It would be a great win if the free software community came up with something like this. If not, I fear that government-imposed rating, or an ugly proprietary scheme, is inevitable. Further, the system must be integrated into the browser if you want enough participation to achieve reasonable coverage. So, clearly mozilla integration is a great opportunity.
I would highly encourage anyone interested to begin thinking about how to make this part of mozilla.
Modulo XML and some small details, what you describe is what already exists. The hard problems are the semantic ones, and XML won't help you a wit there. Nor will XML magically create you a nice, usable interface.
Me too, and I'm still trying to figure out whether the cigarette burns in Fight Club were part of the movie. I must say I left the theatre in a disturbed state.
> It is precisely because security holes in Linux > are announced openly that it is so secure.
Oh, please. Do you actually believe that Alan Cox posts every security bug he finds out about to linux-kerel before he (or someone) fixes it? Go ask him, but I can tell you the answer in advance.
Entrusting security bugs to a respected core group (including not just Netscape employees, I hope) is a perfectly acceptible compromise between open source principles and user safety. Because the source is open, it will almost certainly come to light if they violate this trust. This is strong incentive for the security group to act promptly and responsibly.
Keep in mind that it's quite possible you're seeking BSD's smaller (less featureful or less bloated, depending on your perspective) userland. That makes things faster, too. I imagine that this would be more noticeable to most users than kernel differences.
If you flatly reject all IPC (inter-process communication) as "a can of worms", you're probably stuck. But IPC can be done properly, and is very powerful. A daemon listening on a unix domain socket can reliably verify the identity of the other end of the socket (via the kernel), perform strict and careful validation on the input, and carefully update the user database.
As the microkernel unix people used to say before they finally died out, "We are within X% of the performance of the original".
- Alan Cox
The JIT guys can write all the papers in the world about how they should (yes, really should) be able to get native code speed. "No, that extra level of indirection is really a performance benefit, not a hindrance! I'll believe it when I run Java bytecode and it's not slow.
[slashdot bug reformatted the last post] As the microkernel unix people used to say before they finally died out, "We are within X% of the performance of the original". - Alan Cox The JIT guys can write all the papers in the world about how they should (yes, really should) be able to get native code speed. No, that extra level of indirection is really a performance benefit, not a hindrance! I'll believe it when I run Java bytecode and it's not dog slow.
As the microkernel unix people used to say before they finally died out, "We are within X% of the performance of the original". - Alan Cox The JIT guys can write all the papers in the world about how they should (yes, really should) be able to get native code speed. I'll believe it when I run Java bytecode and it's not slow.
Lynx is a terribly twisted bit of code. Many security holes have been found, most of them suggesting bad design. I guarentee there are many more to be found if anyone looks. I would be petrified to surf with Lynx if anyone cared to target exploits to it.
In answer to roblimo's question, I find that Linux Weekly News (www.lwn.net) already does an excellent job of such roundups. This isn't something slashdot needs to duplicate.
Read again what I said about "half-as-good" free alternatives. That was an important part of the claim.
VMWare had no free competition in sight when it became popular. I gather that Win4Lin is similar to VMWare, so the same applies. We will see whether this changes as Plex86 becomes usable, although head-starts carry some weight.
StarOffice, in my impression, is not that popular. I doubt it would have lasted long as proprietary software (and it may not last long even now that it's free). Most people in my experience use it grudgingly for MS compatibility. Again, no free program offered decent MS compatibility until recently (even now, many Excel and Word documents are not read properly by any free software I know of).
Netscape is an obvious case: there were no decent free graphical browsers until recently. Netscape had a long reign as the only reasonable option, which gives it lots of momentum. If Netscape 4 and mozilla were both new today, Netscape 4 wouldn't have a chance (on Linux).
Games are harder for me because I'm not much of a gamer, and because there are so many genres of game (not to mention that individual games periodically create their own genre). What proprietary games are you thinking of that have taken off on Linux? Shooters, and action games in general, had no free competition (Doom and Quake were eventually freed, but under unusual circumstances; I don't know how this has affected the popularity of newer proprietary shooters, do you?). The best case I can think of for your side is the Civilization genre. Freeciv qualifies as "half-as-good", and is quite popular, but I don't much about the popularity of the Loki Civ ports. You may score a point on this one, but I didn't doubt you'd find some example.
Here's the central argument: the free software community has two desktop/office projects that, while immature, clearly have basic functionality, strong developer support, and "long-term credibilty" (to use a term from the Halloween Document). Under these circumstances, I claim that no proprietary product that significantly overlaps with them will be successful on Linux.
