Also, I think a little less bias in the selection of the stories themselves could be useful. Who on earth made the decision that the release of FreeBSD 4.0 wasn't even worthy of a mention as a "quickie,"
Perhaps you should read a little more of slashdot in future, to catch information about FreeBSD 4.0? Or, if you're more specifically interested in BSD articles only, maybe just read all the articles about BSD? After all, the information is quite available to you.
Yeh, your right, what was I thinking. After All, I've never done any pure assembly coding for windows only pure-DOS! What would I know. The only programming I've done for windows has been in C++! Good for you. I'll admit, I've never coded assembly for windows (and strenuously avoided assembly for DOS). Never found it very enjoyable.
WTF? You think I don't know what msdos.sys's BOOTGUI option does? Do you think I'm a moron? Let me ask you, what the hell does that have to do with whether or not windows is a shell for DOS, and why the hell wouldn't I just hit the F8 key at boot? BeOS starts in MacOS, NT can run in Linux, and Linux can run in NT. Most people don't know about it, or what it does. So, yes, until you've shown otherwise, I assume you don't know.
As for why not just hit F8, the reasponse could simply be: "Well, that's windows just loading command.com at the end, instead of completing the startup." By editing that file, it can not be simply running command.com at the end. Something else must be making it stop.
Lets play another game. Find the file called Loadlin.com on your distro's CD. I know Red Hat has it. Copy the Linux kernel to a DOS partition and type "Loadlin.com Linux.krl or" whatever. What happens? Why? Linux boots from DOS!!!! I guess that means Linux is just a shell for dos! I'm sure Linus will be surprised.
Boring game. Played it in the past, even got the t-shirt somewhere. Had to use it until the AWE32 drivers were done enough that I didn't have to use the dos utilities to initialize the card. Did it for a year or so, roughly. Anyway, the point remains the same: DOS happens well before windows ever pops up. And, from the evidence I've seen, the relationship between the two is still the same: Windows 9x is nothing but a shell which runs on top of DOS.
Oh? What's that? You've never heard of loadlin.com? I guess you must not know very much about Linux then. Actually, you're right. I don't know much about Linux. However, in making that statement, you have to remember that I compare myself to people like Linus, Alan Cox, et al.
As for why I said you've never poked around in that section of the Windows internals before: Based on the evidence I've seen, Windows is a DOS shell. And with a complete lack of information to prove otherwise by you in your original post, I had to assume you'd never before seen the BootGUI line.
Windows 9x does not run on a dos Kernel, the main system controler is vmm32.vxd, I think and it is not based on anyway on Legacy dos code. NT has its own Kernel as well. Just being DOS compatable does not make an operating system a version of DOS. Infact Linux is probably more DOS compatable (with DOSemu) then NT. Now, here's where I have to differ with you. Since you've never dug into the internals of Win9x (or, at least, this piece of the internals), I'm going to give you step by step directions on how to see that Win9x is still just a shell over the top of DOS, just like Win3.1.
attrib -r -h -s c:\msdos.sys
edit msdos.sys with your favorite text only editor
Change the line which reads "BootGUI=1" to read "BootGUI=0"
Save the file
attrib +r +h +s c:\msdos.sys
Reboot your PC
Wait until you get your familiar old DOS prompt back.
Type in "WIN", and press
As you can see, it's not very complicated. Win9x is nothing but a DOS shell, just like Win3.1. The only important difference is that the new version of DOS (version 7.0 as of Win95) automatically loads the shell, whereas with Win3.1 it required extra action to make sure it was loaded either in the form of the install program appending to autoexec.bat, the user doing so, or the user typing "win" every time he wanted to get into windows). Win9x is only the newest version of DOS. WinNT is not, and doesn't use a DOS kernel. Win9x has a DOS kernel.
I'm a programmer. I dream of writing the next big thing to change the world(tm). But, even though my knowledge goes into the writing of the program, the program itself is absolutely worthless, even to me.
My knowledge is what's valuable. My knowledge of how the program works is what's valuable. I'll give away the program for free. Essentially, it's worthless anyway. But, if you want something specific done with it, then we'll start talking about my making money off of it.
