Here's how it was explained on previous episodes of Slashdot versus Microsoft:
Kerberos is designed to be an Authentication protocol. Which means it can verify who you are.
It does not attempt to be an Authorization protocol, which means it has no ability to tell the system what rights you have.
A Unixy example of this is that Kerberos can tell your system that you are in fact a user named 'root'. What it can't tell the system is that you have a UID of 0 and therefore have Superuser rights. (MS is of course using their SID concept instead of UID/GID.)
Now, if you've had any experience with NDS or LanMan Domain security, you'd know that the primary purpose of directory systems is to centralize authorization management: mapping a single user list to a various ACLs and other privledges on hundreds or thousands of boxes around the world. So a non-extended Kerberos would have been useless to Microsoft, or anyone else trying to sell a directory management system to corporate america.
On the flipside, your company probably wasn't using Kerberos for anything unless you had some different system for mapping authentication to authorization in place. What sucks, is that right now, there is no mapping between the MS authorization system and any other system you might want to use. But you can see the possiblities -- a single login system that gets you Local Administrator on your NT box and wheel access on your Unix box. --
And is it not true that that specification had a big fat field in it for vendor extensions? And that Microsoft, as the vendor, used that specified vendor extension field, to hold the extended mapping they needed (the user sid) to get it interoperating with NT? And that Unix vendors had done the same sort of thing with that field in the past?
As far as I can tell, Kerberos was used in a totally legitimate way in ActiveDirectory. The bullshit was that Microsoft published the docs for their vendor field under "public NDA" and then threatened to sue certain discussion boards for reproducing the information. --
(Back in 92-93 you could get Office for Windows for less than the price of the DOS WordPerfect wordprocessor by itself. Once you added Lotus 1-2-3 and Harvard Graphics, the cost was well over 3 times what Microsoft was charging for an more accessible product.) --
And this system exists for a reason -- up until the 1960s, most pop music was not original compositions, but just reinterpretations of the same standard songs. People wanted to go to the local nightclub and hear their local crooner sing "Satin Doll"; people wanted to hear every member of the Rat Pack sing "I Get A Kick Out Of You".
So Congress instituted a standard royalty system. The benefit was that songwriters would actually get paid for their work and not just ripped off, and record companys didn't have to create original product for every record, and everyone was happy for about 50 years. --
Which is *exactly* what was done with software fonts -- You can't copyright a typeface, but since a font file is a "computer program" (is it, or is it just data?), it can be copyrighted. That's the logic anyway. --
My take is that the economies of scale would have forced the CP/M guys to standardize and eventually beat Apple and Commodore without IBM or Microsoft to help. There's lots of basement operators in todays PC market that get along fine, and there were big companies like TI producing CP/M machines.
That is, unless Apple or Commodore would have developed an open hardware design and freely licenced their OS. But if they couldn't do that to beat IBM/Intel/Microsoft, I doubt they would have in an slightly different market. --
Before the IBM PC, "business" microcomputers were a horrible mess of partially compatible hardware, with a mostly compatible common OS (DR CP/M) and applications (WordStar, VisiCalc, etc).
The basic problem was that I couldn't take a disk with my WordStar files from my CP/M machine to your CP/M machine unless they were from the same manufacturer. Even if we null-modemed them together, we might find that our machines didn't agree between their psuedo-ASCII character sets.
So, IBM comes out with a machine that's largely a knock off of the existing CP/M 'standard' boxes, except with a 8088 instead of a 8080. And it ran WordStar and VisiCalc and this Microsoft thing that looked just like CP/M. This did bring order out of chaos because all of a sudden everyone agreed that they would use IBM's standards for disk storage and character sets and expansion slots and so on. But if IBM didn't do it, someone else or an industry consortium would have, because it was a huge fucking problem back in the day. So, the "PC" being a standard was much more about the failings of CP/M machines than it was about the companies that were doing their own thing, such as Apple, Commodore, and Atari. --
Compaq (and later Phoenix) reverse engineered the BIOS without looking at IBM's ROM listings. (The listings were there to make the machine hacker-friendly.) The IBM publications were not "open source" by any means -- using IBM's ROM code would have required a licence.
