Ah, but the only reason that.biz was illegal was because they took money from people without giving back anything in return. Obviously you would only charge the winner of the lottery in my example;.biz was charging everybody just for a chance in the lottery. Big difference.
The reason that I'm not convinced that it's just Verisign is because they haven't been in control of the registry for that long (has it been a year already?) and so the previous operator, NSI, should have cut those domains loose long ago. Although Verisign seems to be continuing along with NSI's poor operations in this regard. Maybe that's a different part of NSI than NSI-the-registry, but I'm not incredibly confident that NSI can maintain a wall between those parts of their company. Their other actions have not been consistent with divorcing the interests of the Registry and the registrar parts of the company.
Where is the load actually coming from? I can't imagine that it is so much of a load to remove a site from their database, therefore it must be the registrar query traffic. Handling this traffic is therefore the priority.
It's understandable that they can't selectively deny access to a registrar, but that just means that they can't deny one registrar while allowing another. That doesn't mean that they have to allow themselves to be essentially DOS'd by those registrars. A temporary ban based on exceeding an allowed query limit per IP address would be a reasonable approach. Even better, allow the registrars to sign up for notification of a certain domain deletion, and then send a broadcast mail. Change the load from a multiple-client-pull to a server-push pattern.
Hmmm, I see your point now - registrars aren't querying for the same name multiple times a minute, they're querying for various different ones. But the mailing list approach would still work. Or you could design a "ticker" app that would have lower latency than email.
For highly-demanded names, maybe divvy them out like concert tickets - take applications from everyone within a half-hour window, and then randomly pick a winner out of that lot. That would take the guesswork out of it for the registrars, thus decreasing their loading of the system with spurious queries.
But whatever they do, Verisign needs to get the names back out there if they're not going to be renewed. Waiting a year or two before putting it back in circulation is a travesty.
I think the point is that it doesn't take a whole day to scan through the config file or look in the docs, and while you're doing actual reading up on the server rather than just doing pointy-and-clicky, you might actually learn something about how it works. Something that will come in handy when you have security problems, or something doesn't work quite right, etc.
The big argument for me against GUI configuration, at least as provided by Microsoft, is that it makes it easy to make big decisions without understanding those decisions. This may make managers and monkeys happy, but it's a mistake in the long run. Configuration of complex things should be complex, or should at least instill into the user the understanding that they are doing a lot more than they think they are.
Now, if the config file is excessively convoluted (moreso than it needs to be) then that's a problem. Sendmail may fit that mold, but httpd.conf, with its abundant comments and consistent syntax, does not.
When a domain
expires it is actually automatically renewed by the registry.
But they're creating their own headaches by doing this! Just delete the name every time, without exception and without fanfare, and let the other registrars figure out that it's gone when it's gone. Better yet, randomize the release time over, say, a week or two, and IP ban anyone who hits the site more than once or twice a minute. This is not exactly rocket science here, you know. If someone really did want to renew it, they're maybe entitled to one or two warning emails and maybe a week's grace period, but after that delete the domain and let them scramble for it.
There are domains that are still registered which have expired more than a year ago, so somebody's not doing their job right. The public shouldn't be denied the right to get those domains just so that NSI/Verisign can coddle a few ruthless domain speculators. And if they had some leadership in realms other than money-grubbing, maybe NSI/Verisign would figure some of this out.
Unity is good. But a unified, centralized root server system run by unscrupulous frauds got us into this mess in the first place. Are you willing to bet that the mistake-that-is-NSI will never happen again?
My prediction: in 10 years DNS is obsolete. It will be replaced by the search-engine-name-system, where you ask your PDA's search engine where to find such-and-such a company, and it sends you to their site. Domain names are just a crutch to find the site; by then we'll have much better crutches.
Of course, at that point there will be lots of squabbles over who gets listed first by which search engine, etc. It's always some damn thing:)
Bingo. It's ridiculous that they are setting this up to handle existing names that haven't expired yet, when there are names which have already been expired for one or two years which cannot be claimed due to various registrars' screwed-up policies.
The whole name registration racket is in dire need of either total decentralization (to empower the customer) or else some real regulation to make sure that all registrars are playing by the same rules. Since I'm not too confident in ICANN's regulation so far, decentralization sounds like the way to go.
Heck, I'd love to see the Commerce Department (or an international disinterested party (you know, like ICANN was supposed to be?)) take back over the actual database, and provide the same access to all registrars alike. As it is now, any one of NSI's bad business ideas are basically unstoppable without a significant court battle.
