As others have pointed out, we already have a cathedral system, but it's masquerading as a bazaar because there are so many of them.
Rather than have a package management system, what really needs to happen is a package management framework. If the basic tools for a package management system were compatible with each other, then you could have RPM, DEB, portage, etc that all understood and could use each others formats.
Unfortunately, this means defining a single COMPREHENSIVE standard, rather than the typical open source "just enough to get by and expand when we need it" approach.
The bazaar is not always the best model. In reality, you can only achieve so much without a cathedral eventually coming into play. You can achieve a LOT with a bazaar, but ultimately there's a point of diminishing returns where conflict between the parties grinds productivity to a halt and crushes ingenuity.
XP is actually significantly more than just 2000 with a firewall, especially when you consider SP2. It's worth the upgrade to XP SP2 for the browser security improvements alone (though you shouldn't have to upgrade to get them.. but that doesn't change the fact that you DO have to upgrade to get them).
Things in XP that I use every day and would go nuts not having if I went back to 2K.
* Tray Icon Hiding. Too many apps put icons on the tray and it's very nice to get rid of them.
* The new Start menu. I can get at pretty much everything in 1 or 2 clicks, without having to minimize everything to get to the desktop (for instance, right click on "My Computer" in the start menu and choose "Manage" to get to computer management. Right click the "Network Neighborhood" in the Start menu and choose properties to view all your network connections, etc..
* Remote Desktop. I use this *ALL* the time. I'd have to run 2000 Server to get terminal services in 2000.
* Volume Shadow Copy client. This lets me version snapshot network drives and get previous versions from various dates and times.
That's not even counting how fast XP boots compared to the typical 3 minute boot time of 2000 in a domain environment.
While Novell's own technology is very good, their biggest problem is that they create absolutely CRAP network clients.
They've got better at some of it. For instance, the clients no longer take over your machine and are impossible to uninstall without reformatting, but there's still big problems.
For instance, even with the current client on our systems, if you don't specify a default server, it blue screens your machine. Stupid.
I'm not a fanboy. I just can't stand when people continue to parot long obsolete information.
Yes, C2 was more or less a useless specification, other than as a checkmark to gain government contracts.
Common Criteria, however, is not and is much more realistice. It addresses your concerns about 'real world' conditions. That's why CC includes things like VPN's and network management tools.
You have a valid point that NT4's Network certification is largely meaningless, but Win2k's CC certification is not, and actually represents real configuration environments.
Opening a new window *IS* a popup. What are you smoking?
Again, I'm not talking about what happens on YOUR browser, I'm talking about what happens on a publicly accessible web Kiosk that is not supposed to be use for browsing anything other than the pages necessary to register for classes.
Sounds like the moron here is you. It's not a matter of "enabling" the back button. There *IS* no back button in Kiosk mode. There's nothing to enable. Why not learn what you're talking about before you start making suggestions and calling people idiots? No, sorry, that would be asking too much of slashdot.
There are legitimate uses for popups. For example, we run "web kiosks" at school that allow students to register for classes and such. Occasionally, we need to send a student outside of our site to a state or federal site and we tend to open those up in new popup windows so that they can get back to our site just by closing the window, otherwise there would be no way for them to get "home" (they don't have a back button either in kiosk mode).
Otherwise, 2 or 3 times a day we have to send someone to the kiosks to reset them to the default page.
Oh, and don't even think about suggesting frames, the java servlets used by our registration system don't work right in frames, not to mention all the other problems related to frames and usability.
No, Microsoft believed (and probably still does) that tabs violate their design guidelines. However, tabs have been requested by so many people that they seem to have relented.
Of course other products (even other MS products) use tabs, such as Excel, but in that case the tabs are keeping related items organized (as do tabbed dialogs). MS has been pushing an SDI interface for almost 10 years for discrete work items (Different documents in Word, for instance) and MDI for related items (code files in Visual Studio, for instance).
While there are certainly some distro's that handle dependancies better than others, gentoo is hardly a "emerge and it's done" kind of thing. It's usually "emerge and wait 6-36 hours, and don't shut down or reboot or anything else while you're doing it and your done."
Then there's the fact that it's a command line, and someone needs to know the incantation necessary to get what they want. Yup, there's GUI tools as well, but most of them are VERY poorly designed and not very useful.
Then you better hope you weren't unlucky enough to get a package in the middle of it being updated. Yeah, gentoo makes this less likely by masking packages during the update process, but it can still happen. Boom/bam it's broke and you're left wondering why.
And speaking of masking, there's no guarantee the package you want is actually unmasked at any given time, which means fidling with more parameters and config files to get them.
