COM is an interface standard, not a API, and it's not specific to Windows or x86 - it has been ported elsewhere. Writing portable code is an old trick, the point is to have portable binaries. If Apache supported this it would make module packaging and installation much better on all platforms. Getting to recompile Apache is a hack, not a feature.
> Apache works fine on Windows - very, very well, actually Apparently the folks in DC disagree.
People have no problem using the open software, but they continue to have small compatibility issues that keep them falling back to Microsoft.
If coordinated IT efforts like City Of Munich and their IBM Con$ultant$ can't adequately solve these "small compatibility issues", it's a lot to ask of the very stretched education IT budgets.
teach kids how to use Linux and the spirit of open source and the "real world" will eventually change.
Maybe you should remind your wife that teaching religion is schools is illegal:P
Blame whoever you want, but Windows has many production web servers, and Apache is not one of them.
> There is, in fact, a standard for such things. It's called Posix.
Apparently Windows 2003 R2 (shipping soon) will have fairly complete POSIX support. The entire rationale for the Apache 2 branch was better Windows support -- if they still haven't got it working, maybe it would be easier for them if they junked it and just targetted the W2003 portability layer.
> Further, Windows doesn't have a source-based build system at all! Shouldn't it?
Windows has a binary standard which allows code to 'snap together like legos', so why would it need the kludge of an entire development and build environment on production systems? There seems to be widespread confusion in the *nix world -- 'make' is not an installer. You have things like XPCOM -- use them.
The [Windows Update] ActiveX control is able to exceed user's security privileges to provide back certain system information that might not otherwise be possible to do otherwise for a given user.
Umm, no. Windows Update can only be run by Administrator users, and administrators can (directly or indirectly) do anything to a system.
ActiveX has enough real problems that there's absolutely no need to manufacture ridiclous falsehoods in order to talk it down.
The trust ratings and user comments need to be safe from poisoning and therefore moderated
Keep in mind that Kazaa was the run-away most popular filesharing client for years, despite all of the well-known spyware it came with.
If you want to moderate all of the "wrong" opinons or just plain spam on this proposed BBS, you might as well just skip a step and put the Cabal directly in charge. (Whether that would be mozilla.org is unlikely, I think.)
And since your proposal relies on hashes, browser support, and some sort of authority, you might as well accept that you've just proposed code signing and you agree with me:)
The upshot is to really do it right, you need operating system support for an application-based security model. No current desktop OS currently supports this kind of security -- they are all firmly rooted in the user-based security model inherited from time-sharing systems.
That means either waiting for Vista or waiting for someone to add this security model to Linux/X11. Hopefully Firefox (and other internet software packages) will mimic IE and also have "low-rights" support on Vista.
No, I don't think signing is a cure-all, but it does minimize one social exploit. Whatever you think about ActiveX, I've never heard about an evil control that pretends to be Windows Update or Macromedia Flash.
If firefox become popular, it's possible there would be a ton of fake "Ad Block" and "Tab Browser" extentions, and signing is pretty much the only way to stop it.
If you want to see an example of this in action, search Google for "eMule", the opensource filesharing client. About 90% of the links go to fake sites which are probably spyware-laden clients. Too bad the official Emule installer doesn't use Authenicode -- I would defiantly check it.
Now it would be nice if code-signing was extended so that things could be "Certified by So-N-So to be Spyware-Free!". But even then, if it's an open system, fake certifiers will come about.
Yup, Firefox isn't using enough memory right now -- let's run Bittorrent in the same process space too! Memory protection is overrated anyway.
How would a bittorrent:// protocol and accompanying extention provide a better experience the current system of http, a torrent file, and a stand-alone client? It's only two fricking clicks to download something -- not like that's preventing people from installing Linux.
Correction -- ActiveX Distribution is officially called "Internet Component Download". Again, this is the thing people dislike the most about IE, not necessarily "ActiveX".
Good point -- it always helps to clear up the termonology before diving too deep into a flamewar. Mozilla has developed a bunch of technologies that have rough equivilance to IE tech:
Netscape Plugins =~ ActiveX control XPInstall =~ "ActiveX Web Distribution" (may not be the official name) Firefox Extentions =~ Browser Helper Objects (BHOs)
The confusion I think is that most BHOs use ActiveX Distribution as the installation mechanism.