Ironically, IBM Microelectronics (a different division than the PC group) was under a court order to cheaply licence all patents it held. This meant that most "clones" used and still use IBM-licenced technology, prime examples being the ISA bus and VGA video. IBM gets a couple bucks back for every non-IBM PC sold, so they aren't complaining too much about it. --
Yup, I should have made note of that. My guess is that the industry is going to abandon SCMS because there's no way for it to tell if the device at the other end will honor the Copy Protect bit. Plus they have the 'new and improved' firewire version out now. --
One thing about the Word spec is that it depends on the "OLE Structured Stream format", which is an undocumented spec, but is of course built into Windows and is an extention to MacOS. That's where the embedded stuff like Spreadsheets and even some 'built-in' stuff like line art come in.
The thing to realize about Microsoft releasing the Word spec is that they very carefully wanted to give 3rd Party vendors enough information to create DOC files that Word could open, *but not* information to open any DOC file that Word created. So what you see is a subset of information that Word is committed to support.
And as another sidenote on the Word issue, I imagine that Microsoft themselves has a few employees dedicated full time to 'reverse-engineering' the Word format when they plan a new release. Even MS has had interoperability problems (for example, Word 97 before the service pack). --
Why do you want to smack RMS? I don't download my shit from gnu.org, I download it from RedHat (for example). I could give a shit if GNU put out their docs in MS Word v2 format - it's the Distributor's problem to put a fucking working help system in place for their "distribution". This is why they are there, is it not? --
Windows DRM can disable your S/PDIF output because there is no way to "secure" S/PDIF.
Now if you had 1394 going to your "secure" Sony amplifier/speaker combo, then Windows would presumably authenticate with the amp and data could leave your computer. --
Those "socialist" practices embodied in SimCity are actually how 90% of American cities operate.
The basis of SimCity is American-style Zoning. The basis of American-style zoning is to protect the property values of the bourgeois. The only way to succeed in SimCity is for your 'citizans' to increase their rent profits, aka property values, while attracting additional capital to your location. Therefore SimCity, is at best, a simulation of State Capitalism. --
If you use the language of a champion, you will project the fact you are the winner by default and people will all beleive you are the winner. That's the best advertising anyone can get. "Use Window2k, it's the future".
And that "language of a champion" is exactly how Microsoft got to where it is today. 2 years ago it was "UNIX? Oh, you're still running that? Expensive isn't it? Shame on you! We can help you get off that legacy platform and onto NT."
Unfortunately for Microsoft that attitude worked against Novell and OS/2, but it isn't working against UNIX. Microsoft thought that Windows 2000 would put them on the offensive against Sun (the company they hate the most) and other big server companies. Now, a year into it, not only have they not made any traction against Sun, they are fighting a defensive war against Linux in the small server market. This apparently has left them so confused that all they can do is babble about how bad UNIX and Linux is. --
Not to mention that forcing a simultaneous x86 release forces Sun's internal developers and others targetting Solaris to write portable code, which is a good thing for Sun because it gives them lots of flexibility.
(Remember a couple years ago when it looked like Sun was serious enough about IA64 that they would ship Intel boxes with their brandname on the front?) --
Yeah, and face it - the people signing the "OS X on Intel" aren't interested in specing out a supported box. What they want is to run the latest cool thing on the box they have already.* And as another poster pointed out, Apple's business model just doesn't work that way - it's easier to build your own "supported system" than to dig around in the PC parts bin and find something.
* Even Linux, which has pretty damn broad hardware support, suffers from the moaners that can't kickout $50 for a real modem. Even if Apple did invest Microsoft-style money in getting drivers ported, there would always be some NIC that didn't work, along with the latest+greatest video and sound cards. Which the "cool toy" people that wanted OS X on Intel to begin with would bitch about to no end. --
Re:Affect hardware sales?
on
OS X on x86?
·
· Score: 1
"My ISA legacy devices are what my business depends on"
As a side note in agreement, a couple years ago I saw a company where this was actually true -- IBM twinax terminal ISA cards in 100s of PC allowing them to get a greenscreen application, complete with IBM mystery drivers from 1983.