Actually, I'll be surprised if this ever gets replied to, due to the age of the article, but what the hell. Here goes:
Richard Jewel was the gentleman who found the bomb at the Atlanta Olympics. Was subsequently investigated as having probably placed the bomb. And then was exonerated by the same investigation. The FBI never apologized for what happened to Jewel (much harassment by the media for an FBI leak), nor what happened to his career. You see, once cleared, the media never bothered to mention that he had been. So, for many people who remember the name, Richard Jewel is that guy who placed the bomb at the Olympics. Even though he didn't, people never got the chance to hear that. And the FBI still doesn't care that this man's livelihood was trashed by them. He's just another nobody to them.
As for why he was mentioned: It's another example of FBI abuse of their powers.
Actually, here's a scenario which could explain their patenting an item: Control.
By patenting it, they guarantee themselves exclusive control of the ideas in the patent for the next couple decades. It prevents somebody else from coming up with and using it on their own.
Why would they need to prevent this? I honestly don't know. But it's a thought, at least.
These questions are related to my others, but are distinct enough to deserve a second posting. Judge Jackson has found that MicroSoft does enjoy monopoly power in the market. He has also found occasions where they have abused this power. It is likely that they will be found to have broken the law, and punishments will be meted out to Microsoft. Now comes the very important part, deciding what to do:
They could have their IP rights stripped from them.
They could have massive fines (to the tune of billions of dollars) assessed.
They could have all existing contracts nullified, with no legal ability to generate new ones for a time. Any licensing which they wish to do would be overseen by an independent body.
They could be broken up into a bunch of BabySofts (BS:)
Of these options, which do you feel are likely to be taken, and why (or why not)? Furthermore, which of these options should be taken, and which shouldn't? Again, why on all of that.
I'd like to thank you all for taking the time to do this. I know enough of the law to get myself in trouble, and appreciate the opportunity to hear from people who have actually studied these areas of law.
I've just got too many questions to ask, but I will try and keep them to as few as possible.
1) I've heard that the finding of fact is pretty much untouchable, and I've heard that it can be moderately easily overturned on appeal. Can you tell me how easily overturned this would be, and why?
2) Judging from the Finding of Fact (and economic factors), how harsh do you believe a final verdict from Judge Jackson will be (assuming that the final verdict is that Microsoft did, indeed break the law)? Not in terms of the details (what fines, etc), but rather in terms of "Wrist Slap" or "MS will need to file for Chapter 11" (and everything in between) type of terms.
3) I've also heard that MicroSoft would be very interested in settling before a final ruling by Judge Jackson. Failure to settle before then would mean that the Finding of Fact would stand as is, and be admissible in other courts as evidence against MicroSoft. This is actually a multi-part question, but here goes: a) When this case goes to the Supreme Court, the justices have the option of overturning any verdict from Judge Jackson (which is likely to be hard on Microsoft, from what I and others have read). Will the Findings of Fact still be admissible in other courts? b) Is this document admissible as evidence until then? c) Will it be admissable in case of a settlement?
Sorry to have asked so many, but they are all closely related, and I felt belonged in the same post as one group of questions.
God, you guys are touchy. All I was saying was that TCO is important. Apparently no one read far enough to see this in the original post:
Actually, I did read that in the original post. Of course, immediately after stating that, you came back with how you need a Linux bartender who would probably be more expensive than the Windows bartender. Here's the quote, in case you've forgotten:
Linux may be cheap, but it ain't free (as in beer) in the sense that you still need a bartender. And the truth is, for a Windows shop, it may actually be more expensive, given retraining costs and the tightness of the labor market in technical areas.
That, to me at least, indicates that TCO for Windows is lower, at least sometimes. And that's the point I'm disputing. I don't honestly believe that Windows TCO can be lower than Linux/*BSD/other UNIX TCO. The various Unices out there, once set up properly, rarely need work. The various flavors of Windows need constant attention and babying. Without it, they tend to self-destruct after a few months. Not a good thing to have at all.
I think some Linux advocates would do better if they didn't act like a bull seeing red every time the see the word Microsoft.