Seriously. If you think gentoo is a usable distro for grandma and grampa, you're smoking something. It's a wonderful distro for an expert, it's horrible for most others.
I was trained on Big Iron, IBM and Amdahl mainframes in the early 80's. Those things only went down when they were replaced.
That may have been true at your installation, but I assure you it wasn't a common occurance in the mainframe world. For example, at one of the shops I worked at, a nightly IPL (equivelent to a reboot) occurred because the CICS code had a memory bug somewhere that IBM couldn't seem to fix. At another shop they had weekly IPL's for a similar reason.
Now, if you were a huge shop that paid IBM a ton of money, they typically had SE's *ON SITE* 24/7 to fix these kinds of issues, but we were a small piss-ant insurance company that IBM couldn't be bothered with unless things broke completely (which still happened way too often).
What this means is that sure, if you have the money to have someone from the company making the hardware and OS on site, things will run smooth, much like thins would run smooth if an experienced MS SE (NOT MCSE) were on site.
Actually, it's not even two-thirds of all web servers. It's two-thirds of all web site *DOMAINS* run on Apache. Remember, domain != server.
Netcraft has done some statistics in the past that shows how this breaks down.
These statistics are grossly out of date, of course. However, given that MS's general percentage of hosts hasn't changed much over the last few years the Windows stats are probably still close to accurate. Also, the Linux percentage is probably much higher because most of the people running Unix webservers have converted to Linux over the last few years (which is likely why the Unix webservers have dropped so much). Of course this is speculation, but would explain the drop in non-apache Unix servers.
What about all those sites out there using blink and marquee tags? Should we just shrug and say "On well, got to support them for life"? OR what about the sites that conform to the crappy and broken Netscape CSS1 support?
At some point (and i'm not saying that point will be anytime real soon) you have to break with legacy to move forward, unless your legacy was designed to be forward compatible (xhtml 1.1 is).
There aren't many sites still supporting gopher:// for instance. You just have to move on when better things come along, and you usually end up dragging the laggards kicking and screaming for the good of all.
As for dynamically generated sites, those sites COULD install filters that would convert code using something like Tidy (of course they might want to do that with their existing code today... ) That's not an exacuse.
Simply put, I don't see anything wrong with sites having to stay up to currently supported standards TO LOOK GOOD. I wouldn't mind if browsers just stripped *ALL* formatting from older versions of HTML (HTML 2 and earlier today, and HTML 3.2 in a few years). That leaves the data still there, though it may not be very readable.. i suppose some basic kind of style must be preserved.
Actually, you're right and wrong. Needs are driven by the website visitor, but they don't know what their needs are.
In effect, the needs are driven by the software they are using, which means the needs are driven by the browser vendor of the software the site visitors are using.
Thus, if the site visitor's software is following the "advancing standard" then so does the visitors need. Since all the major browsers (hopefully including IE) are moving in that direction then so will end users as they upgrade.
What you're saying is "I don't care about all them newfangled horselss carriages, horses are good enough and will be used by everyone I know for years to come". That may be true, but eventually it won't be.
Many things are wrong with current HTML. Well, ok, not CURRENT HTML (4.01) but everything before that.
Since XHTML is just a reformulation of HTML 4.01 into XML, and XHTML 1.1 is just a modularization of XHTML, technically nothhing is wrong with HTML as it exists today. But what about how it exiists tomorrow?
XHTML allows for the easy expansion of the language. Right now, DOCTYPES are the only way to define what version of HTML you're using, which makes it an all or nothing proposition. What if you want to use HTML + SUPERCOOLHTML-Extended? XHTML 1.1 allows you to basically define your own syntax, and more importantly allow the standards body to do so easily).
This way you only have to use as much of the standard as you want to, and if there are two competing standards you can actually choose which one (or ones) to use in a way that the browsers can understand.
Now, i'll grant you that for the typical "my cat fluffy" site, they don't give a rip. HTML 4.01 is just fine. But did you know that HTML 4.01 strict doesn't have font tags? It doesn't have the target attribute on links. It doesn't have a lot of stuff you're used to that are going away.
It's best to get used to XHTML now, HTML won't be improved, but XHTML will.
"strong" is a semantic definition while "bold" is a typographical one. What does "bold" mean to a text-to-speech device? The same holds for "em" or "emphasis".
The purpose of XHTML is to decouple the presentation from the information. "strong" just tells whatever is interpreting it to do whatever it thinks is best to make it stand out.
While I agree that i've not seen any evidence to support the claim of DoS yet, I just wouldn't put it past anyone. People get emotional over things and often run off and do things that are not well advised.