I don't think you can complain about inaccuracy while your post is inaccurate as well. There are no existing U.S. breeder reactors, so anything that's the result of a breeder reactor will not be stored at Yucca Mountain.
Thanks for your comments based on your experiences, although so far they haven't been too much help:(
There is no such product as Visual Studio 2000. There is Visual Studio 6. It came out much before
Windows 2000 time.
My mistake, it is VS6. I wasn't sitting in front of it when typing before. Not making it "2000" like everything else really seems a failure of marketing/branding, don't you think?
COM components can easily be "uninstalled", as in, unregistered and then deleted. COM+ is the same
way. The only reason you wouldn't be able to "uninstall" them would be if there were in-use - IE you
had a VS component open, a debug window open, OR an errant thread.
It wasn't a threaded application, so I don't think it was that. In those other cases, the system should tell you to exit those, rather than just telling you to reboot, wouldn't you think? One would hope, at least.
For network changes I can't think of anything that causes a reboot other than joing or disjoing a
domain, changing the machine name, or changing hardware settings on very old ISA devices.
In fact, IIRC I was changing the netmask and ip addresses associated with one NIC
You say unpatched, which is of course stupid. If you really mean that then you missed about 2 hrs of
patching to get your system "up to date". You need SP2 for Windows, SP2 for Office, and SP5 for VS.
Making sure your development machine has the lastest release quality of code is very important to stable
Windows development.
I mean "unpatched" as in it's what our local Windows dev gurus told me to use, a stock install of Win2k and some version of Visual Studio. I have no idea what SPs that includes. I meant "unpatched" in sense that it's not a patched-up Linux kernel, for instance.
It's true that in both directions, anecdotal arguments aren't particularly convincing. I'm just lashing out in my frustration at all those people who told me that a newer version of Windows would actually Work Right (tm). In my experience, this hasn't been the case, so unfortunately a world of statistical evidence to the contrary doesn't really sway me.
Although I can't really complain about the server uptime, at least (yet) - using this as a development platform, it's been rebooted quite often enough for other reasons, installs, etc. that there hasn't been a chance for Windows itself to fail:)
OK, I could see gnome getting whacked enough that an X login wouldn't work, at which point it would be necessary to switch to a text terminal and revert back to a good gnome build, returning to run level 5 only when that's done. But I still don't see the need to reboot anywhere in there.
Nope, this was an entirely stock Windows 2000 install with Office 2000 and Visual Studio 2000. The reboots have occurred while doing standard network configuration changes, and installing/uninstalling COM components. Nothing patched, no funky hardware, etc.
And if you had to reboot to do anything with Gnome I'd be very surprised - I can't remember the last library upgrade on my Linux machine that required a reboot.
Sorry to pick on you, I'm having a non-fun Windows-using day today:)
Still, I wasn't trying to do anything a normal user wouldn't do, though - just removing the items that Microsoft crams onto the task bar by default. If doing that is enough to crash the damn thing, then I think it's not stable enough for a normal user either.
My point is that I feel that my recent experiences belie the oft-made observation that "But Windows is so much better now!". As far as I can tell, I'm rebooting about as much as in Windows 95 or NT, and it doesn't seem any more stable. So the moral, for me at least, is to not take the constant assertations of Windows' vast improvement at face value.
Fact: all the Microsoft trolls said that Win2000 was sooo much better than Windows NT.
Fact: In my first five minutes of using Windows 2000 the other day, I managed to lock up the frickin' task bar at the bottom of the screen, and had to logout and log back in to fix it. That was only the most immediate problem that I've found in my unfortunate journey into developing on Win32.
Truth: this crap isn't any more stable, it just looks a little prettier and sometimes survives a little longer before it goes belly-up. I don't know why I expected Microsoft apologists to actually tell the truth for once about its stability - I won't make that mistake again:)
Win2000 reboots today in order to alter network settings and add/remove VB components: ~10
Reboots ever required to alter network settings and add/remove components on Unix systems: none that I can recall, EVER.
No, I haven't told anybody anything new either, unfortunately...
That is so true - the last several Microsoft viruses and worms that hit my company didn't do anything to all those NT machines, due to their multi-user permissions and all. We were totally spared from the depredations of all of those viruses, and lost no time or money.
;)
I agree with your point that it's the outcome that really matters, I just disagree that somehow less damage is done on NT. It seems to be plenty vulnerable from here.
If you have ~/bin or . in your PATH before/bin, and a malicious ~/bin/su or./bin/su is created, then yes, it could snag your password, use it to su, and do whatever with it.