And this is something to which I take direct offense. I have actually advocated Microsoft in the past, and will do so again, under the right circumstances. However, one of they very few times I will do so is if the computer will see heavy games usage, which is something that Linux (and other Unices) don't have yet. Outside of the games, though, a properly set up Unix box will be easier to take care of than a properly configured Windows box. Especially if I'm in California, on a dialup, and the PC with the problem is in New York, also on a dialup.
I don't know enough about autorpm to be able to comment correctly about it. apt-get, though, I do know somewhat at least (though by no means am I an expert). I commented on what I knew about for that, and just realized that how I said it could be taken as a slight against RedHat, even though I didn't intend to.
Please, compare apples to apples. RedHat is only one option, and hardly the best for considering automatic updates.
Install Debian, do it once, and do it right. Build an image of the hard drive (using tar, to prevent problems with different sized hard drives), and a quick script later, and you've got multiple, automatic installs without any effort. And they work at least as well (if not better) than any automated install script from M$.
Next, install a local mirror of the Debian distributions. Every night, at say 9pm, it goes and downloads the latest and greatest packages (using apt-get), and gets them installed correctly on itself.
Next, each of the client machines connects to the local mirror (at say, 3am), and does the same thing. You could even throw a reboot in at the end, if you really wanted to. And now the machines are automatically updated (as much as they need to be) on a daily basis. With no human intervention whatsoever.
Okay, let's first kill a few quick myths in THAT post.
First, you don't have to pay somebody to download it. You order it from Cheapbytes. With shipping, you're still under $10.
Second, you don't have to burn multiple copies of the CDs to install to multiple machines simultaneously. You install to one machine, set up a proper NFS server (one option), FTP server (another option), or HTTP server (third option), and you can now install to as many client machines as you want via network (very easy to do, by the way, from my experience).
Third, you don't have to pay somebody to stay at your site constantly. Linux machines, once set up correctly, need much less babying than Windows machines. And you won't ever have to reinstall Linux, unless you have a hard drive crash. All in all, you pay somebody to set up your network (most businesses could be done in under a month, some in under a week), and then pay him to check up on things every so often.
Fourth, your retraining cost is going WAY down, since you've got Star Office, Gnome, KDE, and other Office suites on the way. People will have no problem figuring them out.
TCO is a major thing to consider. But don't cook the numbers. They're quite yummy raw, and lose much of their value once cooked.
1 RedHat Secure Server CD (for all 8 servers), $140
Upgrade or new install, total cost: $210
That's almost affordable enough for me to do on my own, without corporate sponsorship. Any company should be able to afford that quantity very easily, or it won't be up to a medium sized network.
How strange. I find Debian very easy. Unlike the other distro's alot is automated for you. 'update-menus' leaps to mind.
I agree with you wholeheartedly. Using Debian is much easier than other distros. Maintenance can even be automated to a better degree, if you choose (cron, anybody?). However, the very initial installation of Debian is much more difficult than others, I think. It's getting over that hill of "Let's get it installed and working" that is so hard. Once past that, though, it's very easy to get along with.
I'm starting to think there is more to the 'technical/non-technical personality' issue than I've generally accepted, and this explains the differences in perceptions of Debian's installation process.
Would you mind elaborating on this? I consider myself to be a moderately technical person (people around me look at me funny when I say that. But I know who I compare myself to, so I am only moderately technical), and I was stymied by the dselect process at first. It's not so bad anymore, and apparently apt can now do multi-cd, so it should be even easier. Anyway, I'm rambling, so I'll just ask.
What is the "more" that you've noticed to that issue?
Ah, sorry about that. I'd gotten the impression that a journalist might have done some actual research, and gone to the Debian home page. Of course, now that I think about it, a journalist doing research is roughly equivalent to MicroSoft producing a stable, reliable operating system and being honest about any security issues (and even fixing them quickly). Sure, it could happen, but most likely not in this lifetime.
Again, sorry about that. Shoulda thought before I posted that.
I'll speak up as a person who is a fan of maintaining a Debian installation, but who had awful problems installing it the first time.
I bought the Cheapbytes Debian CDs, and was working with them. I had no problems until dselect came into the picture. After that, it was all downhill, for about a week. It took me about an hour of trying (and reading dselect help) to figure out that I was supposed to put CD 2 in, run update, and then put CD 1 in for the install. If I'd caught that up front, the entire process would have been much less painful. In addition, the interface for dslect is very very far from easy to follow (at least for me).
