Is "Linux x86 assembly" any different to
any other kind of "x86 assembly"?
Yes. Although it requires understanding the
CPU's native capabilities to the same degree,
Linux uses AT&T syntax, whereas most of the
Wintel world uses (unsurprisingly) Intel/Microsoft
syntax.
Personally, although I far prefer coding C under
Linux, I prefer Intel syntax assembly. Even
with many years of coding experience, I find
AT&T syntax unneccessarily convoluted and
somewhat difficult to quickly read through.
The larger idea holds, however, regardless of
what assembler you use. I wholeheartedly agree
with the FP - People who know assembly produce
better code by almost any measurement except
"object-oriented-ness", which assembly makes
difficult to an extreme. On that same note,
I consider that as one of the better arguments
against OO code - It simply does not
map well to real-world CPUs, thus introducing
inefficiencies in the translation to something
the CPU does handle natively.
The whole point behind port knocking is
the wrong impression that "open" ports are
more insecure than "closed" ports. This is
totally bogus.
Tell me, then, why we all rush to patch our
SSH installations every time a new weakness
comes out?
I agree, in theory, a visibly open port does
not pose a security risk. Something,
however, has to listen on that port, and
programmers do make mistakes.
Adding something like port knocking (see my
previous post on this topic for my idea
on a more useful generalization of port knocking),
a potential attacker needs to find a weakness in
both the knock sequence and the
SSH (or whatever) server.
Kernel code is not intrinsically more
secure than application level code!.
I'll agree with you wholeheartedly, on that
point. In fact, I'd call kernel code potentially
less secure, since, while bad user-space
code can cause a few problems, bad kernel code
can easily compromise an entire system. And,
while I'd like think that kernel hackers take
extra care for that exact reason, I'll
reiterate - Programmers do make mistakes.
but you don't need port knocking for this.
Just stuff the key in the SYN packet. Don't
accept if it doesn't have the magic word.
Putting the key in a SYN packet, you can snoop.
knocking on a seemingly random set of ports,
with some care in the implementation (such as
using the current time as a component to the
correct sequence), you cannot.
I agree, a "good" implementation of this would
require, at the very least, a firewall module...
Though not necessarily poking holes in it, since
none of those holes actually let traffic through.
Personally, I'd like to see this idea extended
to a general "machine" password, totally separate
from any security associated with a particular
server bound to a port - That way, you don't need
to modify a million and one individual programs
that already speak TCP/IP (both clients
and servers), you just need a single extra
program (a doorbell, if you will). Without
ringing the doorbell, you simply can't connect
to anything listening on that machine.
I agree, though, that this goes beyond what the
link in the FP had in mind. But take that
original idea just a bit further, and you have
something that sounds rather useful.
This is equivalent to to putting a password
on access to the port.
This seems much better than a password,
I would think (Though I certainly would
still use a password as well).
As an analogy, if you want to get into a
house, and find a locked door, you have a
few options... You can try one of those
M x N position key blanks, which will take
a very very long time (exhaustive search).
You can try to pick it (exploit a weakness
in the password algorithm). You can try to
get ahold of a copy of the real key (packet
sniffing, "shoulder surfing", etc). But you
have no doubt that somewhere, a key exists
that will open that door.
Now compare that to a solid block of concrete,
roughly the size of a house. What does it do?
Do helicopters land on it? Does it cover
something, or hold something down? Does
it have something sealed inside it? You'd
never suspect that that, if you utter
the magic phrase "Sim sala bim bamba sala do
saladim", a door will appear in the side of
this large concrete block, allowing those
with a key to gain entrance.
The main difference involves knowing whether or
not a way in exists. With just a passworded
port, an attacker knows that enough effort will
pay off. Adding in port knocking, that attacker
doesn't know whether or not their hard work can
ever gain them entrance, since a port
might well not exist.
Now, in my opinion, the more interesting question
here involves how to hide this from one's ISP
(ie, make it snoop-proof).
Heh... That almost comes across as though
you wrote it with a straight face...
Good one. Heh... Pay a "use tax" on something we
manage to buy without a second (or third)
raping by the IRS (and the individual states',
and counties', and cities' version thereof)?
Hilarious.
It amazes me that people put up with the idea
of sales tax at all. The US constitution
has these funny ideas about "taxation without
representation". They exist for a reason, and
the entire concept of "sales tax" violates that,
big-time.
Anyone but a complete moron either buys online
(until Kansas and the like ruin that for
us), or at the very least saves up "big"
purchases for a trip to a nearby sales-tax
free state.
If we run out of non-mac mods your post will
tank. I've had +5 go down to 0 after a few days
with mac moderators still attacking my posts
moderations.
Same here... I tend to not really care about
my/. karma, since for the most part people seem
to consider my posts decent, and I've kept
"excellent" karma for a couple of years now.
However, I tread rather lightly when it comes to
anything Apple-related... Even guarded criticism
of glaringly obvious facts result in a post
getting moderated into oblivion.
Sadly, I have yet to see such blatant abuse of
the/. mod system when I metamod (in fact, I
don't think I've ever seen anything that
I would consider "unfair" while metamoding...
At worst, a few that I didn't really find all
that funny, but not bad enough to rate as
distinctly "unfunny"). Personally, I think/. needs to shift the focus of metamoderation
to detect abuses of down-mods, rather than
up-mods. I mean, except for glaringly obvious
"I posted from my other account and modded myself
up", people tend not to abuse positive moderation.
But, so it goes. Except for certain topics
(such as anything Apple), the moderation system
seems to work tolerably.
