The FCC (as opposed to the FTC) regulates calls made by the industries you list. The FCC recently voted to support the registry, so that loophole has been closed.
This isn't the forum for a detailed discussion, but I actually prefer enumeration classes now over simple enumerations. The basic idea is to write a class something like:
final class Color {
[snip]
static final Color RED = new Color("red");
static final Color BLUE = new Color("blue");
static final Color GREEN = new Color("green"); }
This idiom works great in some situations, but can be an absolute nightmare in others - see this article for some examples. Real enumerators address some problems that can't be solved using this mechanism.
The type-safe enum pattern shows the correct way of handling enumerations. And you can the Jakarta Commons Lang library [apache.org] to make it a bit easier.
The typesafe enum idiom is broken. It looks pretty but it's not a substitute for real first-class enumerators. In many cases it is actually worse than using plain ints as it can lead to very subtle bugs.
This article discusses some of the problems with it.
The XML is really an completely inappropriate use of it, and makes build files hard to work with and understand.
This is nothing more than opinion - you offer no arguments whatsoever to back it up. I can't believe it got moderated to +5.
Personally, I'm not crazy about reading raw XML in general, but it's not difficult at all in the case of Ant buildfiles. The structure is not very complicated and using XML makes it easy to use a validating editor to check your build files (there are DTDs that work for simple buildfiles out there). This also means that it is easy to write tools to manipulate the build files (XML is easily scriptable and there are lots of parsers already available). Try writing a tool from scratch to deal with Makefile syntax in an hour or two - not very easy. Using XML also makes sense from the standpoing of being able to easily integrate new custom tasks into the tool - each new task doesn't have to invent its own funky syntax. So, I don't think using XML buildfiles is such a bad decision.
Ant is basically a cross-platform scripting language...
No, it's not. It's a build tool and deserves to be evaluated as such. If you're willing to make such blatantly false claims, I wonder how much you really know about Ant or about scripting languages.
If you actually read the link in the story, you will see that it prevents you from using it if you are developping (or aiding development) of an alternative.
This actually means that if I am both a subversion hacker and a kernel hacker, I can't use BitKeeper anymore for my kernel hacking....
Right, and if you read it closely, it also says that if your company develops a configuration management program (even if it's for in-house use and even if you don't work on it), you can't use BitKeeper without buying a license. That means that if you are (for instance), an IBM employee working on the kernel in your spare time, you're hosed. The same applies to other companies, I'm sure.
I think it's fair and it definitely makes sense from BitMover's perspective, but it also shows why it was a mistake for Linus to ignore the implications of requiring (or close to it) the use of commercial software to develop the kernel. Who knows how many more license changes there are in the pipe?
I'm all about the live music experience, but you do realize that small acts make almost nothing from playing clubs and so forth, right? It's fun, but you don't make a lot of money doing it. It seems to me like all your "policy" is going to do is make it financially impractical for a lot of bands to record CDs and tour - and we're not even talking about signed bands.
Besides, is propping up TicketMaster really more noble than supporting your favorite indie artist by buying their CDs?
Maybe I'm wrong, but I don't think this is the solution.
There is no right or wrong way to determine which window recieves focus based on where the mouse is.
No, but usability testing would probably indicate which method is easiest for certain groups of users, from total newbies, to windows users, to CDE users.
So what? That's irrelevent to the point the parent was making unless you're targeting a specific set of these groups. And at that point you're making a design choice, not a technical choice.
Actually, that's exactly what I'm saying. English is crack, too, as are ALL languages, really. The only languages that could ever qualify as crack free are human constructed languages, like Newspeak. Newspeak is flawed in that it doesn't offer a large enough vocabulary, but who can deny that its straight-foward approach and completely uniform grammar are less useful for communicating ideas than the tremendous over-complexity of all non-constructed languages?
I certainly don't deny that it's less useful.;-)
Seriously, though, this is silly. I think it's a question of philosophy to a large extent, but your comment about poetry basically negates your point to my mind. The nuances of human language used in poetry do communicate ideas, and in a way that languages that are not "overly complex" can't do. "Aesthetics" as you put it have meaning and value in terms of communication of ideas. We are not computers after all. While I have to admit that I'm not familiar with Newspeak, I can't imagine how an artificial language with a completely uniform grammar and vocabulary could be more expressive than a natural language which all parties involved understand equally well. It may be easier to understand or clearer (in some sense), but I think that's more a function of having a very limited syntax and vocabulary. Besides, as humans, we're innately pretty good at expressing ourselves. It seems likely to me that the evolution of human language over millions of years would select for whatever makes it easiest for people to get their ideas across to each other. I just don't believe that you or anybody else is going to start from scratch and come up with something better than every existing natural language.