Now some notes about this experience: I've installed Slackware, and several versions of RedHat, all without a hitch (well, except for the first time I installed Slackware. That was due to total inexperience with setting up an OS, though). I'm not exactly a computer newbie, either. I've been working (and playing) with them since the old days of the VIC-20. Just the interface for dselect lacks a lot in terms of usability.
Now, I know that I could have just settled for one of the profiles, but I wanted to actually see what packages would be installed, and maybe modify the package listings. That process was about three hours of reviewing the listings, and determining if I wanted a package or not. Too many packages, I think.
Next issue was during the install stage of dselect. A few hours worth of package configuring (hindered by the fact that at least a third of the packages wanted me to help them out, and didn't have defaults). All in all, an extraordinarily frustrating day.
Finally, I had two other issues, but they can't be blamed on Debian itself. CD 1, it turns out, was a bad CD (meaning it had some bad sectors), so once those failed, making dselect go again (I thought) meant restarting from scratch. That was a very very frustrating experience, I can tell you. As luck would have it, I also had a hard drive (brand new) which had a bad spot on it, and had to be replaced.
All in all, not an easy installation experience, and one that I would steer newbies away from as much as possible.
My $0.02, YMMV, <insert standard disclaimer here>.
Well, of course it referred to GNU/Linux. The proper name for Debian is "Debian GNU/Linux" (their page at http://www.debian.org makes this very clear).
Yes, Suse is available in a box. As is slackware, Mandrake, and certain others (I just know I've seen others, but can't remember the names).
I loved Opera under Windows. Fast, small, easily downloadable. What more could you ask for?
Well, a few kept promises would be a good thing, I think.
Project Magic will be updated monthly. Big surprise, it's updated every two months, unless somebody (like me) leaves them a nastygram in the guestbook telling them they haven't done so in two months.
In April, we were told that at least a beta for linux would be released in 2Q, 1999. Shocking! We've got nothing other than a couple of (easily faked) screenshots of a text based browser!
I've given up. Mozilla still isn't ready for even beta use, Opera is dead (IMO), and I've not yet tried Amaya (though that will be fixed later today). Netscape is buggy (at best). No good browsers for Linux yet, and none on the horizon.
I give up, and will simply stick with Netscape for now.
Unfortunately people who aim for total consistency are rather difficult to argue with, unless you have an equally consistent opposing viewpoint, and often frighteningly persuasive.
Now, maybe I'm reading more into this than I should. But what's wrong with aiming for consistency? Quite often, the people who have a consistent basis for their beliefs (and act accordingly) are among the easiest to deal with. Even the Bible-thumping Christians are easy enough to deal with, in part because you can understand their beliefs. You may not like them, but you know where they're coming from, and why they're acting that way.
Personally, I see nothing wrong with consistency, but maybe I'm missing something important.
Ya know, I can only believe this is pure flamebait, so I'm an idiot for responding, but your description shows zero knowledge of how digital signatures work.
Digital signatures, on the whole use public and private keys. These public and private keys are unique numbers, somewhere on the order of a few hundred digits long (usually). In order for a packet to be signed, that packet must have access to the private key. In order for a packet to be verified, the receiver must have access to the public key.
Now, think about what's been said:
Keys are unique numbers, several hundred digits long.
Both private and public keys are reuired for signatures to work (signing and verification).
It follows, then, that the chip must have both the public key, and the private key, on it at the same time.
Backtracking through definitions, we see that the private key is a unique number, and it must be embedded into the chip.
Now, the number isn't a true "serial number", simply because it doesn't count up in order (in fact, due to other facts, not mentioned here), it can not count up.
Instead, we have something even better: a unique, cryptographically secure (supposedly) identifier attached to each and every computer which Big Blue sells. If there is a backdoor in these chips, then the government will now have a way of tracking, and reading, everything which gets encrypted/signed by these chips.
Well, the main point is not the cipher itself, but rather the public key infrastructure.
In order to use a public key for, well, anything, you must have separate public keys for each application. What I'm suggesting is dropping down to one public key which is easily used in all other applications.