Above some "decent" number, I'd tend to agree
with you. Standard IA32 does not, however,
have anywhere near enough.
"Enough" means not needing to use registers
with clear designated purposes (such as ESI,
EDI, and yes, even EBP) just to keep a triply
nested loop (that actually does something)
from hitting memory.
"Enough" means not needing to remember if CX
gets clobbered on a given shift-like operation,
so we need to preserve it temporarily.
"Enough" means having the ability, on performing
a set of similar operations on a vector (not talking
about using SSE or the like here), to interleave
enough groups of operations to prevent a register
dependancy stall...
Enough already. <G>. "Enough" means "more
than 8" (of which you can only ever use at most
seven, and even that with some difficulty).
Most consumers are going to sitck with 32-bits
for some time to come since it currently offers a
cost advantage and there is no real advantage for
them to move to 64-bit.
I really don't understand why so many people have
decided to speak poorly of 64b CPUs... Sure, not
all that many programs will benefit from them right
at the moment. But some will. And if we
really want to take the "we don't need more bits",
why not go back to 16b? or even 8b? Hell, let's
just use bitslicers - We could probably clock one
of those puppies well into the hundreds of gigahertz,
due to the design simplicity.
The mathematical hobbyists (such as myself) will
switch first, for the obvious number-crunching
advantages. I already look forward to getting
my hands on an Athlon64, to port a few of my
projects to. Even better than the machine
word length, though, I look forward to having
a decent number of registers. Yum.
Then the gamers, as soon as they realize that
64b->64b->256b (CPU to AGP to GPU) performs
quite a lot better than 32b->64b->256b...
Next will come either another boon to the gamers,
or to the artists, depending on whether "killer
game 2005" or "Photoshop 2005" comes out first
with x86-64 optimizations.
Last will come the casual users, but they will only
buy in when they have no choice (really, although 8b
doesn't suffice, 16b works just fine for 99% of
what a PC does.. So why do they even use 32b?).
Overall, I agree with you - you don't need 64b to
browse the web and email grandma. For anything that
either pushes the CPU to its limits, or improves
linearly with the addition of more RAM, 64b has an
obvious advantage.
As an aside, I already mentioned the increased
number of registers in AMD's 64b offerings... While
a lot of people have debated this issue only on the
merits of "why do I need 64b?", the x86 line has
always had a SEVERE shortage of GPRs. Doubling
that number will (not just "could") boost
performance immensely.
Fine, here we go, embarking on an exercise
in futility - educating the ignorant trolls
of the world. Twit.
For paging to disk, which would you say
takes more time: "hit the disk", or
"nothing"? Even if only verifying the
pagefile's integrity at boot (not currently
the case on any OS I know of,
though feel free to name a few, of
which I can save you the trouble and tell
you that you won't find either "Linux", "Windows",
or "OS-X" as an example), that still comes out to
T>0 vs T=0.
For a ramdisk... Wow, where to start?. Okay...
Do you suppose loading file X from RAM, or from
a set of glass disks with mechanical actuators
moving across them, takes longer? Keep in mind
those moving parts read bits off the glass
plates, perform several layers of verification
and error correction, place them in local RAM
on the drive itself, and only then does the
requested data make it to the CPU. So again,
factoring out the "RAM" step common to both,
this reduces to "something" vs "nothing".
Maybe you shouldn't be making a fool of
yourself then.
I have enough confidence in my statements
to post as myself, rather than AC. I occasionally
make mistakes, and people call me on it, but really,
I can't even imagine what you had in mind when you
decided to post your comments...
However, feel free to vindicate yourself by
proving the counterpoint to my assertions? I'd
love to hear a bit about the world you
live in, where a disk performs faster than
RAM...
Oh my. PC's catch up to CP/M circa '81 and
AmigaDOS circa '86. Put your C compiler, sources
and vi in ramdisk and things were just nice,
even on a floppy based system
Heh, I agree this seems like a throwback to an
older generation of computer hardware, but one
that works rather well.
Personally, I don't understand why we still have
virtual memory - Another throwback to the days
of now-obsolete hardware, when memory cost an
arm and a leg yet programs started wanting tons
of it.
The difference between the two, now that we
have oodles of cheap RAM, involves performance.
Disk paging (even totally idle) hurts
performance, while a ramdisk improves it. Yet
all modern Windows version force you to use the
former (though you can technically disable it
in XP, some programs simply will not work if
you do so, regardless of how much physical RAM
you have), and don't even include native support
for the latter.
How do you do ramdisk in W2K/XP (I'm
too lazy to Google it)?
It doesn't come with Windows, but a number
of 3rd-party programs (mostly all based off
a reference MS driver, ironically enough)
exist to do it for you.
Personally, I like QSoft's, which installs
easily and works well. The free version
theoretically has a 64MB limit, but I've
had no problem telling it to use half-a-gig
(and the pulldown menu goes to 1GB, but I
don't have a machine with enough RAM to try
that one).
And, as I mentioned, just 64MB makes a HUGE
difference in responsiveness under XP. Still
helps on 2k, but not nearly as noticeable.
Yup... And while an Itanium 2 currently goes for
$1500, the Opteron 248 comes in at under $1000,
with Athlon64 3400s going for $400.
Oh - and it makes Opteron look like my grandma
when it comes to the real domain of 64-bit
computing - science and engineering.
In theory, the Itanium-2 beats the Opteron, even at
a much lower clockrate. In practice, the benchmarks
I've seen (mostly just summaries of material from Spec)
show at most a 50% difference on comparable hardware.
