IAC AYT is no good because the response arrives as NVT data, but either end could use option negotiation for liveness (pick some harmless option, send IAC WILL FOO and expect IAC DO/DONT FOO).
HTTP with WebDAV offers type negotiation, transparent compression, connection reuse (hardly any FTP clients can use MODE B or MODE C), and cache coherence. Unless you're aware of something FTP still does better, it really is obsolete.
HTTP proxies and NATs tend to time out idle TCP connections. As usual, without micropayments, there's no way to give all the providers between incentive to offer higher quality of service.
But responses have to be one-to-one with requests. A While clients can send overlapped requests (but a proxy might not forward them overlapped!), servers MUST send one complete response to each request in order--no prioritization or unsolicited messages are possible, and since "server push" is implemented as one big multipart response, the client can give no feedback about it other than unexpectedly closing the connection.
IETF recommends that browsers make disabling cookie support possible for privacy (for the same reason I don't walk around with my True Name tattooed on my forehead). Hidden fields are more reliable and don't represent the same threat.
The World-Wide Web Consortium doesn't have any particular influence over the Motion Picture Experts Group, and video compression is such a minefield of patents that a free standard is probably impossible (though I wish luck to the Ogg Tarkin team).
There's nothing wrong with making money. But since the World-Wide Web is about universal access and platform independence, if your business model involves controlling who may use a technology and how, it has no business being recommended by W3C.
See the intrinsic events module for onchange. Remember that XML is case-sensitive, and that user agents may not implement any particular scripting language or have scripting enabled (due to a history of security and privacy exploits in the standard sub-standard browsers).
"New keywords" is an odd complaint about a language that declares abstract methods using "= 0", gives at least three different meanings (non-instance member, no external linkage, not reinitialized at each funcall) to the keyword "static", and resorts to a dummy argument to distinguish prefix from postfix operators.
Did GCC actually produce incorrect code? I seem to recall reading that the KDE or Qt maintainers made brain-damaged assumptions ("the first virtual base class is stored at offset zero in any derived class" or something similar) and got upset when GCC 3 started giving them undefined behavior that was different than the undefined behavior they were expecting.
Don't forget, they ought to take responsibility not only for determining whether the current limit is short enough to be Constitutional but also whether adding twenty years to the limit every twenty years is still "limited" enough to be Constitutional.
Tabs? I don't get it. I already have a perfectly good window manager. Why do I need another one (with undocumented keystrokes, no less) that only applies to one app?
The spec is over 400 pages, and I've been meaning to read it one of these days, but it hasn't been a real priority for me yet--can you be more specific? There are certainly huge differences, but do any of them represent new capabilities (as opposed to "negative variety" due to not-invented-here syndrome)? Assemblies seem to be a workaround for vendors too brain-damaged to rename libraries when they violate old interface contracts, and unsafe code completely subverts the security model. What have they done that's actually an improvement?
Without Microsoft's proprietary classes,.NET offers nothing you couldn't already do with JDK1.1 for years (unless you want to embed whitespace and comments in your RPC messages).
Exchange is allegedly moving away from their own protocol towards WebDAV (mostly so that other apps can easily store documents in "public folders") but I agree it's unlikely Outlook will be allowed to interop well with a competing mail server.
Does Sun make a lot of money on Solaris licenses? I thought they were pretty much staying in the hardware business.
Seems to me you're blaming the firefighters for putting effort into hosing down your house during the emergency but failing to put equal effort into drying your belongings off afterwards. Their real job is to save the neighborhood, not any one house.
The surrogate codepoints (0xD800-0xDFFF) have long been reserved so that UTF-16 can encode codepoints out to U+10FFFF. Codepoints beyond U+FFFF haven't been assigned yet, but a few scripts using them have already been proposed for Unicode 3.2. It's almost certainly going to happen, and if java.lang.String still lets you chop a character in half then it's seriously broken.
I want users to have source. I want them to understand their software and make it do what they want. Even better, I want them to hire me to do it for them.
If some proprietary vendor wants to control their customers and eliminate work-for-hire (the ethical business model for software, just like other professional services) I refuse to subsidize them.
It was legal until very recently. Courts are just starting to rule "equitable estoppel" (basically, you can't do it because it's not fair) against patent holders who wait until an invention becomes an industry standard before pointing out their patent covering it.
If my patent is ruled to be invalid, that means it always was. Can you sue to recover the fraudulent royalties I exacted from you with my invalid patent? Aren't royalties usually put in escrow until the court case and appeals are resolved?
Cable companies do pay networks for the right to use their content, just like local broadcaster affiliates pay their networks. That's why your cable system might not have as many channels as it's capable of. Networks run ads so they can get more money, and it helps that they let cable companies sell some of the ad slots locally.
Cable companies even pay local broadcasters. That's why there are so many crummy little UHF religious and shopping stations--if their signal is above a certain minimum, the local cable monopoly is required to carry their content, which is guaranteed income for them.
Slashdot Mirror
IAC AYT is no good because the response arrives as NVT data, but either end could use option negotiation for liveness (pick some harmless option, send IAC WILL FOO and expect IAC DO/DONT FOO).
