but the guys I know who tuned in said it was a fiasco).
There are plenty of problems, but most people don't do multicast terribly well.
Each packet could then be broadcast as needed down each internet link. Currently most streaming video is a UDP feed that is pretty inefficient. The resolution is crappy, the frame rate is horrible, on most of the video feeds I've ever seen. I might pay $5, however, I'd be more interesting in paying $10, and having you ship me the DVD of the live coverage in the next week. Don't bother editing, just take the recorded live broadcast, break it into 5 minute sections, and ship me the game.
It sure doesn't sound that way. They say you won't be able to send e-mail to ATT.com.
What AT&T is asking is for you to help AT&T to restrict incoming mail
to just our known and trusted sources (e.g., business partners, clients
and customers). Therefore, we need to know which IP address(es) are
used by your outbound e-mail service so we can selectively permit them.
Please send this information to the following e-mail address
(rm-antiattspam@ems.att.com).
I'd think it was cool if they said, look nothing on our mail server is going out over port 23 unless you register with us.
Uhhh, I do business with people on the AT&T network. At least I'm reasonable sure the 1000's of clients who use e-mail to contact me use it. I wonder what I need to do to get on the list.
Complete shock and disbelief at the first e-mail (the dreadfully short message at the bottom).
Has anyone actually called and confirmed with the 1-800 number that this truely is AT&T, and it really is what they are saying? I'm not sure I'll believe it until I see the e-mail actually start bouncing. That's clinically insane. Do they seriously believe they'll be able to pull this off? You mean ever time a small company creates a new mail server they'll have to contact AT&T with the outgoing SMTP servers? If this starts a major trend, you mean I'll have to contact lots of major ISP's to send mail to them?
Assuming this it to stop SPAM (what else could it be?), what's to stop a spammer from just calling up and saying I'm a legit mailer set me up? What do I do when I get assigned the IP from the old spammer? What will there policy be on setting you back up? Will there be an official form? How can they tell the Spammer just isn't dupping them a second time with a fake business?
This sounds like a terrible idea, and like their security people haven't really thought this through. About the only thing I like about it, is that it is a sign that major ISP's are starting to play hardball. I'm curious if one of their net admins was behind some of the major black lists that just got DDoS'ed off the net. I hope they accept e-mail from anybody with a legitimate MX record at least. At least for a little while. I can't believe they aren't going to do a black list instead of a white list.
What's the over-under on how long this takes to get pulled the plug on? There's no way this will last. It'll be a world class disaster. My guess is it won't last 15 business days.
I run Oracle, and well unless you absolutely need it, running in a raw filesystem is a bad idea. I'm pretty sure the caching I'm talking about goes on at the block device level, so it makes no difference if you use raw partitions, or a filesystem, it'll all get cached just the same (unless you use O_DIRECT, or "raw). You might setup raw devices, but that's relatively new in RedHat, I'm not up for it on my production systems.
Even Oracle says not to run on raw partitions. The management headaches aren't worth it. Oh, and try setting up and running it hot standby mode and dealing with the naming translations. I run on RedHat linux, and LVM isn't support, and I want to run off an MD device to get mirroring across controllers. You can't currently partition md devices, and to have the number of partitions under the 2GB limit (so I don't have to worry about bugs in the utilities I use to do my backups), I have to have lots, and lots of 2GB partitions. No thanks. Oracle8i on RH7.1 doesn't do larger then 2GB files now that I think about it. 9i does, but 8i (what I currently use, doesn't). I'm not sure if it'll do a 10GB partition under Linux, but I know it won't do files that large. I'll have to try it and see.
Well, I've got 170GB database, on a six disk disk raid set, that's mirrored to another six disk raid set. Well, lets see, that means I'd need ~80 partitions per disk, and I'd need roughly ~80 md devices. That's a lot of them, I thought you only got 64 minor numbers for md. It's a lot of managment overhead. It's a lot to keep track of. Instead, I let Oracle do it's thing on the filesystem, there are lots of easy to use tools to keep track of those, and that's the recommended install type. Also having very, very large files makes restores take longer in the case of a single file corruption. It means the data is more compartmentalized if you have corruption. They mean that you can't decompress right up to the limit of your filesystem size. (You can decompress 10 1GB files when you couldn't compress the same data as 1 10GB file, been there done that).
Oracle wants better O_DIRECT support, however I don't think even they use it under Linux. Last I heard, it still had some nasty bugs in it. Someone pointed out that the memory might just get mmapped. I've never run an strace on Oracle, or profiled it to find out what it does for memory management. I've always been told that most Oracle stuff goes into shared memory segements, and that Oracle does reads and writes in block size, that it doesn't do mmap, but I could be wrong on that one. However, it could be that it mmaps it in.
Even the two high end Oracle DBA I know, who used to do raw partitions, say it really isn't worth it any more. There are too many other things that can go wrong using raw partitions. Just use filesystems ones until you can't throw more hardware or optimization at the problem. Eventually when you run out of other options, switch over to using raw partitions for that last 2-5% you'll ever get in terms of performance boost.
Once LVM is production ready by RedHat's support, or until the partitioning of MD devices is done, I'll stick with my filesystem thanks (Then I could have 5 MD's with 30 partitions each and make it work).