And that, only on native 64b code. For the desktop
market, where 32b will still exist as a significant
portion of the available programs for at least another
few years, the Opteron and Athlon64 suffer *no*
performance hit on 32b code, while the Itanium line
slows to a crawl.
Overall, I'd agree that the Itanium fills a particular
niche fairly well. With an Intel-optimized Fortran
compiler, I doubt anything else can even come close,
clock-for-clock. For general-purpose computing, however,
AMD has made tradeoffs that I strongly believe will
pay off, and not just over the long-term. For desktop
use, this seems like a no-brainer - Go with AMD.
I can't see a need for it on the desktop just
yet.
One word - Ramdisk.
Set yourself up with a nice 1GB ramdisk,
and make everything with a user-settable
temp and/or cache directory point to it
(most media apps, MSIE, Mozilla/Netscape).
Also point your "temp" and "tmp" environment
variables to it.
Personally, I still use 2K and consider XP a
total dog. Circumstances recently having forced
me to use it quite a bit, however, I found that
it hits the disk constantly. Setting just
a 64MB ramdisk (on a machine with only 256MB)
literally doubled its responsiveness. And
theoretically, this should increase the life,
and decrease the fragmentation, of your HDD.
The one down side, some programs will complain
if they can't fit their data on the temp dir
(so for example, in Netscape, which for some
unknown reason stores your downloads there
before moving them to their real destination,
downloading a 70MB file with only a 64MB temp
ramdisk will make it complain). At a GB or two,
though, I expect that would very rarely
cause any problems.
Another reason, that the recent so-called "benchmark"
failed to consider - We've had 128b and even 256b GPUs
for a while now. AGP (and PCI64, currently rare but 64
bit CPUs should push it into the mainstream) will benefit
GREATLY from native 64b drivers. Perhaps we'll finally
see Doom 3?
Besides, if the PC can also serve as a
PVR or other cool functions, then it's a
doubly-good deal.
I think that point matters the most, to
those of us who can roll our own...
Let's consider price-vs-functionality...
A Tivo or ReplayTV will set you back around $200,
not counting a few hundred more for their
subscription service. So let's say a nice
round $500 for something that will remain useful
for the next five years. A mid-range TV, such
as my household recently purchased (30" HDTV
16:9), $1000. A decent DVD player, $200. Add
in a 6.1 surround decoder, and
tolerable-but-low-end speakers, and you have
another $500. Total of around $2200, and it
only has TV, DVDs, and PVR capabilities (plus
basic audio functionality, such as radio and
CD playing).
For a PC... The TV you really still need, since
PC monitors cost far more per inch than a good
HDTV. However, you don't need the Tivo,
or the DVD player, or the 6.1 decoder (though
still the speakers). PC, $500. Pioneer A-06,
$130. SB Audigy 2, $120. Decent tuner card
(personal choice on these varies a lot, though
Hauppage seems the most common brand of choice),
$200. Total comes out to around $2100 - Already
ahead of the home theater setup.
The PC setup really excells in the
"features" category. First of all, it can
do everything the home theater setup does,
which obviously counts as the basic level
of functionality. It can also play any digital
media format you can imagine (ie, anything a PC
can play)... Some standalone DVD players now
support DivX, but how about XVid? Or AVI? Or
raw MPEG-2 streams? Or QT? Then we have content
that actually runs as a form of program, such as
Flash.
Additinally, a HTPC does work as a general
purpose PC. You won't want to read small text
for several hours a day on even a high-quality
TV, but to look something up on-line, or a quick
email check, they do rather nicely. Enjoy
classic console video games? Well, download an
emulator for your favorite system (you can
get them for anything before the current gen,
up to and including PS1, N64, and GBA), and they
look sweet, without needing to connect
27 dongles to your TV's coax-in, or ever blowing
the dust out of a cartridge again.
Finally, you can burn your stored media to DVD.
A set-top box to do that, even if the
PVR would let you, will set you back another
grand. Network support? A $20 NIC (assuming
your motherboard didn't include one) rather
than a $500 "appliance".
So the price comes out ahead (not even considering
"advanced" features that most people familiar with
PCs take for granted), and the feature set just
destroys a standard home theater system.
Considering this, we shouldn't ask "which costs
less and does more", but rather, "Why the hell
do home theater systems cost so much and do so
little?". Even if you do plan to blow $5k
to $15k on a kickass home theater setup, spend the
cash on a better TV and sound system, not obscenely
priced components, when a $1k PC will literally
"do it all" for you.
Also, anyone seriously wanting a home theater
setup should budget a bit for environmental
considerations as well - Spending $1000 on
light-control and sound suppression will
improve overall viewing enjoyment far more
than. A reflected glare on the screen, or a
noisy HVAC system that kicks on every half
hour, makes more of a difference in quality
than an extra 2 inches of screen or 4ch-vs-6.1
surround.
Professors don't care. In fact in some
cases they are paid to select the more
expensive of two options
Not true. Most professors do care,
and many that I had even took rather extreme
steps to minimize our book costs. My CS
professors used as many free web resources as
possible; my math profs tended to use, for example,
the 1968 classic text that you could get for a
pittance at any little buck-a-book shop; I even had
one psych professor that actually photocopied books
for us, handing out a packet containing the
relevant pages every few classes.
The one exception to this, where I agree with you
100%, involves having a professor who wrote their
own book. Those miserable bastards will gouge
you any way they can. I actually had one "invite"
me to drop his class because I insisted on using
the (only trivially changed) previous edition that
I picked up used on-line for $10.