(And the other confusin is that MS has defined the term "ActiveX" in 9 different ways. Tons of stuff in Windows use COM/ActiveX, but the think people bitch about is the installation mechanism.)
Because in theory, someone educated enough to run Firefox would also be educated enough to not allow it to run untrusted things.
In theory, Firefox is a browser for the masses and is designed to supplant Internet Explorer. If Firefox has a userbase that's more technically sophisticated than other browsers, that only means that there's more work to do.
So please quit blowing yourself by thinking Firefox is l33t d00d software -- it isn't. The whole goal is stripped down and simple for the ordinary IE user.
Now it is true that Extentions are "elite", and they are generally only found on one or two sites. The questions is if the security model will hold up when Firefox gets more popular and users get used to instaling extentions from a varity of sources. I'm sure at some pont a signing mechanism like Authenicode will be deemed necessary.
> In fact, it's an Application in the form of a.app directory.
Even that's too much trouble. Just create a old-style Carbon binary (CFM?), set the file type to APPL, and the file extention will be ignored. (MacOS didn't have the concept of extentions until OS X) Give it the stock JPEG icon and your application will be virtually indistigishable from a regular JPEG.
> By that metric, both Germany and Japan are still occupied nations
Germany was technically under occupation until reunification in 1992 or so.
Japan left legal occupation significantly eariler, and could change their constitution if they wanted to (Article 96), although apparently there's significant popular support for the peace provision.
I certainly don't think it's a matter of resources. It's the marketing problem -- cell phone manufacturers are basically bound to the service providers, and can't compete with their services (like with iTMS). This runs contrary to Apple's mode of operation, where they basically control their entire channel.
Number are totally baloney because Slashdot is served gzipped.
For this example, the total yearly savings for Slashdot would be: $3,650 USD!
OOH, three whole thousand dollars!!! Going to be a great XMas party this year!
Seriously, are you impressed with that? This ain't a mom-n-pop operation. I wouldn't be surprised if the site redesign cost VA Linux $100K. Even if it cost half that, and even with bullshit figures, your cost analysis falls right on its ass.
About the only business reason to justify such a redesign is so/. doesn't look so hypocritical it's daily blathering about how great web standards are. As you can see, when you break down they numbers, the case isn't always there.
Well, I do, but 99.9% of the current web surfing population certianly doesn't remember HTML 1.0. They started using the web well into the age of tables, and therefore have certain expectations about websites, one being accurate layout.
While a "degraded" site may be pefectly acceptable on a cellphone or lynx, it's going to be rarely adequate for anything the client considers to be a popular PC web browser.
To be fair, the modern Single UNIX Specifications cover much more ground (and are more "useful" for application writers) than the original POSIX.1 spec that WinNT went through the motions with.
I think the GP's point was more along the lines of it being useless as a RFP requirement -- federal customers didn't actually care if it was there or not, and if they did, they wouldn't be shopping for Windows to begin with.
Well, I think you somewhat have a point, but Microsoft is playing an entirely different ballgame than Apple. MS is really looking at all of the enterprise server systems migrating to Linux instead of Windows.
On a technical level, "Unix" was important for Mac users because it gave them the robust, modern kernel that Apple had failed to develop in-house. Only a tiny minority of Mac users actually care about running awk or vi - they're just happy they can copy files without the machine grinding to a crawl.
Microsoft already has the underlying OS technology, so SFU is more of a bolted-on-top thing for program compatibility. From a system-administration standpoint, it's still Windows, not Unix.
Although the one thing that Mac OS X has shown is that there's a certain class of desktop users that don't want the full-meal-deal Unix environment like Linux or Solaris, but like to have sh or perl handy next to their MS Office install. I can't see why it would hurt Microsoft to appease these folks.
Slashdot Mirror
COM is an interface standard, not a API, and it's not specific to Windows or x86 - it has been ported elsewhere. Writing portable code is an old trick, the point is to have portable binaries. If Apache supported this it would make module packaging and installation much better on all platforms. Getting to recompile Apache is a hack, not a feature.
> Apache works fine on Windows - very, very well, actually
Apparently the folks in DC disagree.
People have no problem using the open software, but they continue to have small compatibility issues that keep them falling back to Microsoft.
:P
If coordinated IT efforts like City Of Munich and their IBM Con$ultant$ can't adequately solve these "small compatibility issues", it's a lot to ask of the very stretched education IT budgets.
teach kids how to use Linux and the spirit of open source and the "real world" will eventually change.