Still want to argue that I'm not being "realistic" here? I've got lots of good reasons for running it on my filesystem thanks. Not the least of which is, I can easily deal with transferring it from machine to machine (I don't have to remember what kernels are 64bit clean if I move it to an older system). I can easily backup it to a filesystem. I can easily recover/restore/rebuild it on any old machine that has enough disk space. I don't have to go fiddle with the partition of the disks. Until recently Oracle and Linux didn't do 64bit block devices nicely. That's all new stuff since I started my Oracle database. I got my performance another way.
We run SCSI, because IDE drives just die like dogs under the load we put these SCSI devices. I've seen it ha
Actually, I'm sure there are parts of the Manhatten Project that should remain a secret. I'm pretty sure that a lot of things about parts of WWII are kept underwraps just to ensure that everybody is dead before the facts are released. I'm sure there are some commando's from WWII out there who really don't want visits from grandson's of men they massacurred 50 years ago to come find them (that goes for Axis and Allied commando's). I'm really sure that "who shot Kennedy" shouldn't be declassified (assuming we know it), it could cause mass hysteria if we found out it was "X", where X, is the cuban, the mafia, the Russians, the CIA, or the magic bullet from space.
The military has lots of things it keeps under wraps that should be left that way. If there are still encryption algorithms that aren't generally known to the public, and are still considered secure by the NSA, they should probably stay secured. Stuff like that.
Lookup the word utilization in the dictionary...;-).
If you run out of PCI bandwidth because you are using all of it, that's a good thing. It means you have reached the theoretical peak performance of your PCI bus. You have two options, get more busses, or get a different bus (think 64bit PCI, 66Mhz bus, or one of the new I/O busses that Intel is working on).
If you saturate your PCI bus at nearly 100% utilization, it really doesn't matter weather you are using IDE or SCSI or bit-saving squirrels, as your backend. Your bottle neck is no longer SCSI or IDE, it's your PCI bus. If both SCSI and IDE have equally good thru-put on the PCI bus, there really isn't a lot to complain about. In that case, you can use either one, it won't make you run any faster.
If using one has a 100% utilization, but 50% of it is wasted in bus arbitration, or competing for the bus, that's bad. If the other has 100% utilization, and 2% wasted on bus arbitration, that's good. You'd want the latter, instead of the former. That's all. A PCI bus generally has a lot of bandwidth, managing to use it all up is very, very good.
Find me a motherboard that can take 10GB of RAM that you can assemble a disk subsystem for less then $2K (my point was that $500, it's hard to even get your foot in the door on SCSI, a good card, a good cable, and a drive will run you that much, $2K gives you more options to build a better SCSI system, possible a better performing one then you can build with IDE on PC architecture). Better yet, find me 10GB of RAM for less then $2K. Generally speaking, heavy updates, implies you are bound by the speed of the bits to the platter, because those will be sync'ed before allowing the process to move on to a read. A 64MB cached SCSI controller, can make up an awlful lot of ground on an IDE subsystem in terms of performance (plus the disk caches), and Tagged Queing, and offloading the work from your CPU sure do help alot.
If the updates are independent of the reads, you probably have a point that a lot of extra RAM will help. In my experience, you'll need that RAM either way (SCSI or IDE). Then your problem isn't really I/O bound any more. SCSI would be overkill for such a situation. I generally only think SCSI in terms of multiple disks, or highload disks that have lots of concurrent reads/writes going on, not as "backing store" for my RAM that gets written to periodically, at that point any old drive will do. (At that point, investing in a RAM disk to be your journal is a better option).
It was my understanding that high end ATA cards cost roughly ~$300-500USD for one that is comparable to a good SCSI card. The crappy ATA-RAID cards that cost $100 I've seen generally only support 2 or maybe 4 drives. I was comparing controller to controller. Drive to drive is hard to do. Also don't forget, when you buy IDE, you have to pay for the one with the 3-5 year warrantee to compare it to SCSI, those don't run you $100. They are pricer then that. I haven't priced one in a while, but I know the warrantee costs plenty extra.
Having 4-8 cards compete for bandwidth is worse then one having all of the bandwidth. 2 SCSI controllers I can have at least 32 disks, possibly up to 128 disks using only two slots. Given that I like to do mirroring I'd definitely have two SCSI cards. When they compete for bandwidth, you have a problem. You want complete utilization to optimal thru-put. Which means, fewer cards. Fewer cards in IDE moves you a lot closer to SCSI pricing.
Actually one point for builtin SCSI/IDE, is that it actually isn't on the PCI bus (I believe they aren't on the south bridge, or they are on their own individual PCI bus, giving them a large advantage bandwidth wise to the CPU).
No, RAM isn't useless on writes. It's useless on syncronous writes ( a write immediately followed by a fsync(fd); in UNIX parlance). Which is what a lot of I/O bound operations do. It's highly useful on streaming writes (say large downloads). If you don't believe me jigger up an LD_PRELOAD for write that does remaps write(); to write(); fsync();. Buffering your writes allows you to accumulate more writes to a single track, or closely related tracks, thus allowing you to avoid extra seeks. It also allows you to do "delayed" allocations on filesystems that support it, so a temp file that never gets written to the disk unless it has to. In general more RAM is good, assuming the application you are using it's useful. In general, if RAM is good for your file cache, your application isn't that I/O bound.