There's a word for your clients: backups. Oh wait, I see you already know the word:)
Actually, with a named pipe you could probably do this for certain files - you think you're writing to a file, but really you're handing off to another daemon that actually stores the file elsewhere. If you try to delete the file, you just lose the pipe, not the file itself.
Or...the British Government? Some of their sites are already unusable without IE 5+, do you really think they'd have a problem making government services available only through.NET if Microsoft gave them enough of a break on it?
Actually, John Does get sued all the time, with the names filled in once more investigating has been done. Isn't that what happened with some of the DeCSS defendants?
There's a difference between freedom and safety. The statistics do seem to bear out that for the average person, gun-owning may not increase their safety. It's a lot harder to measure the amount that maintaining a well-armed citizenry preserves their freedom.
I suspect I am not the only person finding the comparison to the civil rights protests disgusting. There is
a big difference between getting arrested for drinking at a colored water fountain and getting arresting
for ripping off the latest U2 or Dire Straits tracks.
If you want to listen to music then pay for the damn stuff.
Actually, I sort-of agree with that, since I'm a non-mp3-trading person who's computers would still be affected by the proposed SSSCA. If the folks who just don't believe in copyright would STFU, then the rest of us who give copyright some credence but still like to have fair use and control our own hardware might be in better shape today.
However, the original comment was "what, and go to jail for your beliefs?" to which I was pointing out that yes, that's exactly what's supposed to happen if you truly believe those things. The difficulty of standing up for what you believe in is first a mental and spiritual challenge; once you've decided that you will stay the course, the it doesn't matter if you're going to jail for mp3s or for drinking from the wrong water fountain.
Although I will be the first to admit that so far the civil rights fighters have suffered a lot more for their beliefs than the mp3 traders. Hopefully no one will have to die for their beliefs on the subject of copyright.
Not that I don't sympathize with those folks, and feel that there's no reason to imprison or penalize them, but even that isn't reaching to the level of civil disobedience that's needed. Maybe if every pot smoker in America (of which I am not one) would go downtown, light up in front of city hall, and dare the government to throw them in jail, there would be a change. It really takes a lot of people to affect change. So far the drug resistance has not met that critical mass nationally, although in some states it's probably there or pretty close.
I fail to see the distinction - civil disobedience is by definition disobeying a law because you disagree with it. Sitting at that lunch counter and being arrested was breaking the law and it was civil disobedience. Not sitting at the back of the bus was the same thing. Many Jim Crow laws were upheld in courts at some point, so breaking them was disobeying a court order.
I agree that you can't just dismiss court decisions, and in fact I didn't say that you should. I just said that you should stand up for your beliefs, plan to embrace the consequences of your actions, and think about whether the consequences are really worthwhile to you.
Slashdot Mirror
And your inability to learn from your mistakes affects the rest of us how? I mean, other than providing a morning chuckle or two?
Maybe they left out of disgust over a site that requires Flash? All those years pursuing a doctorate must have taught them something :)
On the plus side, having a Flash intro is probably a good indicator of all the smoke-and-mirrors waiting within...
Ah, but the only reason that .biz was illegal was because they took money from people without giving back anything in return. Obviously you would only charge the winner of the lottery in my example; .biz was charging everybody just for a chance in the lottery. Big difference.
The reason that I'm not convinced that it's just Verisign is because they haven't been in control of the registry for that long (has it been a year already?) and so the previous operator, NSI, should have cut those domains loose long ago. Although Verisign seems to be continuing along with NSI's poor operations in this regard. Maybe that's a different part of NSI than NSI-the-registry, but I'm not incredibly confident that NSI can maintain a wall between those parts of their company. Their other actions have not been consistent with divorcing the interests of the Registry and the registrar parts of the company.
Where is the load actually coming from? I can't imagine that it is so much of a load to remove a site from their database, therefore it must be the registrar query traffic. Handling this traffic is therefore the priority.
It's understandable that they can't selectively deny access to a registrar, but that just means that they can't deny one registrar while allowing another. That doesn't mean that they have to allow themselves to be essentially DOS'd by those registrars. A temporary ban based on exceeding an allowed query limit per IP address would be a reasonable approach. Even better, allow the registrars to sign up for notification of a certain domain deletion, and then send a broadcast mail. Change the load from a multiple-client-pull to a server-push pattern.
Hmmm, I see your point now - registrars aren't querying for the same name multiple times a minute, they're querying for various different ones. But the mailing list approach would still work. Or you could design a "ticker" app that would have lower latency than email.