Maybe you should remind your wife that teaching religion is schools is illegal
Funny, because I worked with a MSCE guy who would endlessly blame any network problems on a couple castoff VA Linux boxes.
> I think we can blame Windows for that.
Blame whoever you want, but Windows has many production web servers, and Apache is not one of them.
> There is, in fact, a standard for such things. It's called Posix.
Apparently Windows 2003 R2 (shipping soon) will have fairly complete POSIX support. The entire rationale for the Apache 2 branch was better Windows support -- if they still haven't got it working, maybe it would be easier for them if they junked it and just targetted the W2003 portability layer.
> Further, Windows doesn't have a source-based build system at all! Shouldn't it?
Windows has a binary standard which allows code to 'snap together like legos', so why would it need the kludge of an entire development and build environment on production systems? There seems to be widespread confusion in the *nix world -- 'make' is not an installer. You have things like XPCOM -- use them.
The [Windows Update] ActiveX control is able to exceed user's security privileges to provide back certain system information that might not otherwise be possible to do otherwise for a given user.
Umm, no. Windows Update can only be run by Administrator users, and administrators can (directly or indirectly) do anything to a system.
ActiveX has enough real problems that there's absolutely no need to manufacture ridiclous falsehoods in order to talk it down.
The trust ratings and user comments need to be safe from poisoning and therefore moderated
:)
Keep in mind that Kazaa was the run-away most popular filesharing client for years, despite all of the well-known spyware it came with.
If you want to moderate all of the "wrong" opinons or just plain spam on this proposed BBS, you might as well just skip a step and put the Cabal directly in charge. (Whether that would be mozilla.org is unlikely, I think.)
And since your proposal relies on hashes, browser support, and some sort of authority, you might as well accept that you've just proposed code signing and you agree with me
The upshot is to really do it right, you need operating system support for an application-based security model. No current desktop OS currently supports this kind of security -- they are all firmly rooted in the user-based security model inherited from time-sharing systems.
That means either waiting for Vista or waiting for someone to add this security model to Linux/X11. Hopefully Firefox (and other internet software packages) will mimic IE and also have "low-rights" support on Vista.
You are misinformed. That info bar was in SP2 betas for months before Firefox shamelessly ripped it off.
No, I don't think signing is a cure-all, but it does minimize one social exploit. Whatever you think about ActiveX, I've never heard about an evil control that pretends to be Windows Update or Macromedia Flash.
If firefox become popular, it's possible there would be a ton of fake "Ad Block" and "Tab Browser" extentions, and signing is pretty much the only way to stop it.
If you want to see an example of this in action, search Google for "eMule", the opensource filesharing client. About 90% of the links go to fake sites which are probably spyware-laden clients. Too bad the official Emule installer doesn't use Authenicode -- I would defiantly check it.
Now it would be nice if code-signing was extended so that things could be "Certified by So-N-So to be Spyware-Free!". But even then, if it's an open system, fake certifiers will come about.
Yup, Firefox isn't using enough memory right now -- let's run Bittorrent in the same process space too! Memory protection is overrated anyway.
How would a bittorrent:// protocol and accompanying extention provide a better experience the current system of http, a torrent file, and a stand-alone client? It's only two fricking clicks to download something -- not like that's preventing people from installing Linux.
Correction -- ActiveX Distribution is officially called "Internet Component Download". Again, this is the thing people dislike the most about IE, not necessarily "ActiveX".
o ad/overview/overview.asp
http://msdn.microsoft.com/workshop/delivery/downl
Good point -- it always helps to clear up the termonology before diving too deep into a flamewar. Mozilla has developed a bunch of technologies that have rough equivilance to IE tech:
Netscape Plugins =~ ActiveX control
XPInstall =~ "ActiveX Web Distribution" (may not be the official name)
Firefox Extentions =~ Browser Helper Objects (BHOs)
The confusion I think is that most BHOs use ActiveX Distribution as the installation mechanism.
(And the other confusin is that MS has defined the term "ActiveX" in 9 different ways. Tons of stuff in Windows use COM/ActiveX, but the think people bitch about is the installation mechanism.)
Or not so important, because Windows users generally run with administration privledges, and that's where the virus problem lies.