No matter that IBM and lots of other people have thought of 100s of other ways of accessing greenscreen apps, they were depending on backcompatible PC AT cruft. I always wonder what's going to happen to them when ISA finally really goes away, or if they just dumped the whole system and rewrote it in VB or something.
But, no, other than some edge cases, businesses don't care about any of the PC legacy support, or flexibility, or other advantages of the architecture. They just want something they can drop on a desk and forget until the budget comes around until they can drop a new one on the desk. Most businesses already pay Apple-style prices for stock Dells or Compaqs that come with the hard disk they come with. And despite all those ISA/PCI slots, most business PCs never ever even get their cases opened. --
Re:Affect hardware sales?
on
OS X on x86?
·
· Score: 3
What matters ultimately is how they run, not how well they fulfill YOUR idea of what a server architecture should be
"Try moving your sound card to a different PCI slot."
"Why won't my motherboard enable 2x AGP?"
"How can I keep my video card and my NIC from sharing IRQ 9 under Windows 2000?"
"My system hangs if I try to hibernate it."
"Ever since I upgraded my BIOS, all of my games run slowly."
All facts of life, day and day out, in the PC world. Check any hardware/gaming oriented message board and see. Sure, it's great if you are a tweaker, but that's not exactly Apple's core market, and it's the sort of thing that their current customers abhor.
OS X on Intel would be another one of those specilty OSes that you would pretty much have to spec out a whole machine for before buying (kinda like OpenStep 4.2!) Apple could solve that problem by selling nicely designed 'certified' boxes, but that wouldn't be open enough for the people who are campaigning for this (who want to run OSX on the box that's sitting on their desk). --
The real issue was not that One Click Buying was obvious to anyone who understood cookies (Think of the "One Click" way you can order your slashboxes on the/. homepage, etc.), it was that it was *too* obvious, and most people went and invented more complex schemes that had better security and lower return rates. But, with the Patent Office, "too obvious" is just as good as "non-obvious" -- I know someone who has patented a device that should have been thought of 100 years ago - it's just a light bulb and pieces of glass and plastic. But that's fair game, and software isn't an exception. --
Re:So when is /. going to get a decent design?
on
Freshmeat II
·
· Score: 1
Actually if he really wanted to make people happy, he could have a Slashdot mode that spit out XML, and then folks could roll their own XSLT sheets to make it do whatever they damn wanted on the client-side (even kill file people). --
Not FUD: The Exchange guys at my old job were sharp, loved Microsoft products, and generally kept Exchange up.
Their points:
1) No version of Exchange had a stable message store until 5.5SP1. According to them, that's at least 3 years on the market, corrupting mail all along! But it does work fine now, and Ex2000 solves the '1 big database' problem.
2) They had weekly maintance downtime to handle the database issues. That meant they took turns coming in on Sunday mornings. Whoop for them.
3) Even so they still occassionally had niggling database consistancy problems which they never could quite work out. When these things were happening, people would get nervous because basically the server could crash anytime. Many times they had to go offline and restore the entire messagestore from tape to solve these things.
Meanwhile, I used to do some Notes stuff. Notes has it's own problems, but at least you could backup and restore mailboxes with the COPY command, as well as solve DB corruption and whitespace issues (which cropped up rarely) with the server online. I never had to come in on the weekends at least. But to prove this isn't FUD, I'd take the Outlook interface over Notes or Netscape any day of the week --
Not even Microsoft is insane enough to suggest Exchange for an ISP solution. (They used to have a product called Commercial Internet Mail Server or something, but that was dropped.)
If you are looking for a pure mail solution, $30-$90/client (+ NT seats if you haven't bought already) is a ripoff. But if you plan to make use of the calendaring and collab features, it really isn't - Lotus/Novell/iPlanet all cost about the same amount, and there aren't any 'open source' solutions that operate on their level yet. --
Slashdot Mirror
Here's how it was explained on previous episodes of Slashdot versus Microsoft:
Kerberos is designed to be an Authentication protocol. Which means it can verify who you are.
It does not attempt to be an Authorization protocol, which means it has no ability to tell the system what rights you have.