Usually on I/O bound operations, they have custom caching that negates most of the cache for the filesystem. Filesystem cache at that point becomes a duplicate, and in a lot of ways overhead (to keep track of it, and to do the copying of the bytes). This is just as true for reads as it is for writes. FS cache is pure overhead in well done database. Most other applications that are performance critical will want to do their own caching, or they aren't performance critical.
Not really. If you move that to $2000 IDE would really suffer. For starters, it's hard to put that many IDE drives on a single machine. Once you do, you will start running out of PCI slots, or PCI bandwidth in a hurry (running out of PCI bandwidth is a good thing in one sense, but only if you are actually utilizing it well).
You could move to a ATA-RAID card, which would save you some slots, but that starts to get into the SCSI price range.
Putting extra RAM in the machine is not really fair. You normally buy enough RAM for your applications, not for spare buffering (at least on my databases that's the case, spare RAM there is a waste, because normally writes are sync'ed, or you have two copies, one in the DB cache, and one in the filesystem cache), so for most of my I/O performance needs, extra RAM is relatively wasteful. It's actually a hindrence under Linux on a 2.4 kernel to have more the 2GB of RAM for I/O performance. I believe it's getting better under 2.6.
Right, I mean duplicating it, and you have to put in a list into your cache (I'd like that to be auto updated). I'd like the complete root zone,.com zone,.net zone, and a.org zone. So if for some reason all of the root servers are down, and I want to get to one and they are up, I can.
If DNS was distributed locally to everyone, there is no point in attacking the root servers.
Stick an Linux box between that the printer and the network, throttle the output of the card running to the ethernet port on the printer to 128kbit/sec.
You'll tie up stuff on the client end, however, your local OS could/should be doing spooling so it's not tying up the entire machine at that point.
I know how to do that at the IP level (traffic shaping and/or rate limiting) that is production ready. I believe there are projects to do that at the ethernet level (to control the speed of the ethernet frames), but I'm not sure how production ready that stuff is.
Actually from what I know of Verisign, and what some other people have posted. He's got a $150Mil in money invested in networking infrastructure to ensure that his root server never goes down (it sounds like the two root servers he has, are backed up by 8-10 clustered backends that are distributed around the country and connected via leased lines). It sounds like his company has the single most reliable root server on the planet. He's spent a lot of money doing it, he should be proud of how reliable it is. Having them the root servers owned by people that understand, that they house a critical piece of infrastructure, and that it's critical that it be maintained, and secured is a very good idea. He's a business man, the thinks the only way to do that is to build a business around it, then you have the means, and the motivation to secure it. What he doesn't understand that is a lot of others could do it too.
One of the things I don't quite follow, is why DNS isn't even more distributed. Why don't Tier one providers copy the root zones, and the.com,.net, and.org zones to local servers, who will push them down to Tier 2 ISP's, so I can suck my own copy of there? Then as I have my copy of the data for as long as I need it. As long as I can upgrade it once a day, it's all good. Yeah, there are hundreds of thousands of domains in there, but if I could at least get my own copy of where the.com and.net zone servers are, that'd be good enough (I know I can do it via script if I so chose).
I've have no objections to more root servers being owned by commerical entities who have financial incentive to keep the running under all circumstances. My only other concern about say NASA, DoD, and some of the other non-commerical entities, is that the quality of the service provided might be up to the admin who admins the machine. I'm not sure the knowledge of how to feed and care for a DNS server is "institutionalized" there, like it would be at say ISC and Verisign would obviously have.
My primary beef with what he said, is that "99% of the traffic is HTTP". Yeah, but that 1% (Which is a inaccurate stat) is important stuff. HTTP isn't the internet. I wish he would see that fooling around with *.com. breaks the assumptions of a *LOT* of software. I mean, why doesn't he just return 9.9.9.9 to you for all name lookups, and then setup a transparent Squid Proxy server there, so he can really "innovate" by controlling the content on the entire WWW.
As so many people have said before the User Agent is where what Site Finder is trying to accomplish should be done. It's a good idea, it's proper to do that at the User agent level. The problem is, that the Verisign doesn't get a piece of the pie that way. Man just go find another way to earn money. You sound clever, just don't fool around with returning bogus names to.com zone requests.
You should tell that to all the guys on the AF base in Bellevue NE. While SAC isn't what it once was, there is still SC (Strategic Command, instead of Strategic Air Command) there. I still see enough military planes around, and this it is where the president flew on 9/11, so clearly it's a pretty damn safe place in the opinion of the military.
Actually height, and more specifically relative height while young, if you were particularly tall when you were young, is correlated strongly with the early development of your brain, and higher IQ's. It's been related to a lot of chemical, and genetic processes that happen early in life that affect you through out the course of your life. I've always liked this fact, that given I was always that tallest kid in grade school. If I recall correctly though, if you end up being 5ft 6in, if you got there by age 10, your are in the "advanced" group. It wasn't purely a height issue, it was how quickly did you reach your final height.
Remember, it was IBM who made the hardware specs open. They thought they could control the platform by controlling the BIOS. Then they tried to put the Genie back in the bottle by releasing microchannel (MCA). It was a forgone conclusion by the time when IBM made open hardware specs, this would happen. Bill was in the right place at the right time, and was smart enough to see a way to endup monopolizing the market.
