Does your SSL client actually verify the certificate? I've seen a number of clients that support "SSL", but are trivially vulnerable to man-in-the-middle attacks.
Lately, I've started to wonder if all those people who say "being gay/lesbian is a choice" (without any real science to back it up) say it because they're really bisexual, and thus see the possibility of choosing one's sexual preference.
Like you said, I am quite convinced that I couldn't choose to be gay (MAYBE I could pretend to be gay for a little while, but that still wouldn't make me BE gay), and, like you said, I suspect that there are a lot of gay people who couldn't choose to be straight.
I'm not sure why it's even an issue, really. One more gay man just means that I have fewer people competing against me for the women!;-)
BusyBox probably is GNU-based (I'll take your word for it) so it's not the best example, but the argument is still valid. IIRC, there is an OS project somewhere that uses Linux as its kernel, but uses a different userland. I tried to find it so I could link to it, but I couldn't remember what it was called.
I suppose this is a better set of examples for what this construction means and how it could be useful: "Win32/9x", "Win32/NT", "POSIX/NT", "OS2/NT", "Java/Win32",...
The notation is also handy for things like "Debian GNU/FreeBSD", which is the Debian GNU userland running on the FreeBSD kernel.
So my argument is that although RMS tossed around the idea of using "GNU/Linux" to give credit to the GNU project (which is not really such a bad idea anyway, IMHO) there at least one other practical use for that construction: the "GNU/Linux" construction can be used to distinguish different Linux-based "operating systems" (in the broad sense) from each other.
Firefox will allow popups that occur right after you click something. The pop-up craziness didn't happen until I clicked in the web page. Either that, or Firefox doesn't block the Flash plugin's attempts to open new windows.
We still have a long way to go before we have *real* web security.
Really? I would bet a large sum of money that the number of people that give a shit about HURD after it's ported to L4 will be largely equivalent to the number of people that give a shit about it now.
Sorry, I meant "only if/when", i.e. being ported to L4 is a necessary, but not sufficient, condition for HURD ever gaining any importance.
One more thing - you seem to think the fact that the kernel could fit in the CPU's cache is important (presumably to performance).
I think you misunderstand me. The major complaint people seem to have about microkernel architectures is that they have poor cache performance resulting from large amounts of context-switching between programs and the microkernel. The nice thing about L4 is that because it's small enough to fit in your CPU's cache, so you get much better cache performance.
At least, that's what I'm told. I'm getting this information second-hand from a friend of mine who knows a lot more about the details than I do.
Part of being a professional involves understanding that language is just a tool that people use to help them communicate; A requirements specification is a form of imperfect communication between you and your customer.
If part of the spec is ambiguous, it's up to you to either contact your customer and get the ambiguities resolved, or if that's not feasible, make a best-effort guess at what the customer intended to put in the requirements document.
While I agree that, in your case, this was partly your teacher's fault, your real-world customer would probably have been just as unhappy -- and just as ready to blame you, regardless of the facts -- as your teacher was. Even if you dig your heels, point at a copy of the contract, and insist that you've met all the requirements, you still risk losing at least one customer.
Your teacher, unwittingly or not, gave you a very real example of how software development works in the real world. Whether or not you learn from that is your choice.
Debian Policy (and Policy is God) is for every binary package to be built with everything enabled,
As a Debian user, I'm not so sure that you are correct. My experience is that this is *not* what happens in Debian. For an example that's just off the top of my head:
$ apt-cache search abiword | grep ^abiword abiword-gtk - WYSIWYG word processor based on GTK abiword - WYSIWYG word processor based on GTK2 abiword-common - WYSIWYG word processor based on GTK2 abiword-doc - Documentation for AbiWord abiword-gnome - WYSIWYG word processor based on GTK2/GNOME2 abiword-help - Online help for AbiWord abiword-plugins - Plugins for AbiWord abiword-plugins-gnome - Plugins for AbiWord (with GNOME dependency)
What section in Debian Policy are you referring to?
HURD will be important if/when it gets ported to L4. Mach is too fat and too slow. L4, on the other hand, is so small it can fit inside your CPU's cache.
Despite what RMS may or may not say, there is at least one good technical reason to use the term "GNU/Linux" over just "Linux": disambiguation. Not every Linux distribution is GNU-based, particularly ones that run in small embedded environments or installation floppies.
If you say "GNU/Linux", you can make certain assumptions:
Your libc is GNU libc
#!/bin/bash will work
Stuff like "ls/bin -l" and "tar xvjf..." will work, because you're using GNU coreutils.
Your C compiler supports GNU extensions, because you're using the GNU compiler collection
None of those assumptions can be made when you are talking about just "Linux".
A similar line of thinking leads people to use the term "TCP/IP" instead of just "Internet Protocol".
Probably the worst thing you can do is use a computer that you care about. It's absolutely critical that the child be allowed to experiment and try new things without worrying that he/she might break things irreparably.
Older computers that had only tapes/floppies were better in that way, since it was pretty hard to ruin media that was either in the drive with write-protect enabled, or in the desk drawer.
You probably also want to have programs (read: games) available that can be changed easily.
I haven't tried Macromedia Flash, but I'd look into it.
Loki closed almost three years ago, and the events that led to its demise occurred over a course of 2-3 years before that. A lot has changed in (approx) 5 years.
Re:Exactly the problem that a lot of people have
on
Does Linux Have Game?