Because in theory, someone educated enough to run Firefox would also be educated enough to not allow it to run untrusted things.
In theory, Firefox is a browser for the masses and is designed to supplant Internet Explorer. If Firefox has a userbase that's more technically sophisticated than other browsers, that only means that there's more work to do.
So please quit blowing yourself by thinking Firefox is l33t d00d software -- it isn't. The whole goal is stripped down and simple for the ordinary IE user.
Now it is true that Extentions are "elite", and they are generally only found on one or two sites. The questions is if the security model will hold up when Firefox gets more popular and users get used to instaling extentions from a varity of sources. I'm sure at some pont a signing mechanism like Authenicode will be deemed necessary.
Probably a better way to put it is that most of the wide-spread Mac viruses died with System 7 through 7.1.
Other than the Autostart worm, there hasn't been a real virus problem on Macs since about 1990 or so. Before then, they were quite prevelant.
> In fact, it's an Application in the form of a .app directory.
Even that's too much trouble. Just create a old-style Carbon binary (CFM?), set the file type to APPL, and the file extention will be ignored. (MacOS didn't have the concept of extentions until OS X) Give it the stock JPEG icon and your application will be virtually indistigishable from a regular JPEG.
> By that metric, both Germany and Japan are still occupied nations
Germany was technically under occupation until reunification in 1992 or so.
Japan left legal occupation significantly eariler, and could change their constitution if they wanted to (Article 96), although apparently there's significant popular support for the peace provision.
I certainly don't think it's a matter of resources. It's the marketing problem -- cell phone manufacturers are basically bound to the service providers, and can't compete with their services (like with iTMS). This runs contrary to Apple's mode of operation, where they basically control their entire channel.
I've talked to a couple who bragged about the "UNIX-based OS", when I'm damn sure they barely know what a 2 button mouse is, much less UNIX.
BSD TCP? I think your head is shoved up too far up inside your computer case.
It has nearly nothing to do with technology and everything to do with Kazaa's business model.
> But in Minnesota we don't consider ice storms natural disasters
Really? Apparently Duluth suffered a "disaster" in 1999 due to severe storms. I wouldn't be suprised if such declarations were reasonably common.
http://www.fema.gov/news/event.fema?id=374
Number are totally baloney because Slashdot is served gzipped.
/. doesn't look so hypocritical it's daily blathering about how great web standards are. As you can see, when you break down they numbers, the case isn't always there.
For this example, the total yearly savings for Slashdot would be: $3,650 USD!
OOH, three whole thousand dollars!!! Going to be a great XMas party this year!
Seriously, are you impressed with that? This ain't a mom-n-pop operation. I wouldn't be surprised if the site redesign cost VA Linux $100K. Even if it cost half that, and even with bullshit figures, your cost analysis falls right on its ass.
About the only business reason to justify such a redesign is so
Remember HTML 1.0?
Well, I do, but 99.9% of the current web surfing population certianly doesn't remember HTML 1.0. They started using the web well into the age of tables, and therefore have certain expectations about websites, one being accurate layout.
While a "degraded" site may be pefectly acceptable on a cellphone or lynx, it's going to be rarely adequate for anything the client considers to be a popular PC web browser.
To be fair, the modern Single UNIX Specifications cover much more ground (and are more "useful" for application writers) than the original POSIX.1 spec that WinNT went through the motions with.
I think the GP's point was more along the lines of it being useless as a RFP requirement -- federal customers didn't actually care if it was there or not, and if they did, they wouldn't be shopping for Windows to begin with.
Well, I think you somewhat have a point, but Microsoft is playing an entirely different ballgame than Apple. MS is really looking at all of the enterprise server systems migrating to Linux instead of Windows.
On a technical level, "Unix" was important for Mac users because it gave them the robust, modern kernel that Apple had failed to develop in-house. Only a tiny minority of Mac users actually care about running awk or vi - they're just happy they can copy files without the machine grinding to a crawl.
Microsoft already has the underlying OS technology, so SFU is more of a bolted-on-top thing for program compatibility. From a system-administration standpoint, it's still Windows, not Unix.
Although the one thing that Mac OS X has shown is that there's a certain class of desktop users that don't want the full-meal-deal Unix environment like Linux or Solaris, but like to have sh or perl handy next to their MS Office install. I can't see why it would hurt Microsoft to appease these folks.