I'm willing to concede that it's entirely possible that Bill Gates didn't say that, if your willing to concede that Bill Gates in 1996 might really not remember saying it, or might really be lying.
It's not like revisionist history is a new concept. In 1981, I could completely see, Bill Gates saying the 640K quote, and have it taken out of context. One of the Watson's (of founding IBM fame, I can't remember if it was Sr, or Jr. I'm guessing Sr), once said that worldwide we'd probably only need 5 computers ever. It's not like he's terrible stupid either.
If you really want to have fun and games, write down a particular fact that you can't remember a specific event ever happening in your childhood. Now, store that piece of paper someplace safe. Now everyday imagine that event happening. Picture in your mind how you would remember it if it happened. Over the course of time, you'll "remember" it as a fact that is just like all of your other memories from childhood. You'll know it's inaccurate, but to your mind you can't tell between a the old true memories, and the newly fabricated memories. It's a simple form of brainwashing. I've specific memories that I know for a fact never happened. I constructed a conversation I never had once for the purpose of trying this out. It's the old adage about a lie repeated often enough becomes true.
I'll willingly admit it's entirely possible Bill never said that, and he surely can't prove he never said it. However, I'll never trust Bill's memory about him not saying it. However, if you tracked down the original references to it and debunk that, now you have something. Somebody has to cite it. It's in the Usenet Archives, or in old papers and trade magazines. Find the originals and debunk them, don't cite Bill saying 15 years later that he didn't say it. That's not debunking.
Here, I'll prove it to you. "I've done some stupid things, and I've done some wrong things, but I was never born. Nobody in the human race would ever say they were born.". Does that "debunk" the fact that I was born or not? I'd say my sitting here, and typing into slashdot is pretty strong evidence I was born at some point in the past.
A number of statistics have been proven to be false, but are cited all the time in the past. If you follow all of the original citations back, you'll find they all start at one single reference. The original person who stated it, either lied, or had something wrong with the way they came to the conclusion. By the time anybody figures that out, it'll be a "fact". I know this happened on stuff reguarding sexual orientation (formely common cited stat that 10% of all men are gay), and I believe it's happened on several other occasions about other commonly cited stats.
Debunking involves getting reasonable close to the source and debunking it. Not asking somebody 20 years later, who has a vested interest in not looking like an idiot, if he said something that's blatantly stupid 20 years ago. Read up on what Bill has said about what he thought of the internet.
I believe it was Cringely who pointed out that Bill always proclaims he was a visionary about the net, and saw ahead of everyone how much that could change the world. Yet when you read his book from that time where he was spouting off about what he thought was the next big things in computers, just as the internet went mainstream he never mentioned it once. Bill's in a position where he can't afford to say, I missed that huge new technology. He's Bill Gate's, he thinks Microsoft single handedly invented the Personal Computer. Just read the end of the article.
Security clearance, and an NDA are completely different things. He quite clearly was talking about an NDA from a Fortune 500 company about Database work. If he said, I've done work under Top Secret clearance for the last 10 years, no I wouldn't recommend committing treason (or whatever level of Federal law). Any issue you have a clearance for, they should have explained quite clearly what it is you can tell. Just call the people who cleared you, and tell them you need to write a resume, I'm sure they'll explain to you in detail what you are allowed to disclose. For issues of National Security, the Federal Gov't will throw your ass in jail, and throw away the key if need be for the security of the nation.
Insider trading, is again a completely different topic from keeping secret high level details about optimization of a database. An NDA, can't cover the fact, that did consulting working on database from vendor X, and during the course of 6 months increased the efficency by Y%, using platform Z, and OS A,B, and C as clients. Also, if you ask 20 leaving employees for what they we're duing, it is you whose guilty of insider trading, not the 20 employees. Your just talking about your old job trying to get hired.
I believe that where I live (Nebraska), has a reasonable strong precedence for "nothing I sign as an NDA or non-compete can force me to give up my right to make a living in my professional field". They can limit areas of what you can do, like doing the same work for direct and indirect competitors.
In essence, write up your resume however you like, ignoring the NDA. (My suspicion, is that either, your NDA is uninforceable if you can violate it in less then the 3 bullet points, and 1 paragraph description of what your job is, or you are writting your Resume incorrectly). If what you do is standard industry practice, nearly every NDA I've ever seen says "anything that is standard industry practice is not covered by this NDA". It's completely uninforceable otherwise. They pay you to tell them common knowledge, the then hold you hostage to never give up that information again for the term of your NDA.
No mister roofer, you put shingles on my roof. For the next 3 years, every roof you build can't have shingles, otherwise I get to sue the bejesus out of you. You can't even tell the next guy you work for, that you know how to apply wooden squares to a angled board to keep water from hitting you. Nope, I've got an iron clad NDA right here. Uh-huh. Judge would throw you out on your ear withing 15 minutes of hearing the arguments for doing that.
For goodness sakes, your describing what's already known to the general population in the industry. No offense, but your not doing anything terrible secret in the area of RDBM's database optimation. Most of the really cool stuff was figured out 15 years ago, and they've been waiting around for the hardware to catch up. Lets see, schema layout, indexing, denormalizing data, princepal of locality (temporial and spacial), and instrumentation/profiling. Everything else probably came straight out of the vendor documentation, and was applied to a particular database. Optimation comes down to "time this, change param A, time again, change A some more, lather, rinse, repeat. Move to parameter B". Maybe you've figured some cool stuff out about ODBM's or something. However, even most of that is just following the bouncing ball after reading the documentation from the Vendor.