·
· Score: 0
There probably *is* an efficient way to implement DirectX in Linux, but I don't think enough programmers care enough, are skilled enough, and are willing to face possible lawsuits (frivolous or otherwise) from MS.
You may copy and distribute the Program (or a work based on it, under Section 2) in object code or executable form under the terms of Sections 1 and 2 above provided that you also...
if I submit a form with informatin to https://blah.com/secret.cgi?this=password;that=por no I can still see where that trafic went because the ssl transaction hasn't started yet.
I think you're thinking of the subject line of encrypted email messages. In HTTPS, SSL negotiation happens as soon as the TCP connection is established, i.e. before requests are made.
Slashdot Mirror
Right?
IIRC tanks and other heavy artillery weren't around during those wars.
The binary CA-trust model used by many SSL implementations is horribly vulnerable to that kind of thing.
Does your SSL client actually verify the certificate? I've seen a number of clients that support "SSL", but are trivially vulnerable to man-in-the-middle attacks.
"Troll? Hey, that's me! Ok."
Like you said, I am quite convinced that I couldn't choose to be gay (MAYBE I could pretend to be gay for a little while, but that still wouldn't make me BE gay), and, like you said, I suspect that there are a lot of gay people who couldn't choose to be straight.
I'm not sure why it's even an issue, really. One more gay man just means that I have fewer people competing against me for the women! ;-)
I suppose this is a better set of examples for what this construction means and how it could be useful: "Win32/9x", "Win32/NT", "POSIX/NT", "OS2/NT", "Java/Win32", ...
The notation is also handy for things like "Debian GNU/FreeBSD", which is the Debian GNU userland running on the FreeBSD kernel.
So my argument is that although RMS tossed around the idea of using "GNU/Linux" to give credit to the GNU project (which is not really such a bad idea anyway, IMHO) there at least one other practical use for that construction: the "GNU/Linux" construction can be used to distinguish different Linux-based "operating systems" (in the broad sense) from each other.
We still have a long way to go before we have *real* web security.
Sorry, I meant "only if/when", i.e. being ported to L4 is a necessary, but not sufficient, condition for HURD ever gaining any importance.
One more thing - you seem to think the fact that the kernel could fit in the CPU's cache is important (presumably to performance).
I think you misunderstand me. The major complaint people seem to have about microkernel architectures is that they have poor cache performance resulting from large amounts of context-switching between programs and the microkernel. The nice thing about L4 is that because it's small enough to fit in your CPU's cache, so you get much better cache performance.
At least, that's what I'm told. I'm getting this information second-hand from a friend of mine who knows a lot more about the details than I do.
This is because you are an idiot.
IHBT, HAND.
If part of the spec is ambiguous, it's up to you to either contact your customer and get the ambiguities resolved, or if that's not feasible, make a best-effort guess at what the customer intended to put in the requirements document.
While I agree that, in your case, this was partly your teacher's fault, your real-world customer would probably have been just as unhappy -- and just as ready to blame you, regardless of the facts -- as your teacher was. Even if you dig your heels, point at a copy of the contract, and insist that you've met all the requirements, you still risk losing at least one customer.
Your teacher, unwittingly or not, gave you a very real example of how software development works in the real world. Whether or not you learn from that is your choice.
There's also the apt-src package, which helps you maintain slightly-patched versions of Debian packages.
As a Debian user, I'm not so sure that you are correct. My experience is that this is *not* what happens in Debian. For an example that's just off the top of my head:
What section in Debian Policy are you referring to?
Of course, the package will be upgraded next time you do an upgrade, so to stop that, you can "hold" the package:
To undo that, you'd run:
Of course, fullscreen package management utilities like aptitude and Synaptic let you do the above with fewer keystrokes.
Alternatively, you can put something like the following into /etc/apt/preferences (see man apt_preferences for more details):
Hope that helps.
HURD will be important if/when it gets ported to L4. Mach is too fat and too slow. L4, on the other hand, is so small it can fit inside your CPU's cache.
You made me kill my Firefox process! Bad troll!
If you say "GNU/Linux", you can make certain assumptions:
None of those assumptions can be made when you are talking about just "Linux".
A similar line of thinking leads people to use the term "TCP/IP" instead of just "Internet Protocol".
Oh yeah! Clearly AMD's engineers are tied up doing this investigation...
</sarcasm>
No text
Older computers that had only tapes/floppies were better in that way, since it was pretty hard to ruin media that was either in the drive with write-protect enabled, or in the desk drawer.
You probably also want to have programs (read: games) available that can be changed easily.
I haven't tried Macromedia Flash, but I'd look into it.
Heh. Yeah, take all your critical infrastructure and place it along a semi-hostile border. Brilliant! ;-)
Uhm.... Your link doesn't contain any of the words: "tank", "paint", "color" (or "colour"), "designation", or "NATO".
Loki closed almost three years ago, and the events that led to its demise occurred over a course of 2-3 years before that. A lot has changed in (approx) 5 years.
There probably *is* an efficient way to implement DirectX in Linux, but I don't think enough programmers care enough, are skilled enough, and are willing to face possible lawsuits (frivolous or otherwise) from MS.
Actually, you do. See section 3 of the GNU General Public License:
I think you're thinking of the subject line of encrypted email messages. In HTTPS, SSL negotiation happens as soon as the TCP connection is established, i.e. before requests are made.