Seriously... I realize new books cost a lot, but
the previous edition (especially if the
campus bookstore will no longer carry it) costs
a pittance, like $5-$10 if you can find it used
online. And you almost always can.
As for the questions changing (the only real
difference between most editions that I
ever noticed), go to the library and photocopy
those two pages per chapter. Adds an extra buck
to the cost of the book, but still comes out
quite a lot less than $80-$200 per book.
However, to address the deeper issues here - Well,
publishers have every right to charge what they want.
And with a captive audience, they just have to dream
of dollar-signs and they get it. At this point,
though, they've gone beyond anything even remotely
realistic, and I for one do not feel the least bit
of guilt for having used exactly the method I describe
above. I'd rather have had the correct books, but
on a college student's budget, the $300+ savings
per semester buys a LOT of instant ramen
noodles.
When I went to get my passport renewed,
people had to surrender their batteries.
Uhhh... Why? You walk in, get a picture taken,
hand them the forms, and leave. Nothing really
"secret" that you could take a picture of, even
if you wanted to.
Of course, ironically enough, they probably would
have let you take an actual camera in, and only
bother about phones to play follow-the-leader in
implementing paranoid policies about camera-phones.
Anyone know if the reason 15% of people falsely
think they have an HDTV television is because they
only have EDTV? Seems like a fair reason, since no
one has heard of EDTV.
I think that it has more to do with the
general lack of TVs with an actual HDTV tuner
in them. For example, my household recently purchased
a new widescreen HDTV, which probably would not count
as HDTV under that criterion. It does 480p, and
DVDs look as good as they physically can on it
(pixel-wise, I don't claim that a $15k system
wouldn't look a tad better). But I have absolutely
no use for an HDTV tuner, so why pay up to $200 more
for it?
If you have Movie Gallery stores in your area, they
let you rent overnight on Wednesdays for a buck.
This counts as a "normal" rental, so it still gives
you points toward free rental coupons. Their normal
bulk renting deals apply as well, so you can literally
rent five non-new releases for three bucks on Wednesdays,
and each month (I think it takes 12 points...),
you get a coupon for a freebie.
If you have Blockbusters (who doesn't?), and still use
a land-line phone, switch to the "MCI Neighborhood Complete"
package... It includes all the standard phone services (CID,
VM, 3-way, CW, etc), unlimited free LD, and even covers your
local service, for just over $50 per month. And, to make
it relevant to my point, you'll get a free rental coupon for
every $25 you spend (ie, 2 per month). In combination with
the BB Rewards card (actually pretty crappy, except that
they send you a coupon for one free rental per month, and
only costs $10 for a year), as well as the occasional special
deal ("Pre-order Movie X, and get 10 free rentals", where
even if Movie X sucks, the free rentals would have cost
more than the movie does), I literally never pay
to rent at Blockbuster.
Anyone know of any other good ways to rent for
a pittance at other stores?
Even a dual-layer DVD can be copied if
you're willing to accept some (usually
negligible) quality loss.
From my experience with backing up my own
collection, you don't even need to suffer
any quality loss... Most dual-layer
discs include the menu, the movie, and a
ton of extras.
The menu doesn't usually
take up much room (usually under a quarter
of a gigabyte), but the extras can easily
go to 2-3GB alone. Add to that a Spanish,
French, and three redundant English soundtracks
(DTS, 6ch, and 2ch occur often... You only need
to keep one of them, and AC3-6ch sounds pretty
damned close to DTS, at half the size).
So, if you remaster it as a movie-only disc,
removing all the extraneous soundtracks, you
can usually fit the movie with no recompression
on single 4.7GB disc.
Here's info on a demo Pioneer did at this
year's CES with a DVR-A06 and some tweaked
firmware.
Ah, that makes me SO happy I went with
the A06... Great drive all around, but this?
Wonderful! Now if they'll just release
that firmware (or someone leak it, I don't
really care about the legality of it).
Not with my positive-nitrogen-pressure DVD
player! Although piping the vent outside so I
didn't suffocate was a pain. And it's a bit of
a hassle lugging around all those nitrogen gas
containers.
Ah, see, you need to thing "renewable", rather
than keep getting more nitrogen. Just use a
sealed positive-pressure glove-box, with an
air-lock to add new discs, to store the player
and your collection. That way, a small tank of
nitrogen will last you a year, easy. You waste
too much in changing discs, see, if you just
have an open positive-pressure system.
And, no vent needed. Even in the even of a
complete breach of the seal, it won't contain
enough nitrogen (assuming you turn off the tank)
to asphyxiate you in any normal-sized room.
Y'know, actually, although I joke about this,
I did intend to set up a 10-gallon
aquarium to do something very much like what
I suggest above. Too bad this flopped before
we could abuse it to its fullest.:-(
Consumers know when they rent something that
it has to go back - they don't own it. They're
accustomed to this paradigm.
Perhaps true, but it doesn't change the fact
that people do forget to return movies on
time. Blockbuster, and presumeably all
major video stores, report that over half
of their profits come from late fees. To me,
that really says something (well, mostly, "don't
return movies late").
So, for someone who knows they will always return
their movies late, paying $7 up front may not
seem all that bad. Personally, I'd suggest they
just return the damn things on time, but hey,
just my opinion.:-)
It didn't support spanning. Also the recovery
record option in winRAR is handy for those who
fancy themselves unlucky.