The fact is, all it takes is for one hothead with the skill and resources to do something like this, and we're all aware that many of us geeks tend to have less than stellar emotional maturity at times (as any trip through the kernel mailing list will tell you).
Maybe it never happened, maybe it did, maybe it was exagerated, maybe it was something misinterpreted... the point is, I'm not going to dismiss the possibility of it happening, only question the facts supporting it.
Better yet, why don't you actually try to figure out what it's trying to do. Download filemon and figure out what it's trying to access and then only give the privs necessary to make that app work.
Printing? It's probably writing to the spool folder for some reason. Give the spool folder write permissions for the users. You don't need to give the users administrator privs.
Really, it's a bit earlier to be gloating about Firefox. It took them something like 10 days to patch the last two vulnerabilities. Now we're gloating about how slow IE is after 1? That's not right.
Further, MS has been known to ship critical fixes outside the normal patch timeframe, especially when they become publicly known, so assuming MS will wait a month is rather childish as well.
I've never had a problem with Publisher 2003 needing systemroot access. If you're running older versions, you don't need to give them root access. All you need to do is give them write permission to the directory without replacing the permissions on the files within, that way nothing alter existing files. There's nothing special about systemroot other than it's a place many system files are stored.. let the user create new files there isn't going to comprimise security any more than letting them create new files somewhere else.
Oh... you have VERY little experience with changing settings in most Linux software.
An example. I'm using postfix virtual mail server facility, and using authdaemon to provide non-unix accounts for email. Authdaemon is the most sensitive PoS i've used, and it's tools constantly screw up the configuration files. Any slight misconfiguration is happily accepted by authdaemon with no errors generated and you're left scratching your head trying to figure out what's wrong and why it's not "just working".
Bah... *SOME* Linux software works wonderfully, other software sucks horribly. It's people like you, that can't understand that your best experience is not equivelent to an average experience that make it hard to sell this platform.
It's not the place of schools to (effectively) set technology policy for business in the future. Schools react to what the business community wants, not what the schools want the businesses to do.
Slashdot Mirror
As others have pointed out, we already have a cathedral system, but it's masquerading as a bazaar because there are so many of them.
Rather than have a package management system, what really needs to happen is a package management framework. If the basic tools for a package management system were compatible with each other, then you could have RPM, DEB, portage, etc that all understood and could use each others formats.
Unfortunately, this means defining a single COMPREHENSIVE standard, rather than the typical open source "just enough to get by and expand when we need it" approach.
The bazaar is not always the best model. In reality, you can only achieve so much without a cathedral eventually coming into play. You can achieve a LOT with a bazaar, but ultimately there's a point of diminishing returns where conflict between the parties grinds productivity to a halt and crushes ingenuity.
XP is actually significantly more than just 2000 with a firewall, especially when you consider SP2. It's worth the upgrade to XP SP2 for the browser security improvements alone (though you shouldn't have to upgrade to get them.. but that doesn't change the fact that you DO have to upgrade to get them).
Things in XP that I use every day and would go nuts not having if I went back to 2K.
* Tray Icon Hiding. Too many apps put icons on the tray and it's very nice to get rid of them.
* The new Start menu. I can get at pretty much everything in 1 or 2 clicks, without having to minimize everything to get to the desktop (for instance, right click on "My Computer" in the start menu and choose "Manage" to get to computer management. Right click the "Network Neighborhood" in the Start menu and choose properties to view all your network connections, etc..
* Remote Desktop. I use this *ALL* the time. I'd have to run 2000 Server to get terminal services in 2000.
* Volume Shadow Copy client. This lets me version snapshot network drives and get previous versions from various dates and times.
That's not even counting how fast XP boots compared to the typical 3 minute boot time of 2000 in a domain environment.
Will Apple use a standard PC BIOS to replace OF, or will they throw something new together?
My bet is on EFI, Intel's new Extensible Firmware Interface.
While Novell's own technology is very good, their biggest problem is that they create absolutely CRAP network clients.
They've got better at some of it. For instance, the clients no longer take over your machine and are impossible to uninstall without reformatting, but there's still big problems.
For instance, even with the current client on our systems, if you don't specify a default server, it blue screens your machine. Stupid.
I'm not a fanboy. I just can't stand when people continue to parot long obsolete information.
Yes, C2 was more or less a useless specification, other than as a checkmark to gain government contracts.
Common Criteria, however, is not and is much more realistice. It addresses your concerns about 'real world' conditions. That's why CC includes things like VPN's and network management tools.
You have a valid point that NT4's Network certification is largely meaningless, but Win2k's CC certification is not, and actually represents real configuration environments.
