Slashdot videos: Now with more Slashdot!
We've improved Slashdot's video section; now you can view our video interviews, product close-ups and site visits with all the usual Slashdot options to comment, share, etc. No more walled garden! It's a work in progress -- we hope you'll check it out (Learn more about the recent updates).
That it could put files wherever it wanted
They could have implemented VirtualStore as early as Windows 95 as a stop-gap measure for write-anywhere programs. Sure, it's an approach with its own problems, but sometimes you have to trade something in for security.
low-level hardware access for sound, graphics, etc
Trap the hardware interrupts in software, then emulate the low-level I/O routines at the OS level. Possibly with a performance penalty, but again: you have to decide where your priorities are.
And yes, I know hindsight is 20/20. Maybe not all these things were obvious back then. I still think that, security-wise, Microsoft spent the whole 90s and a good part of the 00s asleep at the wheel.
I actually agree with you in that the first Unix implementations had a number of security holes, and that the so-called iron-clad Unix security only came many years later with the accumulated experience of dealing with those holes. But I take trouble with the following claim:
to this date, *nix does not support well the concept of application ownership of a file which leads to programs requiring their own user account, which is another kludge.
Would you care to explain what is kludgey in using the uid namespace to also provide per-application ownership? Arguably, it is simply a matter of implementation simplicity; you have a single namespace instead of two. That a given uid might not correspond to an actual, physical user seems to be more of a semantic problem than a design one.
I'm sorry, but getting paid for your work, in a world where money is necessary to survive, is NOT morally wrong.
I'm no hardline stallmanite, but what does this have to do with free software? Nothing in the definition of free software precludes you from making money through it. At most, it forbids certain ways of making money with it.
With "openness" the market will do the job just fine.
The question is whether openness is possible (or rather: assured) in a free market without regulation.
Actually, the problem is that "free market", as currently (ab)used, is an overloaded term that stands for two completely different concepts: absence of regulation, and perfect competition with no barriers to entry nor asymmetric information. That the second is a necessary consequence of the first is a highly contentious point, because, contrary to what libertarians believe, the state is not the only entity that is able to distort a market.
My hypothesis is this: the more high-level information you can give your compiler, the better it can optimize your programs.
It is worthwhile to point out that GCC actually adds some non-standard facilities to the C language that enable the programmer to do so; witness, for instance, the use of the likely() and unlikely() macros (the actual GCC builtins have different names) in the Linux kernel to nudge the compiler into optimizing for the common case of an if statement.
Even if your argument held water (which I don't think it does in a properly managed network), it seems rather silly to trade global end-to-end connectivity and other IPv6 niceties such as autoconfiguration for the convenience of being able to memorize network addresses or pass them over the phone.
The grandparent is not talking about static addresses, he/she is talking about public addresses which are not the same thing.
Except that in this case you are creating a market to deal with artificial scarcity, which is just plain evil. Remember, there is no physical reason for there being only 2^32 addresses. Only the completely accidental fact that someone picked the number 32 some decades ago.
And how is this any different from IPv4, where the residence gets a public address from the ISP? Unless your ISP is using carrier-grade NAT, that is.
I believe the reason the host part has so many bits does not have to do with expected subnet size, but rather serves to ensure the probability that two autoconfiguring hosts coincidentally choose the same address is kept low.
Try reading what I said again.
I never said that the embassy guards do not have the authority to retaliate. I'm just contesting the notion that the actions of independent citizens translate as a formal declaration by their nation state. The United States, of course, are free to interpret such an attack as an act of war and declare war on Libya; but to say that Libya have declared war on the United States when the embassy was attacked flies in the face of international laws and conventions.
citizens of Libya technically declared war on the US.
Citizens do not declare war. States do.
Are you suggesting that, if a bunch of Americans vandalized a foreign embassy on US soil, that shold technically count as a declaration of war on that country by the United States of America?