As for the rest of your comment, I disagree to some extent, but I think you make some interesting points. I couldn't pass this up, though.
We need people who have access to RoadRunner accounts to start making wild accusations about well-known, established, and notoriously litigious companies. After the first wave of defamation suits hits, they will reconsider the position.
You bet they will! Oh wait, you mean Roadrunner?
Unless these people start posting to alt.religion.scientology, nobody is going to waste money on such a frivolous suit against AOL/TW.
tends to cook my legs very well now. If it were able to conduct heat any more efficiently I believe it would become too painful to use. On the other hand that kind of thing could lead to a lawsuit that would get me out of the daily grind and into the life of luxury I deserve.
Get a lap desk. Keeps your legs cool, plus your computer is more stable and has better air flow.
Once you've done that, *then* you can come back and make silly assertions.
Ruby has some nice features, but it'll never be more than a tiny niche language, regardless of whatever is going on in Japan.
Why don't you take your own advice?
There are way too may people making silly assertions on both side of the fence. Different people are interested in different problem domains. People work and think in different ways, so what is a productive environment for one person may not be for another. Your lack of productivity with Ruby means no more than the previous poster's lack of productivity with Python. As far as niches go, I've got news for you - both of them are niche languages. At least they're niches compared to some other general purpose languages like Java or C++. Be happy that you have a chance to use Python in your work; most people don't get that opportunity. I guess I'm just argumentative tonight, but people in the Ruby community have (for the most part) already bored of constantly carrying on this same tiresome argument about Python and Perl. It's just not interesting anymore.
So GPL can't be used by business eh? This would be because obviously reading and understanding the code, then rekeying it in so that it's slightly different would be waaay to much work for a company who wants to profit off of something they got for free...
That would qualify as a modification and be still be protected by the GPL. So, $200 for WinXP is 4% the price of a $500 PC. Great math there Bill.
PC vendors don't pay retail price for the OS they install on their machines.
Yeah, as a GNOME person the fonts look all blurry and hard to read, too.;)
Seriously, tho, to someone with a Windows background and no UN*X background (the Lycoris/Redmond Linux target audience, I'm guessing) the distinction probably seems natural. Just like associating an icon of a hard drive with every partition on your system.;) I think it's a somewhat reasonable decision for Lycoris to make. The easiest thing to use, at least initially, is always going to be whatever is most like what you're used to. It may not be the best technical solution, but familiarity is a good thing if you're trying to make converts of people.
In any case, 90% of what CN broadcasts are reruns; I'm sure that viewership of reruns is where cartoon network gets most of its ad revenue - even for CN original programming. Space Ghost and Dexter's Lab, although available from Kazaa, haven't yet reached the penetrance of Tenchi Muyo, so I'm not tempted to just watch downloaded versions instead.
That's probably true about the reruns, but I think it has as much to do with the format as anything else. Given the number/frequency of CN shows they run, it's not surprising that they're going to run low on content. Animation is more time-intensive to create than the average syndicated sitcom after all. "Very few cartoons are done live, Mr. Simpson. It puts a terrible strain on the animators.":)
Personally, however, p2p really has damped my interest in CN anime offerings - pretty much completely. For shows that I like but don't feel the need to watch as soon as they're newly broadcast - Space Ghost is a prime example - I'd prefer p2p as well, if they were widely available. I think this is a trend.
I think you're right, but what you're really talking about is video (or content) on demand. p2p is just one possible mechanism for delivering it (although right now it is one of the most attractive). Tivo is a step in this direction too. Obviously, it's going to be much more attractive to watch what you want to watch when you want to watch it than to be tied to static content decided on by broadcast execs. The real question is who is going to control what is made available. The attraction of p2p over the other delivery mechanisms available right now is that it puts this control in the hands of the end user.
I can get all the cartoons I want by p2p. In the case of "the Simpsons", I can see a new episode sooner if I turn on the boob tube (well, my tuner card, anyway) but I already see most of these toons (the ones I actually want to watch) subtitled by fans long before cartoon network gets around to rebroadcasting them sans anything raunchy and with lousy voice talent.