For highly-demanded names, maybe divvy them out like concert tickets - take applications from everyone within a half-hour window, and then randomly pick a winner out of that lot. That would take the guesswork out of it for the registrars, thus decreasing their loading of the system with spurious queries.
But whatever they do, Verisign needs to get the names back out there if they're not going to be renewed. Waiting a year or two before putting it back in circulation is a travesty.
I think the point is that it doesn't take a whole day to scan through the config file or look in the docs, and while you're doing actual reading up on the server rather than just doing pointy-and-clicky, you might actually learn something about how it works. Something that will come in handy when you have security problems, or something doesn't work quite right, etc.
The big argument for me against GUI configuration, at least as provided by Microsoft, is that it makes it easy to make big decisions without understanding those decisions. This may make managers and monkeys happy, but it's a mistake in the long run. Configuration of complex things should be complex, or should at least instill into the user the understanding that they are doing a lot more than they think they are.
Now, if the config file is excessively convoluted (moreso than it needs to be) then that's a problem. Sendmail may fit that mold, but httpd.conf, with its abundant comments and consistent syntax, does not.
There are domains that are still registered which have expired more than a year ago, so somebody's not doing their job right. The public shouldn't be denied the right to get those domains just so that NSI/Verisign can coddle a few ruthless domain speculators. And if they had some leadership in realms other than money-grubbing, maybe NSI/Verisign would figure some of this out.
Unity is good. But a unified, centralized root server system run by unscrupulous frauds got us into this mess in the first place. Are you willing to bet that the mistake-that-is-NSI will never happen again?
My prediction: in 10 years DNS is obsolete. It will be replaced by the search-engine-name-system, where you ask your PDA's search engine where to find such-and-such a company, and it sends you to their site. Domain names are just a crutch to find the site; by then we'll have much better crutches.
Of course, at that point there will be lots of squabbles over who gets listed first by which search engine, etc. It's always some damn thing :)
Bingo. It's ridiculous that they are setting this up to handle existing names that haven't expired yet, when there are names which have already been expired for one or two years which cannot be claimed due to various registrars' screwed-up policies.
The whole name registration racket is in dire need of either total decentralization (to empower the customer) or else some real regulation to make sure that all registrars are playing by the same rules. Since I'm not too confident in ICANN's regulation so far, decentralization sounds like the way to go.
Heck, I'd love to see the Commerce Department (or an international disinterested party (you know, like ICANN was supposed to be?)) take back over the actual database, and provide the same access to all registrars alike. As it is now, any one of NSI's bad business ideas are basically unstoppable without a significant court battle.
relevant UserFriendly cartoon
Any UF-ites able to explain why using the "cartoon search" for "gates" didn't turn this up, so that I had to look for it the hard way?
I don't think you can complain about inaccuracy while your post is inaccurate as well. There are no existing U.S. breeder reactors, so anything that's the result of a breeder reactor will not be stored at Yucca Mountain.
Thanks for your comments based on your experiences, although so far they haven't been too much help :(
My mistake, it is VS6. I wasn't sitting in front of it when typing before. Not making it "2000" like everything else really seems a failure of marketing/branding, don't you think?
It wasn't a threaded application, so I don't think it was that. In those other cases, the system should tell you to exit those, rather than just telling you to reboot, wouldn't you think? One would hope, at least.
In fact, IIRC I was changing the netmask and ip addresses associated with one NIC
I mean "unpatched" as in it's what our local Windows dev gurus told me to use, a stock install of Win2k and some version of Visual Studio. I have no idea what SPs that includes. I meant "unpatched" in sense that it's not a patched-up Linux kernel, for instance.
It's true that in both directions, anecdotal arguments aren't particularly convincing. I'm just lashing out in my frustration at all those people who told me that a newer version of Windows would actually Work Right (tm). In my experience, this hasn't been the case, so unfortunately a world of statistical evidence to the contrary doesn't really sway me.
Although I can't really complain about the server uptime, at least (yet) - using this as a development platform, it's been rebooted quite often enough for other reasons, installs, etc. that there hasn't been a chance for Windows itself to fail :)
OK, I could see gnome getting whacked enough that an X login wouldn't work, at which point it would be necessary to switch to a text terminal and revert back to a good gnome build, returning to run level 5 only when that's done. But I still don't see the need to reboot anywhere in there.
Nope, this was an entirely stock Windows 2000 install with Office 2000 and Visual Studio 2000. The reboots have occurred while doing standard network configuration changes, and installing/uninstalling COM components. Nothing patched, no funky hardware, etc.
And if you had to reboot to do anything with Gnome I'd be very surprised - I can't remember the last library upgrade on my Linux machine that required a reboot.