Both fair points, although for a lossy
transmission medium, you can use PAR entirely
separate from RAR. They do integrate well,
but actually don't need to go together.
You can PAR a set of ZIPs or 7Zs just as well
as a set of RARs.
Slashdot Mirror
Is "Linux x86 assembly" any different to any other kind of "x86 assembly"?
Yes. Although it requires understanding the CPU's native capabilities to the same degree, Linux uses AT&T syntax, whereas most of the Wintel world uses (unsurprisingly) Intel/Microsoft syntax.
Personally, although I far prefer coding C under Linux, I prefer Intel syntax assembly. Even with many years of coding experience, I find AT&T syntax unneccessarily convoluted and somewhat difficult to quickly read through.
The larger idea holds, however, regardless of what assembler you use. I wholeheartedly agree with the FP - People who know assembly produce better code by almost any measurement except "object-oriented-ness", which assembly makes difficult to an extreme. On that same note, I consider that as one of the better arguments against OO code - It simply does not map well to real-world CPUs, thus introducing inefficiencies in the translation to something the CPU does handle natively.
The whole point behind port knocking is the wrong impression that "open" ports are more insecure than "closed" ports. This is totally bogus.
Tell me, then, why we all rush to patch our SSH installations every time a new weakness comes out?
I agree, in theory, a visibly open port does not pose a security risk. Something, however, has to listen on that port, and programmers do make mistakes.
Adding something like port knocking (see my previous post on this topic for my idea on a more useful generalization of port knocking), a potential attacker needs to find a weakness in both the knock sequence and the SSH (or whatever) server.
Kernel code is not intrinsically more secure than application level code!.
I'll agree with you wholeheartedly, on that point. In fact, I'd call kernel code potentially less secure, since, while bad user-space code can cause a few problems, bad kernel code can easily compromise an entire system. And, while I'd like think that kernel hackers take extra care for that exact reason, I'll reiterate - Programmers do make mistakes.
but you don't need port knocking for this. Just stuff the key in the SYN packet. Don't accept if it doesn't have the magic word.
Putting the key in a SYN packet, you can snoop. knocking on a seemingly random set of ports, with some care in the implementation (such as using the current time as a component to the correct sequence), you cannot.
I agree, a "good" implementation of this would require, at the very least, a firewall module... Though not necessarily poking holes in it, since none of those holes actually let traffic through.
Personally, I'd like to see this idea extended to a general "machine" password, totally separate from any security associated with a particular server bound to a port - That way, you don't need to modify a million and one individual programs that already speak TCP/IP (both clients and servers), you just need a single extra program (a doorbell, if you will). Without ringing the doorbell, you simply can't connect to anything listening on that machine.
I agree, though, that this goes beyond what the link in the FP had in mind. But take that original idea just a bit further, and you have something that sounds rather useful.
This is equivalent to to putting a password on access to the port.
This seems much better than a password, I would think (Though I certainly would still use a password as well).
As an analogy, if you want to get into a house, and find a locked door, you have a few options... You can try one of those M x N position key blanks, which will take a very very long time (exhaustive search). You can try to pick it (exploit a weakness in the password algorithm). You can try to get ahold of a copy of the real key (packet sniffing, "shoulder surfing", etc). But you have no doubt that somewhere, a key exists that will open that door.
Now compare that to a solid block of concrete, roughly the size of a house. What does it do? Do helicopters land on it? Does it cover something, or hold something down? Does it have something sealed inside it? You'd never suspect that that, if you utter the magic phrase "Sim sala bim bamba sala do saladim", a door will appear in the side of this large concrete block, allowing those with a key to gain entrance.
The main difference involves knowing whether or not a way in exists. With just a passworded port, an attacker knows that enough effort will pay off. Adding in port knocking, that attacker doesn't know whether or not their hard work can ever gain them entrance, since a port might well not exist.
Now, in my opinion, the more interesting question here involves how to hide this from one's ISP (ie, make it snoop-proof).
It is what you should be doing today.
Heh... That almost comes across as though you wrote it with a straight face...
Good one. Heh... Pay a "use tax" on something we manage to buy without a second (or third) raping by the IRS (and the individual states', and counties', and cities' version thereof)? Hilarious.
It amazes me that people put up with the idea of sales tax at all. The US constitution has these funny ideas about "taxation without representation". They exist for a reason, and the entire concept of "sales tax" violates that, big-time.
Anyone but a complete moron either buys online (until Kansas and the like ruin that for us), or at the very least saves up "big" purchases for a trip to a nearby sales-tax free state.
That's not my experience
:-)
Well, welcome to the club, Mr. "offtopic" for a perfectly apropos response to me.
Damn, though, you'd think the Mac fanboys would at least read far enough to see you defending them. Sad.
Oh, and in case you wonder (I know I would, our positions reversed), I had nothing at all to do with that OT mod.
Please explain how the economy will survive when there is no longer a middle class because all the white-collar jobs have been moved over seas.
Simple, really...
"Would you like fries with that?".
Of course, we may have to say that in Hindi, or Mandarin, or Japanese, but the "economy" will survive, in some sense of the word.
If we run out of non-mac mods your post will tank. I've had +5 go down to 0 after a few days with mac moderators still attacking my posts moderations.
/. karma, since for the most part people seem
to consider my posts decent, and I've kept
"excellent" karma for a couple of years now.
/. mod system when I metamod (in fact, I
don't think I've ever seen anything that
I would consider "unfair" while metamoding...