A Unixy example of this is that Kerberos can tell your system that you are in fact a user named 'root'. What it can't tell the system is that you have a UID of 0 and therefore have Superuser rights. (MS is of course using their SID concept instead of UID/GID.)
Now, if you've had any experience with NDS or LanMan Domain security, you'd know that the primary purpose of directory systems is to centralize authorization management: mapping a single user list to a various ACLs and other privledges on hundreds or thousands of boxes around the world. So a non-extended Kerberos would have been useless to Microsoft, or anyone else trying to sell a directory management system to corporate america.
On the flipside, your company probably wasn't using Kerberos for anything unless you had some different system for mapping authentication to authorization in place. What sucks, is that right now, there is no mapping between the MS authorization system and any other system you might want to use. But you can see the possiblities -- a single login system that gets you Local Administrator on your NT box and wheel access on your Unix box.
--
And is it not true that that specification had a big fat field in it for vendor extensions? And that Microsoft, as the vendor, used that specified vendor extension field, to hold the extended mapping they needed (the user sid) to get it interoperating with NT? And that Unix vendors had done the same sort of thing with that field in the past?
As far as I can tell, Kerberos was used in a totally legitimate way in ActiveDirectory. The bullshit was that Microsoft published the docs for their vendor field under "public NDA" and then threatened to sue certain discussion boards for reproducing the information.
--
It was for the price.
(Back in 92-93 you could get Office for Windows for less than the price of the DOS WordPerfect wordprocessor by itself. Once you added Lotus 1-2-3 and Harvard Graphics, the cost was well over 3 times what Microsoft was charging for an more accessible product.)
--
And this system exists for a reason -- up until the 1960s, most pop music was not original compositions, but just reinterpretations of the same standard songs. People wanted to go to the local nightclub and hear their local crooner sing "Satin Doll"; people wanted to hear every member of the Rat Pack sing "I Get A Kick Out Of You".
So Congress instituted a standard royalty system. The benefit was that songwriters would actually get paid for their work and not just ripped off, and record companys didn't have to create original product for every record, and everyone was happy for about 50 years.
--
Which is *exactly* what was done with software fonts -- You can't copyright a typeface, but since a font file is a "computer program" (is it, or is it just data?), it can be copyrighted. That's the logic anyway.
--
My take is that the economies of scale would have forced the CP/M guys to standardize and eventually beat Apple and Commodore without IBM or Microsoft to help. There's lots of basement operators in todays PC market that get along fine, and there were big companies like TI producing CP/M machines.
That is, unless Apple or Commodore would have developed an open hardware design and freely licenced their OS. But if they couldn't do that to beat IBM/Intel/Microsoft, I doubt they would have in an slightly different market.
--
Before the IBM PC, "business" microcomputers were a horrible mess of partially compatible hardware, with a mostly compatible common OS (DR CP/M) and applications (WordStar, VisiCalc, etc).
The basic problem was that I couldn't take a disk with my WordStar files from my CP/M machine to your CP/M machine unless they were from the same manufacturer. Even if we null-modemed them together, we might find that our machines didn't agree between their psuedo-ASCII character sets.
So, IBM comes out with a machine that's largely a knock off of the existing CP/M 'standard' boxes, except with a 8088 instead of a 8080. And it ran WordStar and VisiCalc and this Microsoft thing that looked just like CP/M. This did bring order out of chaos because all of a sudden everyone agreed that they would use IBM's standards for disk storage and character sets and expansion slots and so on. But if IBM didn't do it, someone else or an industry consortium would have, because it was a huge fucking problem back in the day. So, the "PC" being a standard was much more about the failings of CP/M machines than it was about the companies that were doing their own thing, such as Apple, Commodore, and Atari.
--
Compaq (and later Phoenix) reverse engineered the BIOS without looking at IBM's ROM listings. (The listings were there to make the machine hacker-friendly.) The IBM publications were not "open source" by any means -- using IBM's ROM code would have required a licence.
Ironically, IBM Microelectronics (a different division than the PC group) was under a court order to cheaply licence all patents it held. This meant that most "clones" used and still use IBM-licenced technology, prime examples being the ISA bus and VGA video. IBM gets a couple bucks back for every non-IBM PC sold, so they aren't complaining too much about it.