Sorry to pick on you, I'm having a non-fun Windows-using day today :)
Still, I wasn't trying to do anything a normal user wouldn't do, though - just removing the items that Microsoft crams onto the task bar by default. If doing that is enough to crash the damn thing, then I think it's not stable enough for a normal user either.
My point is that I feel that my recent experiences belie the oft-made observation that "But Windows is so much better now!". As far as I can tell, I'm rebooting about as much as in Windows 95 or NT, and it doesn't seem any more stable. So the moral, for me at least, is to not take the constant assertations of Windows' vast improvement at face value.
Fact: all the Microsoft trolls said that Win2000 was sooo much better than Windows NT.
Fact: In my first five minutes of using Windows 2000 the other day, I managed to lock up the frickin' task bar at the bottom of the screen, and had to logout and log back in to fix it. That was only the most immediate problem that I've found in my unfortunate journey into developing on Win32.
Truth: this crap isn't any more stable, it just looks a little prettier and sometimes survives a little longer before it goes belly-up. I don't know why I expected Microsoft apologists to actually tell the truth for once about its stability - I won't make that mistake again :)
Win2000 reboots today in order to alter network settings and add/remove VB components: ~10
Reboots ever required to alter network settings and add/remove components on Unix systems: none that I can recall, EVER.
No, I haven't told anybody anything new either, unfortunately...
That is so true - the last several Microsoft viruses and worms that hit my company didn't do anything to all those NT machines, due to their multi-user permissions and all. We were totally spared from the depredations of all of those viruses, and lost no time or money.
;)
I agree with your point that it's the outcome that really matters, I just disagree that somehow less damage is done on NT. It seems to be plenty vulnerable from here.
If you have ~/bin or . in your PATH before /bin, and a malicious ~/bin/su or ./bin/su is created, then yes, it could snag your password, use it to su, and do whatever with it.
There's a word for your clients: backups. Oh wait, I see you already know the word :)
Actually, with a named pipe you could probably do this for certain files - you think you're writing to a file, but really you're handing off to another daemon that actually stores the file elsewhere. If you try to delete the file, you just lose the pipe, not the file itself.
Or ...the British Government? Some of their sites are already unusable without IE 5+, do you really think they'd have a problem making government services available only through .NET if Microsoft gave them enough of a break on it?
Guns don't kill people, the Government does :)
Me, I press the conveniently-labeled "down arrow" key, but I hear you have to have a fancy newfangled 101-key keyboard to get that trick to work...
Actually, John Does get sued all the time, with the names filled in once more investigating has been done. Isn't that what happened with some of the DeCSS defendants?
There's a difference between freedom and safety. The statistics do seem to bear out that for the average person, gun-owning may not increase their safety. It's a lot harder to measure the amount that maintaining a well-armed citizenry preserves their freedom.
Actually, I sort-of agree with that, since I'm a non-mp3-trading person who's computers would still be affected by the proposed SSSCA. If the folks who just don't believe in copyright would STFU, then the rest of us who give copyright some credence but still like to have fair use and control our own hardware might be in better shape today.
However, the original comment was "what, and go to jail for your beliefs?" to which I was pointing out that yes, that's exactly what's supposed to happen if you truly believe those things. The difficulty of standing up for what you believe in is first a mental and spiritual challenge; once you've decided that you will stay the course, the it doesn't matter if you're going to jail for mp3s or for drinking from the wrong water fountain.
Although I will be the first to admit that so far the civil rights fighters have suffered a lot more for their beliefs than the mp3 traders. Hopefully no one will have to die for their beliefs on the subject of copyright.
Not that I don't sympathize with those folks, and feel that there's no reason to imprison or penalize them, but even that isn't reaching to the level of civil disobedience that's needed. Maybe if every pot smoker in America (of which I am not one) would go downtown, light up in front of city hall, and dare the government to throw them in jail, there would be a change. It really takes a lot of people to affect change. So far the drug resistance has not met that critical mass nationally, although in some states it's probably there or pretty close.
I fail to see the distinction - civil disobedience is by definition disobeying a law because you disagree with it. Sitting at that lunch counter and being arrested was breaking the law and it was civil disobedience. Not sitting at the back of the bus was the same thing. Many Jim Crow laws were upheld in courts at some point, so breaking them was disobeying a court order.
I agree that you can't just dismiss court decisions, and in fact I didn't say that you should. I just said that you should stand up for your beliefs, plan to embrace the consequences of your actions, and think about whether the consequences are really worthwhile to you.