Why do people like you keep reiterating the tired old "without a network" line?
It hasn't been true since at least NT4 SP6a, when NT4 achieved a C2 rating *WITH* network. Windows 2000 achieved CC both with and without networking.
The NT4 link is no longer around on MS's site, but there are still some pages out there that reference it:
Such as this one
And here is Win2k
Opening a new window *IS* a popup. What are you smoking?
Again, I'm not talking about what happens on YOUR browser, I'm talking about what happens on a publicly accessible web Kiosk that is not supposed to be use for browsing anything other than the pages necessary to register for classes.
Sounds like the moron here is you. It's not a matter of "enabling" the back button. There *IS* no back button in Kiosk mode. There's nothing to enable. Why not learn what you're talking about before you start making suggestions and calling people idiots? No, sorry, that would be asking too much of slashdot.
There are legitimate uses for popups. For example, we run "web kiosks" at school that allow students to register for classes and such. Occasionally, we need to send a student outside of our site to a state or federal site and we tend to open those up in new popup windows so that they can get back to our site just by closing the window, otherwise there would be no way for them to get "home" (they don't have a back button either in kiosk mode).
Otherwise, 2 or 3 times a day we have to send someone to the kiosks to reset them to the default page.
Oh, and don't even think about suggesting frames, the java servlets used by our registration system don't work right in frames, not to mention all the other problems related to frames and usability.
No, Microsoft believed (and probably still does) that tabs violate their design guidelines. However, tabs have been requested by so many people that they seem to have relented.
Of course other products (even other MS products) use tabs, such as Excel, but in that case the tabs are keeping related items organized (as do tabbed dialogs). MS has been pushing an SDI interface for almost 10 years for discrete work items (Different documents in Word, for instance) and MDI for related items (code files in Visual Studio, for instance).
While there are certainly some distro's that handle dependancies better than others, gentoo is hardly a "emerge and it's done" kind of thing. It's usually "emerge and wait 6-36 hours, and don't shut down or reboot or anything else while you're doing it and your done."
Then there's the fact that it's a command line, and someone needs to know the incantation necessary to get what they want. Yup, there's GUI tools as well, but most of them are VERY poorly designed and not very useful.
Then you better hope you weren't unlucky enough to get a package in the middle of it being updated. Yeah, gentoo makes this less likely by masking packages during the update process, but it can still happen. Boom/bam it's broke and you're left wondering why.
And speaking of masking, there's no guarantee the package you want is actually unmasked at any given time, which means fidling with more parameters and config files to get them.
Seriously. If you think gentoo is a usable distro for grandma and grampa, you're smoking something. It's a wonderful distro for an expert, it's horrible for most others.
I was trained on Big Iron, IBM and Amdahl mainframes in the early 80's. Those things only went down when they were replaced.
That may have been true at your installation, but I assure you it wasn't a common occurance in the mainframe world. For example, at one of the shops I worked at, a nightly IPL (equivelent to a reboot) occurred because the CICS code had a memory bug somewhere that IBM couldn't seem to fix. At another shop they had weekly IPL's for a similar reason.
Now, if you were a huge shop that paid IBM a ton of money, they typically had SE's *ON SITE* 24/7 to fix these kinds of issues, but we were a small piss-ant insurance company that IBM couldn't be bothered with unless things broke completely (which still happened way too often).
What this means is that sure, if you have the money to have someone from the company making the hardware and OS on site, things will run smooth, much like thins would run smooth if an experienced MS SE (NOT MCSE) were on site.
Actually, it's not even two-thirds of all web servers. It's two-thirds of all web site *DOMAINS* run on Apache. Remember, domain != server.
Netcraft has done some statistics in the past that shows how this breaks down.
These statistics are grossly out of date, of course. However, given that MS's general percentage of hosts hasn't changed much over the last few years the Windows stats are probably still close to accurate. Also, the Linux percentage is probably much higher because most of the people running Unix webservers have converted to Linux over the last few years (which is likely why the Unix webservers have dropped so much). Of course this is speculation, but would explain the drop in non-apache Unix servers.
What about all those sites out there using blink and marquee tags? Should we just shrug and say "On well, got to support them for life"? OR what about the sites that conform to the crappy and broken Netscape CSS1 support?
At some point (and i'm not saying that point will be anytime real soon) you have to break with legacy to move forward, unless your legacy was designed to be forward compatible (xhtml 1.1 is).
There aren't many sites still supporting gopher:// for instance. You just have to move on when better things come along, and you usually end up dragging the laggards kicking and screaming for the good of all.