--
Yup, I should have made note of that. My guess is that the industry is going to abandon SCMS because there's no way for it to tell if the device at the other end will honor the Copy Protect bit. Plus they have the 'new and improved' firewire version out now.
--
One thing about the Word spec is that it depends on the "OLE Structured Stream format", which is an undocumented spec, but is of course built into Windows and is an extention to MacOS. That's where the embedded stuff like Spreadsheets and even some 'built-in' stuff like line art come in.
The thing to realize about Microsoft releasing the Word spec is that they very carefully wanted to give 3rd Party vendors enough information to create DOC files that Word could open, *but not* information to open any DOC file that Word created. So what you see is a subset of information that Word is committed to support.
And as another sidenote on the Word issue, I imagine that Microsoft themselves has a few employees dedicated full time to 'reverse-engineering' the Word format when they plan a new release. Even MS has had interoperability problems (for example, Word 97 before the service pack).
--
Why do you want to smack RMS? I don't download my shit from gnu.org, I download it from RedHat (for example). I could give a shit if GNU put out their docs in MS Word v2 format - it's the Distributor's problem to put a fucking working help system in place for their "distribution". This is why they are there, is it not?
--
Because: It's a patent, so it's enforcable even if it's reverse engineered.
VirutalDub was forced to remove ASF support because of this patent protection.
(Sure a file format patent can probably be challenged, but who has time and money to fight Microsoft on that one when you could just use AVI.)
--
Windows DRM can disable your S/PDIF output because there is no way to "secure" S/PDIF.
Now if you had 1394 going to your "secure" Sony amplifier/speaker combo, then Windows would presumably authenticate with the amp and data could leave your computer.
--
Which brings up the question of when Linux vendors will takes steps to deal with the horrible documentation format problem on most distros.
(Translation: I don't care what format GNU wants to use. I got this shit from you, so you fix the problem. I just want man to work.)
--
Those "socialist" practices embodied in SimCity are actually how 90% of American cities operate.
The basis of SimCity is American-style Zoning. The basis of American-style zoning is to protect the property values of the bourgeois. The only way to succeed in SimCity is for your 'citizans' to increase their rent profits, aka property values, while attracting additional capital to your location. Therefore SimCity, is at best, a simulation of State Capitalism.
--
If you use the language of a champion, you will project the fact you are the winner by default and people will all beleive you are the winner. That's the best advertising anyone can get. "Use Window2k, it's the future".
And that "language of a champion" is exactly how Microsoft got to where it is today. 2 years ago it was "UNIX? Oh, you're still running that? Expensive isn't it? Shame on you! We can help you get off that legacy platform and onto NT."
Unfortunately for Microsoft that attitude worked against Novell and OS/2, but it isn't working against UNIX. Microsoft thought that Windows 2000 would put them on the offensive against Sun (the company they hate the most) and other big server companies. Now, a year into it, not only have they not made any traction against Sun, they are fighting a defensive war against Linux in the small server market. This apparently has left them so confused that all they can do is babble about how bad UNIX and Linux is.
--
Not to mention that forcing a simultaneous x86 release forces Sun's internal developers and others targetting Solaris to write portable code, which is a good thing for Sun because it gives them lots of flexibility.
(Remember a couple years ago when it looked like Sun was serious enough about IA64 that they would ship Intel boxes with their brandname on the front?)
--
Yeah, and face it - the people signing the "OS X on Intel" aren't interested in specing out a supported box. What they want is to run the latest cool thing on the box they have already.* And as another poster pointed out, Apple's business model just doesn't work that way - it's easier to build your own "supported system" than to dig around in the PC parts bin and find something.
* Even Linux, which has pretty damn broad hardware support, suffers from the moaners that can't kickout $50 for a real modem. Even if Apple did invest Microsoft-style money in getting drivers ported, there would always be some NIC that didn't work, along with the latest+greatest video and sound cards. Which the "cool toy" people that wanted OS X on Intel to begin with would bitch about to no end.