HTTP with WebDAV offers type negotiation, transparent compression, connection reuse (hardly any FTP clients can use MODE B or MODE C), and cache coherence. Unless you're aware of something FTP still does better, it really is obsolete.
HTTP proxies and NATs tend to time out idle TCP connections. As usual, without micropayments, there's no way to give all the providers between incentive to offer higher quality of service.
But responses have to be one-to-one with requests. A While clients can send overlapped requests (but a proxy might not forward them overlapped!), servers MUST send one complete response to each request in order--no prioritization or unsolicited messages are possible, and since "server push" is implemented as one big multipart response, the client can give no feedback about it other than unexpectedly closing the connection.
IETF recommends that browsers make disabling cookie support possible for privacy (for the same reason I don't walk around with my True Name tattooed on my forehead). Hidden fields are more reliable and don't represent the same threat.
The World-Wide Web Consortium doesn't have any particular influence over the Motion Picture Experts Group, and video compression is such a minefield of patents that a free standard is probably impossible (though I wish luck to the Ogg Tarkin team).
There's nothing wrong with making money. But since the World-Wide Web is about universal access and platform independence, if your business model involves controlling who may use a technology and how, it has no business being recommended by W3C.
See the intrinsic events module for onchange. Remember that XML is case-sensitive, and that user agents may not implement any particular scripting language or have scripting enabled (due to a history of security and privacy exploits in the standard sub-standard browsers).
The DMCA makes circumventing "effective" copy prevention a "crime" of its own--it doesn't matter whether the underlying use is infringement or not.
"New keywords" is an odd complaint about a language that declares abstract methods using "= 0", gives at least three different meanings (non-instance member, no external linkage, not reinitialized at each funcall) to the keyword "static", and resorts to a dummy argument to distinguish prefix from postfix operators.
Most of the shows the article mentioned don't even require basic cable--they're regularly broadcast to everyone within range.
Or maybe he's a character on the show.
Did GCC actually produce incorrect code? I seem to recall reading that the KDE or Qt maintainers made brain-damaged assumptions ("the first virtual base class is stored at offset zero in any derived class" or something similar) and got upset when GCC 3 started giving them undefined behavior that was different than the undefined behavior they were expecting.
Don't forget, they ought to take responsibility not only for determining whether the current limit is short enough to be Constitutional but also whether adding twenty years to the limit every twenty years is still "limited" enough to be Constitutional.
Tabs? I don't get it. I already have a perfectly good window manager. Why do I need another one (with undocumented keystrokes, no less) that only applies to one app?
The spec is over 400 pages, and I've been meaning to read it one of these days, but it hasn't been a real priority for me yet--can you be more specific? There are certainly huge differences, but do any of them represent new capabilities (as opposed to "negative variety" due to not-invented-here syndrome)? Assemblies seem to be a workaround for vendors too brain-damaged to rename libraries when they violate old interface contracts, and unsafe code completely subverts the security model. What have they done that's actually an improvement?
Without Microsoft's proprietary classes, .NET offers nothing you couldn't already do with JDK1.1 for years (unless you want to embed whitespace and comments in your RPC messages).
Exchange is allegedly moving away from their own protocol towards WebDAV (mostly so that other apps can easily store documents in "public folders") but I agree it's unlikely Outlook will be allowed to interop well with a competing mail server.
Does Sun make a lot of money on Solaris licenses? I thought they were pretty much staying in the hardware business.
Seems to me you're blaming the firefighters for putting effort into hosing down your house during the emergency but failing to put equal effort into drying your belongings off afterwards. Their real job is to save the neighborhood, not any one house.
The surrogate codepoints (0xD800-0xDFFF) have long been reserved so that UTF-16 can encode codepoints out to U+10FFFF. Codepoints beyond U+FFFF haven't been assigned yet, but a few scripts using them have already been proposed for Unicode 3.2. It's almost certainly going to happen, and if java.lang.String still lets you chop a character in half then it's seriously broken.
I want users to have source. I want them to understand their software and make it do what they want. Even better, I want them to hire me to do it for them.
If some proprietary vendor wants to control their customers and eliminate work-for-hire (the ethical business model for software, just like other professional services) I refuse to subsidize them.
That's what a license is--their agreement to let you do something they would otherwise have the power to prevent.
It was legal until very recently. Courts are just starting to rule "equitable estoppel" (basically, you can't do it because it's not fair) against patent holders who wait until an invention becomes an industry standard before pointing out their patent covering it.
If my patent is ruled to be invalid, that means it always was. Can you sue to recover the fraudulent royalties I exacted from you with my invalid patent? Aren't royalties usually put in escrow until the court case and appeals are resolved?
Cable companies do pay networks for the right to use their content, just like local broadcaster affiliates pay their networks. That's why your cable system might not have as many channels as it's capable of. Networks run ads so they can get more money, and it helps that they let cable companies sell some of the ad slots locally.
Cable companies even pay local broadcasters. That's why there are so many crummy little UHF religious and shopping stations--if their signal is above a certain minimum, the local cable monopoly is required to carry their content, which is guaranteed income for them.