Slashdot Mirror
It would be a great win if the free software community came up with something like this. If not, I fear that government-imposed rating, or an ugly proprietary scheme, is inevitable. Further, the system must be integrated into the browser if you want enough participation to achieve reasonable coverage. So, clearly mozilla integration is a great opportunity.
I would highly encourage anyone interested to begin thinking about how to make this part of mozilla.
Modulo XML and some small details, what you describe is what already exists. The hard problems are the semantic ones, and XML won't help you a wit there. Nor will XML magically create you a nice, usable interface.
Me too, and I'm still trying to figure out whether the cigarette burns in Fight Club were part of the movie. I must say I left the theatre in a disturbed state.
> It is precisely because security holes in Linux
> are announced openly that it is so secure.
Oh, please. Do you actually believe that Alan Cox posts every security bug he finds out about to linux-kerel before he (or someone) fixes it? Go ask him, but I can tell you the answer in advance.
Entrusting security bugs to a respected core group (including not just Netscape employees, I hope) is a perfectly acceptible compromise between open source principles and user safety. Because the source is open, it will almost certainly come to light if they violate this trust. This is strong incentive for the security group to act promptly and responsibly.
Keep in mind that it's quite possible you're seeking BSD's smaller (less featureful or less bloated, depending on your perspective) userland. That makes things faster, too. I imagine that this would be more noticeable to most users than kernel differences.
If you flatly reject all IPC (inter-process communication) as "a can of worms", you're probably stuck. But IPC can be done properly, and is very powerful. A daemon listening on a unix domain socket can reliably verify the identity of the other end of the socket (via the kernel), perform strict and careful validation on the input, and carefully update the user database.
[slashdot bug reformatted first two posts]
- Alan CoxThe JIT guys can write all the papers in the world about how they should (yes, really should) be able to get native code speed. "No, that extra level of indirection is really a performance benefit, not a hindrance! I'll believe it when I run Java bytecode and it's not slow.
[slashdot bug reformatted the last post] As the microkernel unix people used to say before they finally died out, "We are within X% of the performance of the original". - Alan Cox The JIT guys can write all the papers in the world about how they should (yes, really should) be able to get native code speed. No, that extra level of indirection is really a performance benefit, not a hindrance! I'll believe it when I run Java bytecode and it's not dog slow.
As the microkernel unix people used to say before they finally died out, "We are within X% of the performance of the original". - Alan Cox The JIT guys can write all the papers in the world about how they should (yes, really should) be able to get native code speed. I'll believe it when I run Java bytecode and it's not slow.
You actually might be even more disgusted ;)
Lynx is a terribly twisted bit of code. Many security holes have been found, most of them suggesting bad design. I guarentee there are many more to be found if anyone looks. I would be petrified to surf with Lynx if anyone cared to target exploits to it.
But I still love it. Slick browser.
In answer to roblimo's question, I find that Linux Weekly News (www.lwn.net) already does an excellent job of such roundups. This isn't something slashdot needs to duplicate.
VMWare had no free competition in sight when it became popular. I gather that Win4Lin is similar to VMWare, so the same applies. We will see whether this changes as Plex86 becomes usable, although head-starts carry some weight.
StarOffice, in my impression, is not that popular. I doubt it would have lasted long as proprietary software (and it may not last long even now that it's free). Most people in my experience use it grudgingly for MS compatibility. Again, no free program offered decent MS compatibility until recently (even now, many Excel and Word documents are not read properly by any free software I know of).
Netscape is an obvious case: there were no decent free graphical browsers until recently. Netscape had a long reign as the only reasonable option, which gives it lots of momentum. If Netscape 4 and mozilla were both new today, Netscape 4 wouldn't have a chance (on Linux).
Games are harder for me because I'm not much of a gamer, and because there are so many genres of game (not to mention that individual games periodically create their own genre). What proprietary games are you thinking of that have taken off on Linux? Shooters, and action games in general, had no free competition (Doom and Quake were eventually freed, but under unusual circumstances; I don't know how this has affected the popularity of newer proprietary shooters, do you?). The best case I can think of for your side is the Civilization genre. Freeciv qualifies as "half-as-good", and is quite popular, but I don't much about the popularity of the Loki Civ ports. You may score a point on this one, but I didn't doubt you'd find some example.
Here's the central argument: the free software community has two desktop/office projects that, while immature, clearly have basic functionality, strong developer support, and "long-term credibilty" (to use a term from the Halloween Document). Under these circumstances, I claim that no proprietary product that significantly overlaps with them will be successful on Linux.