It isn't even that much trouble - taken as a total, turning on whatever software and running it overnight is less time than I would spend watching commercials if I watched them on TV. Why would I ever?
Seriously; what value added does Cartoon Network provide?
There is more to Cartoon Network than anime. Space Ghost, Dexter's Lab, Powerpuff Girls, Samurai Jack, etc., etc... all of them are Cartoon Network originals. Your tastes may differ, but these are all fairly high quality shows, particularly the first 3, and none of them would exist w/o Cartoon Network. I think that qualifies as value added. Note that I'm not against anime, I just personally don't give a rat's flea-bitten ass about it. Also, I find that there's something about kicking back on the couch with a beer and watching a real TV that watching an MPEG or DivX on the computer just doesn't provide. YMMV, tho...
I've been working on a computationally expensive project where I see gcc-3.0.3 (with CFLAGS = -O2 -mcpu=ultrasparc) beat out Sun Workshop 5.0 (CFLAGS = -fast).
This is an interesting result. I've looked at Sun's benchmarks for gcc 2.95.3 against Sun Forte 6 (the sucessor to Workshop 5) and, predictably enough, Sun's compiler was generally better - sometimes much better. gcc did reasonably well though, considering that (as with intel) Sun produced both the compiler and the architecture. Here's the link.
It's also worth mentioning that these tests were done using C++ as the implementation language, so it's very possible that gcc 3 would do better than 2.95. My understanding is that C++ support is better in gcc 3.
Workshop 5, on the other hand, is a buggy piece of junk. We're currently using it on the project I'm working on and we're having to jump through all kinds of hoops to keep it from miscompiling the program. gcc 2.95 on Linux and VAC++ for AIX both work fine.
Yet Russ Cooper posted this exploit AND links to the code earlier this morning to NTBugTraq. If they are so irresponsible, then why didn't he edit/moderate the content they submitted?
According to the quote I saw from Cooper (in the same article, IIRC), he decided not to moderate their content since it was released simultaneously through a number of other channels. The genie was already out of the bottle. Makes some sense from a newsworthiness standpoint, I guess.
That's nice to say, but all current window managers currently suck in some way, and I don't ever see that changing. All we need is a more specific standard for talking between apps and deciding on icons, menus, drag/drop, copy/paste, etc (and some good docs for writing window managers using ICCCM, since I'm quite confused).
You obviously are a little confused. All of the improvements you mention relate to desktop environments rather than window managers. IMO there are a number of perfectly capable Linux window managers - Sawfish, kwin, and Enlightenment, among others. The desktop environments are a whole other thing, though and I think your comments in that direction are good ones.
even if they are generated by server side code, the server still outputs HTML. The HTML can be saved to disk.
In the case of sites that use the URL Query String to fetch pages out of the database. As long as the 'bot can physically cache the resultant file to disk and maintain a table
of hard-coded URL Query Strings that point to the saved file, it shouldn't be much of a problem.
I really don't think it's possible. Think about this for a second. The output generated by a particular database query is dynamic, remember. The output generated by a such a query is worthless for the purpose of mirroring because its content is good only for so long as the database state is the same. Think about a page with a search form. Or a submission field that changes the database state. You're screwed. Without a copy of the database, you just cannot mirror a DB driven site.
Even leaving aside the question of database state, query strings are not the only thing that can determine what HTML a particular person using a particular browser will see. What about pages with HTML generated by the client's IP or browser or whatever. There's no way you can simulate this without having the code.
Leaving aside the question of C++'s portability, it may have had something to do with the state of flux in the GTK+ C++ bindings. As I understand it, some questionable design decisions have been made with gtk--, and now there is Havoc Pennington's Inti API too.
Personally, I seriously doubt that the "must be written in the same language" thing had anything to do with it. That just doesn't make a lot of sense to me.
I'd never though about it, but this is really an excellent point.
Not only is there no standard binary format for C++ object files, there's really not a whole lot of standards compliance among compilers (yet). To write portable code, you have to aim for the lowest common denominator.
Also, as the poster pointed out, some vendors don't even supply a C++ compiler (or sell it only as an add on - e.g. Sun). Yes, g++ is out there and it's pretty good, but no vendor is going to package and support it just to run KDE, particularly since they presumably already have a desktop and they're trying to sell their own compiler. This isn't a problem with C.