Once we've gotten that far down the road, it's a trivial process to attach a stream cipher to it. Of course, for what it's worth, even a block cipher would work fine, since each block is, more or less, independent of each other block. While not technically true, individual blocks could be passed from server to client easily.
Basically, at teh heart of my idea, is the idea that all traffic which can be encrypted, will be encrypted.
Now, to try and address your points, one by one: 1) Already addressed the streaming cipher issue.
2) Very true. But I'm not talking about just browsers. Every program, every daemon, every connection, employs encryption under this idea. SSL is nice. I still will push towards GPG though, out of personal preference.
3)GPG is free, period. If we could implement a GPG/(SSH|SSL) library, then people could link against that, and get full encryption functionality without dealing with the internals much at all. GPG in this case provides the certificates/key pairs, and SSH|SSL provides the socket.
4) Well, in the purest sense, virtually all protocols which use TCP can be almost interchangeable, as most of them (all of them?) are based on text interchanges between machines. As such, SSL would work. SSH would work. Any protocol should be able to sit on top of TCP, and provide a reliable connection. Whether or not it's encrypted... That's another story:)
Actually, a failing of the idea is that once PICS becomes mandatory, sites which strip it out will be blocked. Effectively nullifying the entire idea all at once, and making the service useless.
Slashdot Mirror
Also, I think a little less bias in the selection of the stories themselves could be useful. Who on earth made the decision that the release of FreeBSD 4.0 wasn't even worthy of a mention as a "quickie,"
Perhaps you should read a little more of slashdot in future, to catch information about FreeBSD 4.0? Or, if you're more specifically interested in BSD articles only, maybe just read all the articles about BSD? After all, the information is quite available to you.
Good for you. I'll admit, I've never coded assembly for windows (and strenuously avoided assembly for DOS). Never found it very enjoyable.
WTF? You think I don't know what msdos.sys's BOOTGUI option does? Do you think I'm a moron? Let me ask you, what the hell does that have to do with whether or not windows is a shell for DOS, and why the hell wouldn't I just hit the F8 key at boot? BeOS starts in MacOS, NT can run in Linux, and Linux can run in NT.
Most people don't know about it, or what it does. So, yes, until you've shown otherwise, I assume you don't know.
As for why not just hit F8, the reasponse could simply be: "Well, that's windows just loading command.com at the end, instead of completing the startup." By editing that file, it can not be simply running command.com at the end. Something else must be making it stop.
Lets play another game. Find the file called Loadlin.com on your distro's CD. I know Red Hat has it. Copy the Linux kernel to a DOS partition and type "Loadlin.com Linux.krl or" whatever. What happens? Why? Linux boots from DOS!!!! I guess that means Linux is just a shell for dos! I'm sure Linus will be surprised.
Boring game. Played it in the past, even got the t-shirt somewhere. Had to use it until the AWE32 drivers were done enough that I didn't have to use the dos utilities to initialize the card. Did it for a year or so, roughly. Anyway, the point remains the same: DOS happens well before windows ever pops up. And, from the evidence I've seen, the relationship between the two is still the same: Windows 9x is nothing but a shell which runs on top of DOS.
Oh? What's that? You've never heard of loadlin.com? I guess you must not know very much about Linux then.
Actually, you're right. I don't know much about Linux. However, in making that statement, you have to remember that I compare myself to people like Linus, Alan Cox, et al.
As for why I said you've never poked around in that section of the Windows internals before: Based on the evidence I've seen, Windows is a DOS shell. And with a complete lack of information to prove otherwise by you in your original post, I had to assume you'd never before seen the BootGUI line.
Now, here's where I have to differ with you. Since you've never dug into the internals of Win9x (or, at least, this piece of the internals), I'm going to give you step by step directions on how to see that Win9x is still just a shell over the top of DOS, just like Win3.1.
- attrib -r -h -s c:\msdos.sys
- edit msdos.sys with your favorite text only editor
- Change the line which reads "BootGUI=1" to read "BootGUI=0"
- Save the file
- attrib +r +h +s c:\msdos.sys
- Reboot your PC
- Wait until you get your familiar old DOS prompt back.