If they come after you, just explain to them, you'll take your six figure settlement now, or you'll find a nice attorney who'll work on contingency for 1/3 of the 7 figure settlement after the trial.
I'm not a lawyer, this isn't legal advice. However, they can't compel you to give up your right to earn a living. Barring a matter of national security at least.
Slashdot Mirror
Me, I earned the right to stay out late when I came home on time regularly without causing any trouble.
I earned the right to decide what activities I wanted to participate in when my parents felt I was responsible and intelligent enough to decide what I thought was a constructive use my spare time.
I earned the right to watch cartoons on Saturday mornings when I got my chores done.
I earned the right to mow lawns for money after I'd done a good job on the maintence on the mower and lawn at home.
I earned the right to not eat with my family when I had the responibilities of working a job.
I earned all kinds of rights as a child. The state gives them to you at specific ages, because they can't deny you the rights given to by the Constitution forever, but it's clear your average 10 year old shouldn't be allowed to vote.
I wouldn't mind having a citizenship test for all American Citizens, that until you can pass the test, you aren't eligble to vote. That or the old Roman citizenship test where you had to serve the country or be sponsored and have a vote on your worthyness.
Kirby
Rights come with responsibilities. Children are inheriently irresponsible, precisely because they are children, thus they lack rights. Until they come of an age to take care of the associated repsonibilities they do NOT have the rights an adult has.
You live in fantasy land if you truely believe children have every right an adult does.
Even the Bill of Rights is limited in it's application to children.
It is the job of the child to earn those responsibilites, and the adults should nuture and enable the child to be able to handle responsibilities. However, should the parent not do so, the child is at fault when they come of majority age if they do not appropriately live withing the rights and responsibilities.
A child should learn to deal with those rights and responsibilites irrespective of the parents and the upbringing they receive. The fault lies with the child, not with the parent. While we may condemn the parent for the lack of parenting, when the child becomes an adult, it is the former child whom is punished, not the adult that failed to instruct the child.
Kirby
I've got no problem with them using them. I've got no problem with somebody handing me one. Hell, I use one at work to get me in and out of buildings. It's an RFID in my wallet everywhere I go all day long. If they want to keep track of me, using my cellphone logs to tell where I was would be a good way.
My only issue would be if they are used to track physically where the children are in the building at a specific time. So if little Johnny was in the wrong place at the wrong time it can't be used as judge and jury when it comes time to discipline kids (well the RFID says you where there little Johnny). That'd be wrong. To take attendence with, and use to track what kids bought in the lunch line? Geez, do any of you people use a credit card? You realize that this isn't much different then the clock you punch into at work, and the credit card you use to pay for stuff.
I'd be more worried about them tying the proxy logs of where the kids visit on the web in the library then this.
The part I'm curious about is what do you have to do if the kid loses the card? What if they don't have it that day. Is there a way to override it? Is there a way to deal with it?
Kirby
Nope, your ISP, didn't get root on your box. They gave you software, that connects you to their network, and asked you to log in as root to install it. The software does not self propogate. The software doesn't send secrets from computer A to computer B. Next you'll be telling me it's illegal for if Netscape just set Netscape to be your default browser without asking you. Netscape doesn't, and never has. IE I think at one point did, and the only reason that's illegal is because Microsoft was found to be a monopoly.
I suppose you think installing Service packs is hacking because they fool with settings that seem unrelated, poorly documented, and at various points break other perfectly fine software.
AOL didn't "hack" into your computer. You willingly ran the software they gave you. Now, I find what they are doing somewhat suspect, because they quietly changed a setting, that if I was a Windows user I would have found the knob for it long ago.
Kirby
I'd be pissed if pppd did that if it wasn't documented clearly (for a variety of reasons, upto and including the fact that I forgot to turn off telnet on a machine I ran). Mostly because the people who wrote pppd shouldn't be fiddling with my inetd.conf settings.
I didn't get the impression from the Slashdot story that they are doing it in software. However, that makes me think you are correct, it's FUD. Goodness, is it a crime to install software which enables IIS for you, because enabling IIS has security flaws? I'm pretty sure various pieces of software enable IIM for you when you install them. No 17 year old kid convinces you to install highly useful software, and pay them for a subscription service, and also happens to install BackOrifice on your computer. If it was documented to install BackOrifice, I don't think they'd even have a complaint until somebody actually logged into BackOrifice.
If they wanted to be on the up and up about it, they'd refuse to install AOL until the messagner service was turned off and give you instructions about how to do it. Possible have a dialog box that was set up for you to click okay to approve it, or uncheck this box to leave the service running.
Kirby
Part of the problem, is that desipte what you say, the technology still isn't terrible good. It's surely not as good as it obviously could be. If it was as good as it could be, you'd request subscription from your gateway to multi-cast address X port Y. It would keep asking up the router stream until it got to the source. I've never seen a video feed that looked good live. I really don't want to know what it would be like if there might be 50,000 people who are interested in it. I know the Victoria's Secret special is always a fiasco (never seen it, but the guys I know who tuned in said it was a fiasco).