These kids are at an age where they can be responsible for their own actions and where seeing pr0n probably won't scar them for life.
No, they're not. This is a junior high school. We're talking about children as young as 12 or 13. You may think they're responsible enough to have access to pornography, but I have a feeling most people (including their parents) would disagree.
It's a public facility and there's no more obligation to censor/filter material for minors than a magazine store that happens to sell copies of Playboy and Penthouse.
Whoa there. First of all, we're not talking about the likes of Playboy and Penthouse. Those are publications that are entirely legal pretty much everywhere. There's a huge difference between the kind of nudity featured in newsstand magazines and the kind of hardcore porn you can find on the web. If a 5 year old sees the cover of Playboy at Barnes and Noble, who cares. If he walks into a laundromat and sees a photograph of a woman having sex with a dog (or an autopsy photograph, or whatever) that some punk-ass teenager pulled up and left on the screen as a joke, there are going to be problems. The laundromat may not have a legal responsibility to filter that kind of stuff out, but you can be sure they're going to take a lot more heat than they want if they don't take some kind of action to prevent things like that from happening.
Here's another illustration. My summer job is providing support for public web terminals at a mall. We have filtering software for one simple reason - if we didn't and someone complained about their kid looking at porn, we'd be kicked out of the mall in about 5 seconds flat. It would happen - no question about it. It doesn't matter if it's fair or not. That's the way things work in the "real world."
I don't like filtering any more than anyone else, but the reality is that it really is necessary sometimes. If you're offering public web access, you're taking a huge risk if you don't at least make some effort to restrict content. No, filtering is not perfect and never will be. My point is that there are situations where it's the only answer.
Locking the tenent out was illegal. Not Locking anyone out was not.
Actually, it probably was. A landlord has a legal responsibility to provide a reasonable assurance of safety and security to his or her tenants - making sure all the doors can be locked, for example. The idea is to give people some legal recourse against slum lords and lazy landlords, I guess.
Anyway, I agree with everything else you said, just wanted to point that out.
In conclusion: Support the only true form of participatory democracy: Assassination.
Is this supposed to be funny? It's not. There are plenty of places where assassination is a political tool. Be glad you're not living in one of them. I know I am.
Slashdot Mirror
The "raise window" option is still there. You can turn it on using the message notification plugin.
You're wrong about that.
The FCC (as opposed to the FTC) regulates calls made by the industries you list. The FCC recently voted to support the registry, so that loophole has been closed.
This CNN article mentions it.
This isn't the forum for a detailed discussion, but I actually prefer enumeration classes now over simple enumerations. The basic idea is to write a class something like:
final class Color {
[snip]
static final Color RED = new Color("red");
static final Color BLUE = new Color("blue");
static final Color GREEN = new Color("green");
}
This idiom works great in some situations, but can be an absolute nightmare in others - see this article for some examples. Real enumerators address some problems that can't be solved using this mechanism.
The type-safe enum pattern shows the correct way of handling enumerations. And you can the Jakarta Commons Lang library [apache.org] to make it a bit easier.
The typesafe enum idiom is broken. It looks pretty but it's not a substitute for real first-class enumerators. In many cases it is actually worse than using plain ints as it can lead to very subtle bugs.
This article discusses some of the problems with it.
This is nothing more than opinion - you offer no arguments whatsoever to back it up. I can't believe it got moderated to +5.
Personally, I'm not crazy about reading raw XML in general, but it's not difficult at all in the case of Ant buildfiles. The structure is not very complicated and using XML makes it easy to use a validating editor to check your build files (there are DTDs that work for simple buildfiles out there). This also means that it is easy to write tools to manipulate the build files (XML is easily scriptable and there are lots of parsers already available). Try writing a tool from scratch to deal with Makefile syntax in an hour or two - not very easy. Using XML also makes sense from the standpoing of being able to easily integrate new custom tasks into the tool - each new task doesn't have to invent its own funky syntax. So, I don't think using XML buildfiles is such a bad decision.
No, it's not. It's a build tool and deserves to be evaluated as such. If you're willing to make such blatantly false claims, I wonder how much you really know about Ant or about scripting languages.