- Type in "WIN", and press
As you can see, it's not very complicated. Win9x is nothing but a DOS shell, just like Win3.1. The only important difference is that the new version of DOS (version 7.0 as of Win95) automatically loads the shell, whereas with Win3.1 it required extra action to make sure it was loaded either in the form of the install program appending to autoexec.bat, the user doing so, or the user typing "win" every time he wanted to get into windows).Win9x is only the newest version of DOS. WinNT is not, and doesn't use a DOS kernel. Win9x has a DOS kernel.
I'm a programmer. I dream of writing the next big thing to change the world(tm). But, even though my knowledge goes into the writing of the program, the program itself is absolutely worthless, even to me.
My knowledge is what's valuable. My knowledge of how the program works is what's valuable. I'll give away the program for free. Essentially, it's worthless anyway. But, if you want something specific done with it, then we'll start talking about my making money off of it.
That's my dream, personally.
Richard Jewel was the gentleman who found the bomb at the Atlanta Olympics. Was subsequently investigated as having probably placed the bomb. And then was exonerated by the same investigation. The FBI never apologized for what happened to Jewel (much harassment by the media for an FBI leak), nor what happened to his career. You see, once cleared, the media never bothered to mention that he had been. So, for many people who remember the name, Richard Jewel is that guy who placed the bomb at the Olympics. Even though he didn't, people never got the chance to hear that. And the FBI still doesn't care that this man's livelihood was trashed by them. He's just another nobody to them.
As for why he was mentioned: It's another example of FBI abuse of their powers.
In addition, configure them as identically as possible (Star Office instead of MS Office, for instance).
You're right, names don't matter. Let the users tell you which ones was better for them (oh, keep games to a minimum, in my opinion).
By patenting it, they guarantee themselves exclusive control of the ideas in the patent for the next couple decades. It prevents somebody else from coming up with and using it on their own.
Why would they need to prevent this? I honestly don't know. But it's a thought, at least.
Of these options, which do you feel are likely to be taken, and why (or why not)? Furthermore, which of these options should be taken, and which shouldn't? Again, why on all of that.
I'd like to thank you all for taking the time to do this. I know enough of the law to get myself in trouble, and appreciate the opportunity to hear from people who have actually studied these areas of law.
1) I've heard that the finding of fact is pretty much untouchable, and I've heard that it can be moderately easily overturned on appeal. Can you tell me how easily overturned this would be, and why?
2) Judging from the Finding of Fact (and economic factors), how harsh do you believe a final verdict from Judge Jackson will be (assuming that the final verdict is that Microsoft did, indeed break the law)? Not in terms of the details (what fines, etc), but rather in terms of "Wrist Slap" or "MS will need to file for Chapter 11" (and everything in between) type of terms.
3) I've also heard that MicroSoft would be very interested in settling before a final ruling by Judge Jackson. Failure to settle before then would mean that the Finding of Fact would stand as is, and be admissible in other courts as evidence against MicroSoft. This is actually a multi-part question, but here goes: a) When this case goes to the Supreme Court, the justices have the option of overturning any verdict from Judge Jackson (which is likely to be hard on Microsoft, from what I and others have read). Will the Findings of Fact still be admissible in other courts? b) Is this document admissible as evidence until then? c) Will it be admissable in case of a settlement?
Sorry to have asked so many, but they are all closely related, and I felt belonged in the same post as one group of questions.
Actually, I did read that in the original post. Of course, immediately after stating that, you came back with how you need a Linux bartender who would probably be more expensive than the Windows bartender. Here's the quote, in case you've forgotten:
Linux may be cheap, but it ain't free (as in beer) in the sense that you still need a bartender. And the truth is, for a Windows shop, it may actually be more expensive, given retraining costs and the tightness of the labor market in technical areas.
That, to me at least, indicates that TCO for Windows is lower, at least sometimes. And that's the point I'm disputing. I don't honestly believe that Windows TCO can be lower than Linux/*BSD/other UNIX TCO. The various Unices out there, once set up properly, rarely need work. The various flavors of Windows need constant attention and babying. Without it, they tend to self-destruct after a few months. Not a good thing to have at all.