As for dynamically generated sites, those sites COULD install filters that would convert code using something like Tidy (of course they might want to do that with their existing code today... ) That's not an exacuse.
Simply put, I don't see anything wrong with sites having to stay up to currently supported standards TO LOOK GOOD. I wouldn't mind if browsers just stripped *ALL* formatting from older versions of HTML (HTML 2 and earlier today, and HTML 3.2 in a few years). That leaves the data still there, though it may not be very readable.. i suppose some basic kind of style must be preserved.
Actually, you're right and wrong. Needs are driven by the website visitor, but they don't know what their needs are.
In effect, the needs are driven by the software they are using, which means the needs are driven by the browser vendor of the software the site visitors are using.
Thus, if the site visitor's software is following the "advancing standard" then so does the visitors need. Since all the major browsers (hopefully including IE) are moving in that direction then so will end users as they upgrade.
What you're saying is "I don't care about all them newfangled horselss carriages, horses are good enough and will be used by everyone I know for years to come". That may be true, but eventually it won't be.
Many things are wrong with current HTML. Well, ok, not CURRENT HTML (4.01) but everything before that.
Since XHTML is just a reformulation of HTML 4.01 into XML, and XHTML 1.1 is just a modularization of XHTML, technically nothhing is wrong with HTML as it exists today. But what about how it exiists tomorrow?
XHTML allows for the easy expansion of the language. Right now, DOCTYPES are the only way to define what version of HTML you're using, which makes it an all or nothing proposition. What if you want to use HTML + SUPERCOOLHTML-Extended? XHTML 1.1 allows you to basically define your own syntax, and more importantly allow the standards body to do so easily).
This way you only have to use as much of the standard as you want to, and if there are two competing standards you can actually choose which one (or ones) to use in a way that the browsers can understand.
Now, i'll grant you that for the typical "my cat fluffy" site, they don't give a rip. HTML 4.01 is just fine. But did you know that HTML 4.01 strict doesn't have font tags? It doesn't have the target attribute on links. It doesn't have a lot of stuff you're used to that are going away.
It's best to get used to XHTML now, HTML won't be improved, but XHTML will.
"strong" is a semantic definition while "bold" is a typographical one. What does "bold" mean to a text-to-speech device? The same holds for "em" or "emphasis".
The purpose of XHTML is to decouple the presentation from the information. "strong" just tells whatever is interpreting it to do whatever it thinks is best to make it stand out.
While I agree that i've not seen any evidence to support the claim of DoS yet, I just wouldn't put it past anyone. People get emotional over things and often run off and do things that are not well advised.
The fact is, all it takes is for one hothead with the skill and resources to do something like this, and we're all aware that many of us geeks tend to have less than stellar emotional maturity at times (as any trip through the kernel mailing list will tell you).
Maybe it never happened, maybe it did, maybe it was exagerated, maybe it was something misinterpreted... the point is, I'm not going to dismiss the possibility of it happening, only question the facts supporting it.
Better yet, why don't you actually try to figure out what it's trying to do. Download filemon and figure out what it's trying to access and then only give the privs necessary to make that app work.
Printing? It's probably writing to the spool folder for some reason. Give the spool folder write permissions for the users. You don't need to give the users administrator privs.
No, you did not get that right. Reread what I wrote.
Really, it's a bit earlier to be gloating about Firefox. It took them something like 10 days to patch the last two vulnerabilities. Now we're gloating about how slow IE is after 1? That's not right.
Further, MS has been known to ship critical fixes outside the normal patch timeframe, especially when they become publicly known, so assuming MS will wait a month is rather childish as well.
I've never had a problem with Publisher 2003 needing systemroot access. If you're running older versions, you don't need to give them root access. All you need to do is give them write permission to the directory without replacing the permissions on the files within, that way nothing alter existing files. There's nothing special about systemroot other than it's a place many system files are stored.. let the user create new files there isn't going to comprimise security any more than letting them create new files somewhere else.
Oh... you have VERY little experience with changing settings in most Linux software.
An example. I'm using postfix virtual mail server facility, and using authdaemon to provide non-unix accounts for email. Authdaemon is the most sensitive PoS i've used, and it's tools constantly screw up the configuration files. Any slight misconfiguration is happily accepted by authdaemon with no errors generated and you're left scratching your head trying to figure out what's wrong and why it's not "just working".
Bah... *SOME* Linux software works wonderfully, other software sucks horribly. It's people like you, that can't understand that your best experience is not equivelent to an average experience that make it hard to sell this platform.
It's not the place of schools to (effectively) set technology policy for business in the future. Schools react to what the business community wants, not what the schools want the businesses to do.