There are plenty of problems, but most people don't do multicast terribly well.
Each packet could then be broadcast as needed down each internet link. Currently most streaming video is a UDP feed that is pretty inefficient. The resolution is crappy, the frame rate is horrible, on most of the video feeds I've ever seen. I might pay $5, however, I'd be more interesting in paying $10, and having you ship me the DVD of the live coverage in the next week. Don't bother editing, just take the recorded live broadcast, break it into 5 minute sections, and ship me the game.
Kirby
Kirby
I'd think it was cool if they said, look nothing on our mail server is going out over port 23 unless you register with us.
Kirby
Complete shock and disbelief at the first e-mail (the dreadfully short message at the bottom).
Has anyone actually called and confirmed with the 1-800 number that this truely is AT&T, and it really is what they are saying? I'm not sure I'll believe it until I see the e-mail actually start bouncing. That's clinically insane. Do they seriously believe they'll be able to pull this off? You mean ever time a small company creates a new mail server they'll have to contact AT&T with the outgoing SMTP servers? If this starts a major trend, you mean I'll have to contact lots of major ISP's to send mail to them?
Assuming this it to stop SPAM (what else could it be?), what's to stop a spammer from just calling up and saying I'm a legit mailer set me up? What do I do when I get assigned the IP from the old spammer? What will there policy be on setting you back up? Will there be an official form? How can they tell the Spammer just isn't dupping them a second time with a fake business?
This sounds like a terrible idea, and like their security people haven't really thought this through. About the only thing I like about it, is that it is a sign that major ISP's are starting to play hardball. I'm curious if one of their net admins was behind some of the major black lists that just got DDoS'ed off the net. I hope they accept e-mail from anybody with a legitimate MX record at least. At least for a little while. I can't believe they aren't going to do a black list instead of a white list.
What's the over-under on how long this takes to get pulled the plug on? There's no way this will last. It'll be a world class disaster. My guess is it won't last 15 business days.
Kirby
Even Oracle says not to run on raw partitions. The management headaches aren't worth it. Oh, and try setting up and running it hot standby mode and dealing with the naming translations. I run on RedHat linux, and LVM isn't support, and I want to run off an MD device to get mirroring across controllers. You can't currently partition md devices, and to have the number of partitions under the 2GB limit (so I don't have to worry about bugs in the utilities I use to do my backups), I have to have lots, and lots of 2GB partitions. No thanks. Oracle8i on RH7.1 doesn't do larger then 2GB files now that I think about it. 9i does, but 8i (what I currently use, doesn't). I'm not sure if it'll do a 10GB partition under Linux, but I know it won't do files that large. I'll have to try it and see.
Well, I've got 170GB database, on a six disk disk raid set, that's mirrored to another six disk raid set. Well, lets see, that means I'd need ~80 partitions per disk, and I'd need roughly ~80 md devices. That's a lot of them, I thought you only got 64 minor numbers for md. It's a lot of managment overhead. It's a lot to keep track of. Instead, I let Oracle do it's thing on the filesystem, there are lots of easy to use tools to keep track of those, and that's the recommended install type. Also having very, very large files makes restores take longer in the case of a single file corruption. It means the data is more compartmentalized if you have corruption. They mean that you can't decompress right up to the limit of your filesystem size. (You can decompress 10 1GB files when you couldn't compress the same data as 1 10GB file, been there done that).
Oracle wants better O_DIRECT support, however I don't think even they use it under Linux. Last I heard, it still had some nasty bugs in it. Someone pointed out that the memory might just get mmapped. I've never run an strace on Oracle, or profiled it to find out what it does for memory management. I've always been told that most Oracle stuff goes into shared memory segements, and that Oracle does reads and writes in block size, that it doesn't do mmap, but I could be wrong on that one. However, it could be that it mmaps it in.
Even the two high end Oracle DBA I know, who used to do raw partitions, say it really isn't worth it any more. There are too many other things that can go wrong using raw partitions. Just use filesystems ones until you can't throw more hardware or optimization at the problem. Eventually when you run out of other options, switch over to using raw partitions for that last 2-5% you'll ever get in terms of performance boost.
Once LVM is production ready by RedHat's support, or until the partitioning of MD devices is done, I'll stick with my filesystem thanks (Then I could have 5 MD's with 30 partitions each and make it work).
Still want to argue that I'm not being "realistic" here? I've got lots of good reasons for running it on my filesystem thanks. Not the least of which is, I can easily deal with transferring it from machine to machine (I don't have to remember what kernels are 64bit clean if I move it to an older system). I can easily backup it to a filesystem. I can easily recover/restore/rebuild it on any old machine that has enough disk space. I don't have to go fiddle with the partition of the disks. Until recently Oracle and Linux didn't do 64bit block devices nicely. That's all new stuff since I started my Oracle database. I got my performance another way.
We run SCSI, because IDE drives just die like dogs under the load we put these SCSI devices. I've seen it ha
The military has lots of things it keeps under wraps that should be left that way. If there are still encryption algorithms that aren't generally known to the public, and are still considered secure by the NSA, they should probably stay secured. Stuff like that.
Kirby
If you run out of PCI bandwidth because you are using all of it, that's a good thing. It means you have reached the theoretical peak performance of your PCI bus. You have two options, get more busses, or get a different bus (think 64bit PCI, 66Mhz bus, or one of the new I/O busses that Intel is working on).