I think some Linux advocates would do better if they didn't act like a bull seeing red every time the see the word Microsoft.
And this is something to which I take direct offense. I have actually advocated Microsoft in the past, and will do so again, under the right circumstances. However, one of they very few times I will do so is if the computer will see heavy games usage, which is something that Linux (and other Unices) don't have yet. Outside of the games, though, a properly set up Unix box will be easier to take care of than a properly configured Windows box. Especially if I'm in California, on a dialup, and the PC with the problem is in New York, also on a dialup.
Sorry about that.
Install Debian, do it once, and do it right. Build an image of the hard drive (using tar, to prevent problems with different sized hard drives), and a quick script later, and you've got multiple, automatic installs without any effort. And they work at least as well (if not better) than any automated install script from M$.
Next, install a local mirror of the Debian distributions. Every night, at say 9pm, it goes and downloads the latest and greatest packages (using apt-get), and gets them installed correctly on itself.
Next, each of the client machines connects to the local mirror (at say, 3am), and does the same thing. You could even throw a reboot in at the end, if you really wanted to. And now the machines are automatically updated (as much as they need to be) on a daily basis. With no human intervention whatsoever.
First, you don't have to pay somebody to download it. You order it from Cheapbytes. With shipping, you're still under $10.
Second, you don't have to burn multiple copies of the CDs to install to multiple machines simultaneously. You install to one machine, set up a proper NFS server (one option), FTP server (another option), or HTTP server (third option), and you can now install to as many client machines as you want via network (very easy to do, by the way, from my experience).
Third, you don't have to pay somebody to stay at your site constantly. Linux machines, once set up correctly, need much less babying than Windows machines. And you won't ever have to reinstall Linux, unless you have a hard drive crash. All in all, you pay somebody to set up your network (most businesses could be done in under a month, some in under a week), and then pay him to check up on things every so often.
Fourth, your retraining cost is going WAY down, since you've got Star Office, Gnome, KDE, and other Office suites on the way. People will have no problem figuring them out.
TCO is a major thing to consider. But don't cook the numbers. They're quite yummy raw, and lose much of their value once cooked.
1 RedHat CD (for all 200 machines), $70
1 RedHat Secure Server CD (for all 8 servers), $140
Upgrade or new install, total cost: $210
That's almost affordable enough for me to do on my own, without corporate sponsorship. Any company should be able to afford that quantity very easily, or it won't be up to a medium sized network.
I agree with you wholeheartedly. Using Debian is much easier than other distros. Maintenance can even be automated to a better degree, if you choose (cron, anybody?). However, the very initial installation of Debian is much more difficult than others, I think. It's getting over that hill of "Let's get it installed and working" that is so hard. Once past that, though, it's very easy to get along with.
Would you mind elaborating on this? I consider myself to be a moderately technical person (people around me look at me funny when I say that. But I know who I compare myself to, so I am only moderately technical), and I was stymied by the dselect process at first. It's not so bad anymore, and apparently apt can now do multi-cd, so it should be even easier. Anyway, I'm rambling, so I'll just ask.
What is the "more" that you've noticed to that issue?
Again, sorry about that. Shoulda thought before I posted that.
I bought the Cheapbytes Debian CDs, and was working with them. I had no problems until dselect came into the picture. After that, it was all downhill, for about a week. It took me about an hour of trying (and reading dselect help) to figure out that I was supposed to put CD 2 in, run update, and then put CD 1 in for the install. If I'd caught that up front, the entire process would have been much less painful. In addition, the interface for dslect is very very far from easy to follow (at least for me).
Now some notes about this experience: I've installed Slackware, and several versions of RedHat, all without a hitch (well, except for the first time I installed Slackware. That was due to total inexperience with setting up an OS, though). I'm not exactly a computer newbie, either. I've been working (and playing) with them since the old days of the VIC-20. Just the interface for dselect lacks a lot in terms of usability.
Now, I know that I could have just settled for one of the profiles, but I wanted to actually see what packages would be installed, and maybe modify the package listings. That process was about three hours of reviewing the listings, and determining if I wanted a package or not. Too many packages, I think.