At worst, a few that I didn't really find all
that funny, but not bad enough to rate as
distinctly "unfunny"). Personally, I think /. needs to shift the focus of metamoderation
to detect abuses of down-mods, rather than
up-mods. I mean, except for glaringly obvious
"I posted from my other account and modded myself
up", people tend not to abuse positive moderation.
Same here... I tend to not really care about my
However, I tread rather lightly when it comes to anything Apple-related... Even guarded criticism of glaringly obvious facts result in a post getting moderated into oblivion.
Sadly, I have yet to see such blatant abuse of the
But, so it goes. Except for certain topics (such as anything Apple), the moderation system seems to work tolerably.
Registers are not as important as you may think.
Above some "decent" number, I'd tend to agree with you. Standard IA32 does not, however, have anywhere near enough.
"Enough" means not needing to use registers with clear designated purposes (such as ESI, EDI, and yes, even EBP) just to keep a triply nested loop (that actually does something) from hitting memory.
"Enough" means not needing to remember if CX gets clobbered on a given shift-like operation, so we need to preserve it temporarily.
"Enough" means having the ability, on performing a set of similar operations on a vector (not talking about using SSE or the like here), to interleave enough groups of operations to prevent a register dependancy stall...
Enough already. <G>. "Enough" means "more than 8" (of which you can only ever use at most seven, and even that with some difficulty).
Most consumers are going to sitck with 32-bits for some time to come since it currently offers a cost advantage and there is no real advantage for them to move to 64-bit.
I really don't understand why so many people have decided to speak poorly of 64b CPUs... Sure, not all that many programs will benefit from them right at the moment. But some will. And if we really want to take the "we don't need more bits", why not go back to 16b? or even 8b? Hell, let's just use bitslicers - We could probably clock one of those puppies well into the hundreds of gigahertz, due to the design simplicity.
The mathematical hobbyists (such as myself) will switch first, for the obvious number-crunching advantages. I already look forward to getting my hands on an Athlon64, to port a few of my projects to. Even better than the machine word length, though, I look forward to having a decent number of registers. Yum.
Then the gamers, as soon as they realize that 64b->64b->256b (CPU to AGP to GPU) performs quite a lot better than 32b->64b->256b...
Next will come either another boon to the gamers, or to the artists, depending on whether "killer game 2005" or "Photoshop 2005" comes out first with x86-64 optimizations.
Last will come the casual users, but they will only buy in when they have no choice (really, although 8b doesn't suffice, 16b works just fine for 99% of what a PC does.. So why do they even use 32b?).
Overall, I agree with you - you don't need 64b to browse the web and email grandma. For anything that either pushes the CPU to its limits, or improves linearly with the addition of more RAM, 64b has an obvious advantage.
As an aside, I already mentioned the increased number of registers in AMD's 64b offerings... While a lot of people have debated this issue only on the merits of "why do I need 64b?", the x86 line has always had a SEVERE shortage of GPRs. Doubling that number will (not just "could") boost performance immensely.
Prove it.
Prove it? Um... huh?
Fine, here we go, embarking on an exercise in futility - educating the ignorant trolls of the world. Twit.
For paging to disk, which would you say takes more time: "hit the disk", or "nothing"? Even if only verifying the pagefile's integrity at boot (not currently the case on any OS I know of, though feel free to name a few, of which I can save you the trouble and tell you that you won't find either "Linux", "Windows", or "OS-X" as an example), that still comes out to T>0 vs T=0.
For a ramdisk... Wow, where to start?. Okay... Do you suppose loading file X from RAM, or from a set of glass disks with mechanical actuators moving across them, takes longer? Keep in mind those moving parts read bits off the glass plates, perform several layers of verification and error correction, place them in local RAM on the drive itself, and only then does the requested data make it to the CPU. So again, factoring out the "RAM" step common to both, this reduces to "something" vs "nothing".
Maybe you shouldn't be making a fool of yourself then.
I have enough confidence in my statements to post as myself, rather than AC. I occasionally make mistakes, and people call me on it, but really, I can't even imagine what you had in mind when you decided to post your comments...
However, feel free to vindicate yourself by proving the counterpoint to my assertions? I'd love to hear a bit about the world you live in, where a disk performs faster than RAM...
Oh my. PC's catch up to CP/M circa '81 and AmigaDOS circa '86. Put your C compiler, sources and vi in ramdisk and things were just nice, even on a floppy based system
Heh, I agree this seems like a throwback to an older generation of computer hardware, but one that works rather well.
Personally, I don't understand why we still have virtual memory - Another throwback to the days of now-obsolete hardware, when memory cost an arm and a leg yet programs started wanting tons of it.
The difference between the two, now that we have oodles of cheap RAM, involves performance. Disk paging (even totally idle) hurts performance, while a ramdisk improves it. Yet all modern Windows version force you to use the former (though you can technically disable it in XP, some programs simply will not work if you do so, regardless of how much physical RAM you have), and don't even include native support for the latter.
How do you do ramdisk in W2K/XP (I'm too lazy to Google it)?
It doesn't come with Windows, but a number of 3rd-party programs (mostly all based off a reference MS driver, ironically enough) exist to do it for you.
Personally, I like QSoft's, which installs easily and works well. The free version theoretically has a 64MB limit, but I've had no problem telling it to use half-a-gig (and the pulldown menu goes to 1GB, but I don't have a machine with enough RAM to try that one).
And, as I mentioned, just 64MB makes a HUGE difference in responsiveness under XP. Still helps on 2k, but not nearly as noticeable.