--
"My ISA legacy devices are what my business depends on"
As a side note in agreement, a couple years ago I saw a company where this was actually true -- IBM twinax terminal ISA cards in 100s of PC allowing them to get a greenscreen application, complete with IBM mystery drivers from 1983.
No matter that IBM and lots of other people have thought of 100s of other ways of accessing greenscreen apps, they were depending on backcompatible PC AT cruft. I always wonder what's going to happen to them when ISA finally really goes away, or if they just dumped the whole system and rewrote it in VB or something.
But, no, other than some edge cases, businesses don't care about any of the PC legacy support, or flexibility, or other advantages of the architecture. They just want something they can drop on a desk and forget until the budget comes around until they can drop a new one on the desk. Most businesses already pay Apple-style prices for stock Dells or Compaqs that come with the hard disk they come with. And despite all those ISA/PCI slots, most business PCs never ever even get their cases opened.
--
What matters ultimately is how they run, not how well they fulfill YOUR idea of what a server architecture should be
"Try moving your sound card to a different PCI slot."
"Why won't my motherboard enable 2x AGP?"
"How can I keep my video card and my NIC from sharing IRQ 9 under Windows 2000?"
"My system hangs if I try to hibernate it."
"Ever since I upgraded my BIOS, all of my games run slowly."
All facts of life, day and day out, in the PC world. Check any hardware/gaming oriented message board and see. Sure, it's great if you are a tweaker, but that's not exactly Apple's core market, and it's the sort of thing that their current customers abhor.
OS X on Intel would be another one of those specilty OSes that you would pretty much have to spec out a whole machine for before buying (kinda like OpenStep 4.2!) Apple could solve that problem by selling nicely designed 'certified' boxes, but that wouldn't be open enough for the people who are campaigning for this (who want to run OSX on the box that's sitting on their desk).
--
The real issue was not that One Click Buying was obvious to anyone who understood cookies (Think of the "One Click" way you can order your slashboxes on the /. homepage, etc.), it was that it was *too* obvious, and most people went and invented more complex schemes that had better security and lower return rates. But, with the Patent Office, "too obvious" is just as good as "non-obvious" -- I know someone who has patented a device that should have been thought of 100 years ago - it's just a light bulb and pieces of glass and plastic. But that's fair game, and software isn't an exception.
--
Actually if he really wanted to make people happy, he could have a Slashdot mode that spit out XML, and then folks could roll their own XSLT sheets to make it do whatever they damn wanted on the client-side (even kill file people).
--
Not FUD: The Exchange guys at my old job were sharp, loved Microsoft products, and generally kept Exchange up.
Their points:
1) No version of Exchange had a stable message store until 5.5SP1. According to them, that's at least 3 years on the market, corrupting mail all along! But it does work fine now, and Ex2000 solves the '1 big database' problem.
2) They had weekly maintance downtime to handle the database issues. That meant they took turns coming in on Sunday mornings. Whoop for them.
3) Even so they still occassionally had niggling database consistancy problems which they never could quite work out. When these things were happening, people would get nervous because basically the server could crash anytime. Many times they had to go offline and restore the entire messagestore from tape to solve these things.
Meanwhile, I used to do some Notes stuff. Notes has it's own problems, but at least you could backup and restore mailboxes with the COPY command, as well as solve DB corruption and whitespace issues (which cropped up rarely) with the server online. I never had to come in on the weekends at least. But to prove this isn't FUD, I'd take the Outlook interface over Notes or Netscape any day of the week
--
I use OWA 5.5 for 9 hours a day at work. It crashes Netscape 4.76 within 5 minutes and 6.0 immedeately.
But it also crashes IE 5.01 about once a day. Turns out that IE can't reliably handle lots of javascript day-in and day-out.
--
Not even Microsoft is insane enough to suggest Exchange for an ISP solution. (They used to have a product called Commercial Internet Mail Server or something, but that was dropped.)
If you are looking for a pure mail solution, $30-$90/client (+ NT seats if you haven't bought already) is a ripoff. But if you plan to make use of the calendaring and collab features, it really isn't - Lotus/Novell/iPlanet all cost about the same amount, and there aren't any 'open source' solutions that operate on their level yet.
--