If you saturate your PCI bus at nearly 100% utilization, it really doesn't matter weather you are using IDE or SCSI or bit-saving squirrels, as your backend. Your bottle neck is no longer SCSI or IDE, it's your PCI bus. If both SCSI and IDE have equally good thru-put on the PCI bus, there really isn't a lot to complain about. In that case, you can use either one, it won't make you run any faster.
If using one has a 100% utilization, but 50% of it is wasted in bus arbitration, or competing for the bus, that's bad. If the other has 100% utilization, and 2% wasted on bus arbitration, that's good. You'd want the latter, instead of the former. That's all. A PCI bus generally has a lot of bandwidth, managing to use it all up is very, very good.
Kirby
If the updates are independent of the reads, you probably have a point that a lot of extra RAM will help. In my experience, you'll need that RAM either way (SCSI or IDE). Then your problem isn't really I/O bound any more. SCSI would be overkill for such a situation. I generally only think SCSI in terms of multiple disks, or highload disks that have lots of concurrent reads/writes going on, not as "backing store" for my RAM that gets written to periodically, at that point any old drive will do. (At that point, investing in a RAM disk to be your journal is a better option).
Kirby
Having 4-8 cards compete for bandwidth is worse then one having all of the bandwidth. 2 SCSI controllers I can have at least 32 disks, possibly up to 128 disks using only two slots. Given that I like to do mirroring I'd definitely have two SCSI cards. When they compete for bandwidth, you have a problem. You want complete utilization to optimal thru-put. Which means, fewer cards. Fewer cards in IDE moves you a lot closer to SCSI pricing.
Actually one point for builtin SCSI/IDE, is that it actually isn't on the PCI bus (I believe they aren't on the south bridge, or they are on their own individual PCI bus, giving them a large advantage bandwidth wise to the CPU).
No, RAM isn't useless on writes. It's useless on syncronous writes ( a write immediately followed by a fsync(fd); in UNIX parlance). Which is what a lot of I/O bound operations do. It's highly useful on streaming writes (say large downloads). If you don't believe me jigger up an LD_PRELOAD for write that does remaps write(); to write(); fsync();. Buffering your writes allows you to accumulate more writes to a single track, or closely related tracks, thus allowing you to avoid extra seeks. It also allows you to do "delayed" allocations on filesystems that support it, so a temp file that never gets written to the disk unless it has to. In general more RAM is good, assuming the application you are using it's useful. In general, if RAM is good for your file cache, your application isn't that I/O bound.
Usually on I/O bound operations, they have custom caching that negates most of the cache for the filesystem. Filesystem cache at that point becomes a duplicate, and in a lot of ways overhead (to keep track of it, and to do the copying of the bytes). This is just as true for reads as it is for writes. FS cache is pure overhead in well done database. Most other applications that are performance critical will want to do their own caching, or they aren't performance critical.
Kirby
You could move to a ATA-RAID card, which would save you some slots, but that starts to get into the SCSI price range.
Putting extra RAM in the machine is not really fair. You normally buy enough RAM for your applications, not for spare buffering (at least on my databases that's the case, spare RAM there is a waste, because normally writes are sync'ed, or you have two copies, one in the DB cache, and one in the filesystem cache), so for most of my I/O performance needs, extra RAM is relatively wasteful. It's actually a hindrence under Linux on a 2.4 kernel to have more the 2GB of RAM for I/O performance. I believe it's getting better under 2.6.
Kirby
If DNS was distributed locally to everyone, there is no point in attacking the root servers.
Kirby
You'll tie up stuff on the client end, however, your local OS could/should be doing spooling so it's not tying up the entire machine at that point.
I know how to do that at the IP level (traffic shaping and/or rate limiting) that is production ready. I believe there are projects to do that at the ethernet level (to control the speed of the ethernet frames), but I'm not sure how production ready that stuff is.
Best of luck,
Kirby
One of the things I don't quite follow, is why DNS isn't even more distributed. Why don't Tier one providers copy the root zones, and the .com, .net, and .org zones to local servers, who will push them down to Tier 2 ISP's, so I can suck my own copy of there? Then as I have my copy of the data for as long as I need it. As long as I can upgrade it once a day, it's all good. Yeah, there are hundreds of thousands of domains in there, but if I could at least get my own copy of where the .com and .net zone servers are, that'd be good enough (I know I can do it via script if I so chose).
I've have no objections to more root servers being owned by commerical entities who have financial incentive to keep the running under all circumstances. My only other concern about say NASA, DoD, and some of the other non-commerical entities, is that the quality of the service provided might be up to the admin who admins the machine. I'm not sure the knowledge of how to feed and care for a DNS server is "institutionalized" there, like it would be at say ISC and Verisign would obviously have.
My primary beef with what he said, is that "99% of the traffic is HTTP". Yeah, but that 1% (Which is a inaccurate stat) is important stuff. HTTP isn't the internet. I wish he would see that fooling around with *.com. breaks the assumptions of a *LOT* of software. I mean, why doesn't he just return 9.9.9.9 to you for all name lookups, and then setup a transparent Squid Proxy server there, so he can really "innovate" by controlling the content on the entire WWW.