Right, and if you read it closely, it also says that if your company develops a configuration management program (even if it's for in-house use and even if you don't work on it), you can't use BitKeeper without buying a license. That means that if you are (for instance), an IBM employee working on the kernel in your spare time, you're hosed. The same applies to other companies, I'm sure.
I think it's fair and it definitely makes sense from BitMover's perspective, but it also shows why it was a mistake for Linus to ignore the implications of requiring (or close to it) the use of commercial software to develop the kernel. Who knows how many more license changes there are in the pipe?
I'm all about the live music experience, but you do realize that small acts make almost nothing from playing clubs and so forth, right? It's fun, but you don't make a lot of money doing it. It seems to me like all your "policy" is going to do is make it financially impractical for a lot of bands to record CDs and tour - and we're not even talking about signed bands.
Besides, is propping up TicketMaster really more noble than supporting your favorite indie artist by buying their CDs?
Maybe I'm wrong, but I don't think this is the solution.
So what? That's irrelevent to the point the parent was making unless you're targeting a specific set of these groups. And at that point you're making a design choice, not a technical choice.
I certainly don't deny that it's less useful.
Seriously, though, this is silly. I think it's a question of philosophy to a large extent, but your comment about poetry basically negates your point to my mind. The nuances of human language used in poetry do communicate ideas, and in a way that languages that are not "overly complex" can't do. "Aesthetics" as you put it have meaning and value in terms of communication of ideas. We are not computers after all. While I have to admit that I'm not familiar with Newspeak, I can't imagine how an artificial language with a completely uniform grammar and vocabulary could be more expressive than a natural language which all parties involved understand equally well. It may be easier to understand or clearer (in some sense), but I think that's more a function of having a very limited syntax and vocabulary. Besides, as humans, we're innately pretty good at expressing ourselves. It seems likely to me that the evolution of human language over millions of years would select for whatever makes it easiest for people to get their ideas across to each other. I just don't believe that you or anybody else is going to start from scratch and come up with something better than every existing natural language.
As for the rest of your comment, I disagree to some extent, but I think you make some interesting points. I couldn't pass this up, though.
You bet they will! Oh wait, you mean Roadrunner?
Unless these people start posting to alt.religion.scientology, nobody is going to waste money on such a frivolous suit against AOL/TW.
Get a lap desk. Keeps your legs cool, plus your computer is more stable and has better air flow.
Ruby has some nice features, but it'll never be more than a tiny niche language, regardless of whatever is going on in Japan.
Why don't you take your own advice?
There are way too may people making silly assertions on both side of the fence. Different people are interested in different problem domains. People work and think in different ways, so what is a productive environment for one person may not be for another. Your lack of productivity with Ruby means no more than the previous poster's lack of productivity with Python. As far as niches go, I've got news for you - both of them are niche languages. At least they're niches compared to some other general purpose languages like Java or C++. Be happy that you have a chance to use Python in your work; most people don't get that opportunity. I guess I'm just argumentative tonight, but people in the Ruby community have (for the most part) already bored of constantly carrying on this same tiresome argument about Python and Perl. It's just not interesting anymore.
So GPL can't be used by business eh? This would be because obviously reading and understanding the code, then rekeying it in so that it's slightly different would be waaay to much work for a company who wants to profit off of something they got for free...
That would qualify as a modification and be still be protected by the GPL.
So, $200 for WinXP is 4% the price of a $500 PC. Great math there Bill.
PC vendors don't pay retail price for the OS they install on their machines.
They don't develop it or sell it anymore, but everyone is forgetting that MS once sold Windows NT for the DEC Alpha. That's a 64-bit architecture.
Of course, I don't know how well it worked or whether MS abandoned it for technical reasons or because NT Alpha machines weren't selling.
Yeah, as a GNOME person the fonts look all blurry and hard to read, too. ;)
;) I think it's a somewhat reasonable decision for Lycoris to make. The easiest thing to use, at least initially, is always going to be whatever is most like what you're used to. It may not be the best technical solution, but familiarity is a good thing if you're trying to make converts of people.
Seriously, tho, to someone with a Windows background and no UN*X background (the Lycoris/Redmond Linux target audience, I'm guessing) the distinction probably seems natural. Just like associating an icon of a hard drive with every partition on your system.