And Intel had 64-bits long before AMD
Yup... And while an Itanium 2 currently goes for $1500, the Opteron 248 comes in at under $1000, with Athlon64 3400s going for $400.
Oh - and it makes Opteron look like my grandma when it comes to the real domain of 64-bit computing - science and engineering.
In theory, the Itanium-2 beats the Opteron, even at a much lower clockrate. In practice, the benchmarks I've seen (mostly just summaries of material from Spec) show at most a 50% difference on comparable hardware. And that, only on native 64b code. For the desktop market, where 32b will still exist as a significant portion of the available programs for at least another few years, the Opteron and Athlon64 suffer *no* performance hit on 32b code, while the Itanium line slows to a crawl.
Overall, I'd agree that the Itanium fills a particular niche fairly well. With an Intel-optimized Fortran compiler, I doubt anything else can even come close, clock-for-clock. For general-purpose computing, however, AMD has made tradeoffs that I strongly believe will pay off, and not just over the long-term. For desktop use, this seems like a no-brainer - Go with AMD.
I can't see a need for it on the desktop just yet.
One word - Ramdisk.
Set yourself up with a nice 1GB ramdisk, and make everything with a user-settable temp and/or cache directory point to it (most media apps, MSIE, Mozilla/Netscape). Also point your "temp" and "tmp" environment variables to it.
Personally, I still use 2K and consider XP a total dog. Circumstances recently having forced me to use it quite a bit, however, I found that it hits the disk constantly. Setting just a 64MB ramdisk (on a machine with only 256MB) literally doubled its responsiveness. And theoretically, this should increase the life, and decrease the fragmentation, of your HDD.
The one down side, some programs will complain if they can't fit their data on the temp dir (so for example, in Netscape, which for some unknown reason stores your downloads there before moving them to their real destination, downloading a 70MB file with only a 64MB temp ramdisk will make it complain). At a GB or two, though, I expect that would very rarely cause any problems.
Another reason, that the recent so-called "benchmark" failed to consider - We've had 128b and even 256b GPUs for a while now. AGP (and PCI64, currently rare but 64 bit CPUs should push it into the mainstream) will benefit GREATLY from native 64b drivers. Perhaps we'll finally see Doom 3?
Besides, if the PC can also serve as a PVR or other cool functions, then it's a doubly-good deal.
I think that point matters the most, to those of us who can roll our own...
Let's consider price-vs-functionality...
A Tivo or ReplayTV will set you back around $200, not counting a few hundred more for their subscription service. So let's say a nice round $500 for something that will remain useful for the next five years. A mid-range TV, such as my household recently purchased (30" HDTV 16:9), $1000. A decent DVD player, $200. Add in a 6.1 surround decoder, and tolerable-but-low-end speakers, and you have another $500. Total of around $2200, and it only has TV, DVDs, and PVR capabilities (plus basic audio functionality, such as radio and CD playing).
For a PC... The TV you really still need, since PC monitors cost far more per inch than a good HDTV. However, you don't need the Tivo, or the DVD player, or the 6.1 decoder (though still the speakers). PC, $500. Pioneer A-06, $130. SB Audigy 2, $120. Decent tuner card (personal choice on these varies a lot, though Hauppage seems the most common brand of choice), $200. Total comes out to around $2100 - Already ahead of the home theater setup.
The PC setup really excells in the "features" category. First of all, it can do everything the home theater setup does, which obviously counts as the basic level of functionality. It can also play any digital media format you can imagine (ie, anything a PC can play)... Some standalone DVD players now support DivX, but how about XVid? Or AVI? Or raw MPEG-2 streams? Or QT? Then we have content that actually runs as a form of program, such as Flash.
Additinally, a HTPC does work as a general purpose PC. You won't want to read small text for several hours a day on even a high-quality TV, but to look something up on-line, or a quick email check, they do rather nicely. Enjoy classic console video games? Well, download an emulator for your favorite system (you can get them for anything before the current gen, up to and including PS1, N64, and GBA), and they look sweet, without needing to connect 27 dongles to your TV's coax-in, or ever blowing the dust out of a cartridge again.
Finally, you can burn your stored media to DVD. A set-top box to do that, even if the PVR would let you, will set you back another grand. Network support? A $20 NIC (assuming your motherboard didn't include one) rather than a $500 "appliance".
So the price comes out ahead (not even considering "advanced" features that most people familiar with PCs take for granted), and the feature set just destroys a standard home theater system. Considering this, we shouldn't ask "which costs less and does more", but rather, "Why the hell do home theater systems cost so much and do so little?". Even if you do plan to blow $5k to $15k on a kickass home theater setup, spend the cash on a better TV and sound system, not obscenely priced components, when a $1k PC will literally "do it all" for you.
Also, anyone seriously wanting a home theater setup should budget a bit for environmental considerations as well - Spending $1000 on light-control and sound suppression will improve overall viewing enjoyment far more than. A reflected glare on the screen, or a noisy HVAC system that kicks on every half hour, makes more of a difference in quality than an extra 2 inches of screen or 4ch-vs-6.1 surround.
Professors don't care. In fact in some cases they are paid to select the more expensive of two options
Not true. Most professors do care, and many that I had even took rather extreme steps to minimize our book costs. My CS professors used as many free web resources as possible; my math profs tended to use, for example, the 1968 classic text that you could get for a pittance at any little buck-a-book shop; I even had one psych professor that actually photocopied books for us, handing out a packet containing the relevant pages every few classes.