As so many people have said before the User Agent is where what Site Finder is trying to accomplish should be done. It's a good idea, it's proper to do that at the User agent level. The problem is, that the Verisign doesn't get a piece of the pie that way. Man just go find another way to earn money. You sound clever, just don't fool around with returning bogus names to .com zone requests.
Kirby
Kirby
Kirby
Kirby
http://leaf.sourceforge.net
It is the successor of the LRP project.
Kirby
It's not like revisionist history is a new concept. In 1981, I could completely see, Bill Gates saying the 640K quote, and have it taken out of context. One of the Watson's (of founding IBM fame, I can't remember if it was Sr, or Jr. I'm guessing Sr), once said that worldwide we'd probably only need 5 computers ever. It's not like he's terrible stupid either.
If you really want to have fun and games, write down a particular fact that you can't remember a specific event ever happening in your childhood. Now, store that piece of paper someplace safe. Now everyday imagine that event happening. Picture in your mind how you would remember it if it happened. Over the course of time, you'll "remember" it as a fact that is just like all of your other memories from childhood. You'll know it's inaccurate, but to your mind you can't tell between a the old true memories, and the newly fabricated memories. It's a simple form of brainwashing. I've specific memories that I know for a fact never happened. I constructed a conversation I never had once for the purpose of trying this out. It's the old adage about a lie repeated often enough becomes true.
I'll willingly admit it's entirely possible Bill never said that, and he surely can't prove he never said it. However, I'll never trust Bill's memory about him not saying it. However, if you tracked down the original references to it and debunk that, now you have something. Somebody has to cite it. It's in the Usenet Archives, or in old papers and trade magazines. Find the originals and debunk them, don't cite Bill saying 15 years later that he didn't say it. That's not debunking.
Here, I'll prove it to you. "I've done some stupid things, and I've done some wrong things, but I was never born. Nobody in the human race would ever say they were born.". Does that "debunk" the fact that I was born or not? I'd say my sitting here, and typing into slashdot is pretty strong evidence I was born at some point in the past.
A number of statistics have been proven to be false, but are cited all the time in the past. If you follow all of the original citations back, you'll find they all start at one single reference. The original person who stated it, either lied, or had something wrong with the way they came to the conclusion. By the time anybody figures that out, it'll be a "fact". I know this happened on stuff reguarding sexual orientation (formely common cited stat that 10% of all men are gay), and I believe it's happened on several other occasions about other commonly cited stats.
Debunking involves getting reasonable close to the source and debunking it. Not asking somebody 20 years later, who has a vested interest in not looking like an idiot, if he said something that's blatantly stupid 20 years ago. Read up on what Bill has said about what he thought of the internet.
I believe it was Cringely who pointed out that Bill always proclaims he was a visionary about the net, and saw ahead of everyone how much that could change the world. Yet when you read his book from that time where he was spouting off about what he thought was the next big things in computers, just as the internet went mainstream he never mentioned it once. Bill's in a position where he can't afford to say, I missed that huge new technology. He's Bill Gate's, he thinks Microsoft single handedly invented the Personal Computer. Just read the end of the article.
Kirby
Insider trading, is again a completely different topic from keeping secret high level details about optimization of a database. An NDA, can't cover the fact, that did consulting working on database from vendor X, and during the course of 6 months increased the efficency by Y%, using platform Z, and OS A,B, and C as clients. Also, if you ask 20 leaving employees for what they we're duing, it is you whose guilty of insider trading, not the 20 employees. Your just talking about your old job trying to get hired.
Kirby
In essence, write up your resume however you like, ignoring the NDA. (My suspicion, is that either, your NDA is uninforceable if you can violate it in less then the 3 bullet points, and 1 paragraph description of what your job is, or you are writting your Resume incorrectly). If what you do is standard industry practice, nearly every NDA I've ever seen says "anything that is standard industry practice is not covered by this NDA". It's completely uninforceable otherwise. They pay you to tell them common knowledge, the then hold you hostage to never give up that information again for the term of your NDA.
No mister roofer, you put shingles on my roof. For the next 3 years, every roof you build can't have shingles, otherwise I get to sue the bejesus out of you. You can't even tell the next guy you work for, that you know how to apply wooden squares to a angled board to keep water from hitting you. Nope, I've got an iron clad NDA right here. Uh-huh. Judge would throw you out on your ear withing 15 minutes of hearing the arguments for doing that.
For goodness sakes, your describing what's already known to the general population in the industry. No offense, but your not doing anything terrible secret in the area of RDBM's database optimation. Most of the really cool stuff was figured out 15 years ago, and they've been waiting around for the hardware to catch up. Lets see, schema layout, indexing, denormalizing data, princepal of locality (temporial and spacial), and instrumentation/profiling. Everything else probably came straight out of the vendor documentation, and was applied to a particular database. Optimation comes down to "time this, change param A, time again, change A some more, lather, rinse, repeat. Move to parameter B". Maybe you've figured some cool stuff out about ODBM's or something. However, even most of that is just following the bouncing ball after reading the documentation from the Vendor.
If they come after you, just explain to them, you'll take your six figure settlement now, or you'll find a nice attorney who'll work on contingency for 1/3 of the 7 figure settlement after the trial.
I'm not a lawyer, this isn't legal advice. However, they can't compel you to give up your right to earn a living. Barring a matter of national security at least.
Kirby