BSD, Linux and MacOS allow you to do that, and even delete or overwrite the file while it's still locked without causing problems. Moving, deleting or renaming a file affects only a hardlink to the file and not the file itself; and overwriting a file is actually just deleting a hardlink and writing to a completely new file.

Give the user his own private key and honour it properly. That's all that is required. You don't need to disclose vendor private keys.

That's a fair point. I know nothing about Swedish civil law, but if they feel that their trademark should preclude Mojang's then I would imagine the appropriate thing to do would be to lodge an objection directly with the trademark office. If Bethesda's claim has merit then the application would be denied. Applying for trademark before using the mark is a prudent, diligent, and above-board way of doing things, and making Mojang withdraw their application or be forced to defend themselves in court strikes me as an denying Mojang access to due process. I'm not a lawyer and know very little about these matters, but Bethesda's course of action so far does not sit well with me. Still, I have a lot of respect for both parties and hope that they can work this out without bloodshed (unless it's in Quake, of course).

They can't lose something that they have never had. Bethesda do not have "Scrolls" as a trademark for a computer game, and do not stand to lose "The Elder Scrolls" by doing nothing here. Furthermore, all Notch has done is apply for trademark on a name that he plans to use. There's nothing wrong with that - it's just diligence - and filing a suit about that is rather arbitrary. If Bethesda's representatives have an issue with that then they should just contest the application.

Pardon the spelling - I meant to say, "rarefied", the opposite of "condensed".

Ah - as in "rarified". That makes sense. Thanks for the explanation.

Thank you for that. I thought the summary seemed bizarre - since when do we have laws against jailbreaking, and doesn't SAFACT normally focus on stopping sale of unauthorised duplicates? It's nice to see that things are despite Slashdot headlines things are actually working the way they should.

No, it is a purely practical matter. It's the difference between code working reliably and not.

NULL is a macro who's definition is implementation-defined. It is most commonly defined as 0 or 0L, but I have also seen (void *)0 and even (char *)0. I just looked at string.h from gcc 4.3.4 and it is defined as 0. You're looking at this and thinking (void *)0 is reinterpreting the number 0 as a pointer. It is not, and none of the languages that you mention allow you to convert integers into pointers. 0L, in this context, is implicitly a pointer which may or may not have a numeric value of 0 (look at DOS - it had at least different kinds of pointers with different sizes, some of which were tuples and therefore didn't evaluate to single numeric values at all).

It changes constantly, but it's more of an implementation thing than a platform thing. (Of course, the platform plays a big role in the implementer's decision.)

Those are the very systems which tend to define null as something other than a pointer to memory at address 0. C, C++ and Objective-C do not allow you to assign integer values to pointers anyway - literal zero integers implicitly evaluate to null pointers (whatever those are) which can be anything. Let me give you an example.

int zero = 0;
int null_is_zero = (0 == *(void **)(&zero)); /* nonzero if and only if null is defined as zero. Might break if size of int and void* are different. */
char *pointer = 0; /* pointer is null, which is who-knows-where */
pointer = zero; /* error. zero has a value of 0, but there is no implicit integer-to-pointer conversion. */
pointer = 1; /* error - you can't assign integers to pointers
pointer = *(char **)(&zero); /* dangerous because sizeof(char*) might be different to sizeof(int). pointer now points to memory with address 0, and may or may not be a null pointer */

If a and b are rational numbers, then is ab rational? That depends on the closure of rational numbers, which I would argue is not part of the premise at all. You are extrapolating from assumed properties of the word "rational" which are not given in the statement.

Consider this: "if a and b are imaginary numbers, then ab is imaginary". The same syntax, but not even true. What if we define "rational numbers" to be imaginary? Then the original example isn't true either.

A variation of the original example that is tautological would be "if a and b being rational numbers is sufficient for ab to be rational, and a and b are rational numbers, then ab is rational".

Be careful about thinking of null as 0. Null pointers and zero pointers are different concepts that just happen to coincide sometimes. Although it's usually the convention on x86 that zero pointers are considered null, it's not strictly the case in general. Quite often there's a completely different convention, like 0 being a valid pointer and -1 being null. Part of the confusion comes from the fact that C uses the '0' character as a symbol for null to avoid having an extra keyword. The C code "char *pointer = 0;" doesn't necessarily give the pointer a zero value; it assigns whatever value the compiler is configured to treat as null.

The person to whom you responded is accurate, as far as I know, but "it's stored in /data which is off limits to any app, unless you've rooted the phone." still does not imply that rooting the phone causes /data to cease being off-limits to apps. You're drawing a false conclusion.

No such weakness? What weakness are you talking about? GPP did not imply that rooting the phone causes /data to cease being off-limits to apps.

False. Counterexample: rules against anticompetitive practices make it easier to compete - that's the reason such regulations exist, after all.

Name-based hosting makes it a little more complicated, since (as far as I am aware - please correct me if I'm wrong) URL's don't currently have a way of specifying the domain and the IP address at the same time. You can get a link to the correct host, but because you used an IP address instead of a domain name the server will present the wrong website.

The headline is very misleading. The big news here is that there is a Gallium3D back-end that gives tolerable performance for modern AMD cards. Gallium3D is a still-immature hardware abstraction architecture which could do amazing things for the industry if it manages to get over its chicken-and-egg problem.