That's probably true about the reruns, but I think it has as much to do with the format as anything else. Given the number/frequency of CN shows they run, it's not surprising that they're going to run low on content. Animation is more time-intensive to create than the average syndicated sitcom after all. "Very few cartoons are done live, Mr. Simpson. It puts a terrible strain on the animators."
I think you're right, but what you're really talking about is video (or content) on demand. p2p is just one possible mechanism for delivering it (although right now it is one of the most attractive). Tivo is a step in this direction too. Obviously, it's going to be much more attractive to watch what you want to watch when you want to watch it than to be tied to static content decided on by broadcast execs. The real question is who is going to control what is made available. The attraction of p2p over the other delivery mechanisms available right now is that it puts this control in the hands of the end user.
There is more to Cartoon Network than anime. Space Ghost, Dexter's Lab, Powerpuff Girls, Samurai Jack, etc., etc... all of them are Cartoon Network originals. Your tastes may differ, but these are all fairly high quality shows, particularly the first 3, and none of them would exist w/o Cartoon Network. I think that qualifies as value added. Note that I'm not against anime, I just personally don't give a rat's flea-bitten ass about it. Also, I find that there's something about kicking back on the couch with a beer and watching a real TV that watching an MPEG or DivX on the computer just doesn't provide. YMMV, tho...
This is an interesting result. I've looked at Sun's benchmarks for gcc 2.95.3 against Sun Forte 6 (the sucessor to Workshop 5) and, predictably enough, Sun's compiler was generally better - sometimes much better. gcc did reasonably well though, considering that (as with intel) Sun produced both the compiler and the architecture. Here's the link.
It's also worth mentioning that these tests were done using C++ as the implementation language, so it's very possible that gcc 3 would do better than 2.95. My understanding is that C++ support is better in gcc 3.
Workshop 5, on the other hand, is a buggy piece of junk. We're currently using it on the project I'm working on and we're having to jump through all kinds of hoops to keep it from miscompiling the program. gcc 2.95 on Linux and VAC++ for AIX both work fine.
According to the quote I saw from Cooper (in the same article, IIRC), he decided not to moderate their content since it was released simultaneously through a number of other channels. The genie was already out of the bottle. Makes some sense from a newsworthiness standpoint, I guess.
You obviously are a little confused. All of the improvements you mention relate to desktop environments rather than window managers. IMO there are a number of perfectly capable Linux window managers - Sawfish, kwin, and Enlightenment, among others. The desktop environments are a whole other thing, though and I think your comments in that direction are good ones.
In the case of sites that use the URL Query String to fetch pages out of the database. As long as the 'bot can physically cache the resultant file to disk and maintain a table of hard-coded URL Query Strings that point to the saved file, it shouldn't be much of a problem.
I really don't think it's possible. Think about this for a second. The output generated by a particular database query is dynamic , remember. The output generated by a such a query is worthless for the purpose of mirroring because its content is good only for so long as the database state is the same. Think about a page with a search form. Or a submission field that changes the database state. You're screwed. Without a copy of the database, you just cannot mirror a DB driven site.
Even leaving aside the question of database state, query strings are not the only thing that can determine what HTML a particular person using a particular browser will see. What about pages with HTML generated by the client's IP or browser or whatever. There's no way you can simulate this without having the code.
Personally, I seriously doubt that the "must be written in the same language" thing had anything to do with it. That just doesn't make a lot of sense to me.
Not only is there no standard binary format for C++ object files, there's really not a whole lot of standards compliance among compilers (yet). To write portable code, you have to aim for the lowest common denominator.
Also, as the poster pointed out, some vendors don't even supply a C++ compiler (or sell it only as an add on - e.g. Sun). Yes, g++ is out there and it's pretty good, but no vendor is going to package and support it just to run KDE, particularly since they presumably already have a desktop and they're trying to sell their own compiler. This isn't a problem with C.
Here's another illustration. My summer job is providing support for public web terminals at a mall. We have filtering software for one simple reason - if we didn't and someone complained about their kid looking at porn, we'd be kicked out of the mall in about 5 seconds flat. It would happen - no question about it. It doesn't matter if it's fair or not. That's the way things work in the "real world."
I don't like filtering any more than anyone else, but the reality is that it really is necessary sometimes. If you're offering public web access, you're taking a huge risk if you don't at least make some effort to restrict content. No, filtering is not perfect and never will be. My point is that there are situations where it's the only answer.
Anyway, I agree with everything else you said, just wanted to point that out.