The one exception to this, where I agree with you 100%, involves having a professor who wrote their own book. Those miserable bastards will gouge you any way they can. I actually had one "invite" me to drop his class because I insisted on using the (only trivially changed) previous edition that I picked up used on-line for $10.
Seriously... I realize new books cost a lot, but the previous edition (especially if the campus bookstore will no longer carry it) costs a pittance, like $5-$10 if you can find it used online. And you almost always can.
As for the questions changing (the only real difference between most editions that I ever noticed), go to the library and photocopy those two pages per chapter. Adds an extra buck to the cost of the book, but still comes out quite a lot less than $80-$200 per book.
However, to address the deeper issues here - Well, publishers have every right to charge what they want. And with a captive audience, they just have to dream of dollar-signs and they get it. At this point, though, they've gone beyond anything even remotely realistic, and I for one do not feel the least bit of guilt for having used exactly the method I describe above. I'd rather have had the correct books, but on a college student's budget, the $300+ savings per semester buys a LOT of instant ramen noodles.
When I went to get my passport renewed, people had to surrender their batteries.
Uhhh... Why? You walk in, get a picture taken, hand them the forms, and leave. Nothing really "secret" that you could take a picture of, even if you wanted to.
Of course, ironically enough, they probably would have let you take an actual camera in, and only bother about phones to play follow-the-leader in implementing paranoid policies about camera-phones.
Sad.
Anyone know if the reason 15% of people falsely think they have an HDTV television is because they only have EDTV? Seems like a fair reason, since no one has heard of EDTV.
I think that it has more to do with the general lack of TVs with an actual HDTV tuner in them. For example, my household recently purchased a new widescreen HDTV, which probably would not count as HDTV under that criterion. It does 480p, and DVDs look as good as they physically can on it (pixel-wise, I don't claim that a $15k system wouldn't look a tad better). But I have absolutely no use for an HDTV tuner, so why pay up to $200 more for it?
Paying $1.99 is fine for me for most movies.
If you have Movie Gallery stores in your area, they let you rent overnight on Wednesdays for a buck. This counts as a "normal" rental, so it still gives you points toward free rental coupons. Their normal bulk renting deals apply as well, so you can literally rent five non-new releases for three bucks on Wednesdays, and each month (I think it takes 12 points...), you get a coupon for a freebie.
If you have Blockbusters (who doesn't?), and still use a land-line phone, switch to the "MCI Neighborhood Complete" package... It includes all the standard phone services (CID, VM, 3-way, CW, etc), unlimited free LD, and even covers your local service, for just over $50 per month. And, to make it relevant to my point, you'll get a free rental coupon for every $25 you spend (ie, 2 per month). In combination with the BB Rewards card (actually pretty crappy, except that they send you a coupon for one free rental per month, and only costs $10 for a year), as well as the occasional special deal ("Pre-order Movie X, and get 10 free rentals", where even if Movie X sucks, the free rentals would have cost more than the movie does), I literally never pay to rent at Blockbuster.
Anyone know of any other good ways to rent for a pittance at other stores?
Even a dual-layer DVD can be copied if you're willing to accept some (usually negligible) quality loss.
From my experience with backing up my own collection, you don't even need to suffer any quality loss... Most dual-layer discs include the menu, the movie, and a ton of extras.
The menu doesn't usually take up much room (usually under a quarter of a gigabyte), but the extras can easily go to 2-3GB alone. Add to that a Spanish, French, and three redundant English soundtracks (DTS, 6ch, and 2ch occur often... You only need to keep one of them, and AC3-6ch sounds pretty damned close to DTS, at half the size).
So, if you remaster it as a movie-only disc, removing all the extraneous soundtracks, you can usually fit the movie with no recompression on single 4.7GB disc.
Here's info on a demo Pioneer did at this year's CES with a DVR-A06 and some tweaked firmware.
Ah, that makes me SO happy I went with the A06... Great drive all around, but this? Wonderful! Now if they'll just release that firmware (or someone leak it, I don't really care about the legality of it).
Not with my positive-nitrogen-pressure DVD player! Although piping the vent outside so I didn't suffocate was a pain. And it's a bit of a hassle lugging around all those nitrogen gas containers.
:-(
Ah, see, you need to thing "renewable", rather than keep getting more nitrogen. Just use a sealed positive-pressure glove-box, with an air-lock to add new discs, to store the player and your collection. That way, a small tank of nitrogen will last you a year, easy. You waste too much in changing discs, see, if you just have an open positive-pressure system.
And, no vent needed. Even in the even of a complete breach of the seal, it won't contain enough nitrogen (assuming you turn off the tank) to asphyxiate you in any normal-sized room.
Y'know, actually, although I joke about this, I did intend to set up a 10-gallon aquarium to do something very much like what I suggest above. Too bad this flopped before we could abuse it to its fullest.
Consumers know when they rent something that it has to go back - they don't own it. They're accustomed to this paradigm.
:-)
Perhaps true, but it doesn't change the fact that people do forget to return movies on time. Blockbuster, and presumeably all major video stores, report that over half of their profits come from late fees. To me, that really says something (well, mostly, "don't return movies late").
So, for someone who knows they will always return their movies late, paying $7 up front may not seem all that bad. Personally, I'd suggest they just return the damn things on time, but hey, just my opinion.
It didn't support spanning. Also the recovery record option in winRAR is handy for those who fancy themselves unlucky.
Both fair points, although for a lossy transmission medium, you can use PAR entirely separate from RAR. They do integrate well, but actually don't need to go together. You can PAR a set of ZIPs or 7Zs just as well as a set of RARs.