Next issue was during the install stage of dselect. A few hours worth of package configuring (hindered by the fact that at least a third of the packages wanted me to help them out, and didn't have defaults). All in all, an extraordinarily frustrating day.
Finally, I had two other issues, but they can't be blamed on Debian itself. CD 1, it turns out, was a bad CD (meaning it had some bad sectors), so once those failed, making dselect go again (I thought) meant restarting from scratch. That was a very very frustrating experience, I can tell you. As luck would have it, I also had a hard drive (brand new) which had a bad spot on it, and had to be replaced.
All in all, not an easy installation experience, and one that I would steer newbies away from as much as possible.
My $0.02, YMMV, <insert standard disclaimer here>.
Yes, Suse is available in a box. As is slackware, Mandrake, and certain others (I just know I've seen others, but can't remember the names).
Two links of note, found thanks to this article:
Well, a few kept promises would be a good thing, I think.
- Project Magic will be updated monthly. Big surprise, it's updated every two months, unless somebody (like me) leaves them a nastygram in the guestbook telling them they haven't done so in two months.
- In April, we were told that at least a beta for linux would be released in 2Q, 1999. Shocking! We've got nothing other than a couple of (easily faked) screenshots of a text based browser!
I've given up. Mozilla still isn't ready for even beta use, Opera is dead (IMO), and I've not yet tried Amaya (though that will be fixed later today). Netscape is buggy (at best). No good browsers for Linux yet, and none on the horizon.I give up, and will simply stick with Netscape for now.
Now, maybe I'm reading more into this than I should. But what's wrong with aiming for consistency? Quite often, the people who have a consistent basis for their beliefs (and act accordingly) are among the easiest to deal with. Even the Bible-thumping Christians are easy enough to deal with, in part because you can understand their beliefs. You may not like them, but you know where they're coming from, and why they're acting that way.
Personally, I see nothing wrong with consistency, but maybe I'm missing something important.
Digital signatures, on the whole use public and private keys. These public and private keys are unique numbers, somewhere on the order of a few hundred digits long (usually). In order for a packet to be signed, that packet must have access to the private key. In order for a packet to be verified, the receiver must have access to the public key.
Now, think about what's been said:
Now, the number isn't a true "serial number", simply because it doesn't count up in order (in fact, due to other facts, not mentioned here), it can not count up.
Instead, we have something even better: a unique, cryptographically secure (supposedly) identifier attached to each and every computer which Big Blue sells. If there is a backdoor in these chips, then the government will now have a way of tracking, and reading, everything which gets encrypted/signed by these chips.
Can you see the problem yet?
Well, the main point is not the cipher itself, but rather the public key infrastructure.
:)
In order to use a public key for, well, anything, you must have separate public keys for each application. What I'm suggesting is dropping down to one public key which is easily used in all other applications.
Once we've gotten that far down the road, it's a trivial process to attach a stream cipher to it. Of course, for what it's worth, even a block cipher would work fine, since each block is, more or less, independent of each other block. While not technically true, individual blocks could be passed from server to client easily.
Basically, at teh heart of my idea, is the idea that all traffic which can be encrypted, will be encrypted.
Now, to try and address your points, one by one:
1) Already addressed the streaming cipher issue.
2) Very true. But I'm not talking about just browsers. Every program, every daemon, every connection, employs encryption under this idea. SSL is nice. I still will push towards GPG though, out of personal preference.
3)GPG is free, period. If we could implement a GPG/(SSH|SSL) library, then people could link against that, and get full encryption functionality without dealing with the internals much at all. GPG in this case provides the certificates/key pairs, and SSH|SSL provides the socket.
4) Well, in the purest sense, virtually all protocols which use TCP can be almost interchangeable, as most of them (all of them?) are based on text interchanges between machines. As such, SSL would work. SSH would work. Any protocol should be able to sit on top of TCP, and provide a reliable connection. Whether or not it's encrypted... That's another story
Well, DNS can naturally be broken down into smaller databases. For instance, the .com db, and the .net db, etc.
Keys don't neatly fall into that category. Though you could use the email addressing on those keys to break it down, I suppose.
Actually, a failing of the idea is that once PICS becomes mandatory, sites which strip it out will be blocked. Effectively nullifying the entire idea all at once, and making the service useless.