Buffer overruns have absolutely nothing to do with garbage collection.
I disagree. Programmers use fixed-sized buffers because it is difficult to properly allocate and manage dynamically-sized buffers in "C".
Say I want to write a function that returns multiple foo items (parsed from a file).
In Java, I just create a new foo[] array or an ArrayList and return it.
In "C", I can either: dynamically alloc() space for a returned array or linked list of foo items, and hope that all possible callers free() it at the appropriate time (which may well require reference counts, etc. and will almost inevitably lead to leaks or bad pointers) ; or I can just cram them into a fixed-size array that the caller passed me. Guess what programmers end up doing 99% of the time (and too often without checking the bounds)...
Java's been successful, and not "free" under RMS's definition. That's obviously driving him nuts.
And the irony is: with Java you do have RMS's desired freedoms, the important ones anyway.
Thanks to the bytecode design and decompilers like "jad", you can almost always get a workable approximation of the original source (which has saved my butt more than once, when dealing with buggy third-party jars for which I didn't have actual source). Try doing that with C or C++.
I'm a lot more free to run Java programs than "C" on the hardware of my choice, because I generally don't have to port or recompile them; as opposed to dealing with bizarre and nonstandard, broken, or outright unavailable POSIX functionality on a dozen different platforms (Surprise, "rename()" doesn't work as expected on OS/400, you have to use "Qp0lRenameUnlink()". Surprise, the "timezone" variable doesn't work on z/OS. Surprise, Solaris lets you forget to link against -lpthread, then blows up at runtime. etc. etc. etc.)
But Oh No! Java's not "free-as-in-freebase" (which is what RMS seems to be high on) -- let's stick to "C" instead, nevermind if our software keeps crashing with buffer overruns (ultimately due to lack of garbage-collection) or from unexpected system call behavior. We can't use better technology, because it's [shudder] proprietary.
Yes, and their product demo page is especially unhelpful. There's not a single photo of the actual case, just a bunch of animated figures. I wonder if they even have the product ready, or if this is just a cheap attempt to sell pre-ordered vaporware.
Except that there shouldn't be anything inherently wrong with a corporation giving a candidate money. It's the candidate's responsibility to behave correctly.
I see. Do you think it should be illegal for corporations to bribe law enforcement or courtroom judges? After all, it's also their responsibility to behave correctly.
If it's not okay to bribe the people who enforce or judge the laws, why is it okay to bribe the people who write the laws?
Perhaps slashdot (and similar high-traffic portals) could check for a.slashdot-policy.txt file, with options like:
cached (cache only, please!)
direct (link straight 2 us, we can take it)
denied (please do not cache or link our content)
Sites with no.slashdot-policy.txt will be treated like ones with denied ; editors will refrain from posting any front-page links to them until/unless the site owner can be contacted to set a policy.
That's true for filters that are based on the message content. IMHO, the only reliable way to define spam is: "messages that show up at spamtrap addresses".
You're right. It is almost certain that the outside world should be able to see your "shadow puppet" at night.
Instead of having the fibers run straight through the block, maybe they could twist or braid them; so that light could still seep through but would form random shadow blobs instead of clear silhouettes.
Gosh...an ISP might decide to not carry alt.binaries groups whose main purpose is to distribute copyrighted material without permission?
ObCynic: If ISPs stopped carrying alt.binaries.whatever,
the pirates would just flood other groups. Do you propose
ISP just stop carrying all non-moderated groups?
Roads are designed so the water runs off them onto the sides (where there are usually gutters or ditches). Every time it rains, it's washing a little more of the radioactive dust away from the center of the road and towards the edges.
Nanotubes are much stronger than steel, so won't they slice through the grommets; and the polyurethane uppers of your shoes; and the meat and bones of your feet?
[ Dr. Wierd ] "Gentlemen! Behold! My new... stumps!"
If the store had a sufficiently dense network of RFID scanners, perhaps they could repeatedly poll everything in the store, every 30 seconds or so, to see if anything's not answering (and to alert store security if so). They presumably have video of the item's last known location as well...
Oh, great. One of the proponents is a bulk-emailer called "Goodmail", who wants this system because if they pay to send out spam (with the postage going to ISPs), the ISPs will have a financial incentive not to block them.
I live in a town with one local non-cable ISP. Who would I switch to if they got listed? Answer: noone. Since paying for a remote colo is outside my budget, I'd have no options except to cease operations.
Untrue: you could "smarthost" ; continue to use the spam-friendly local ISP to connect to the net, but relay your outgoing mail through a more responsible ISP somewhere outside your town. SPEWS filterers would now see the mail coming from an IP address that's not blocked.
(You would of course have to pay for the smarthost ISP
account, in addition to your corrupt local ISP account;
that's why local monopolies suck.)
Re:Journey to the center of the earth
on
The Year In Ideas
·
· Score: 1
The one thing i did not understand [...] is how would you attach a sensor package with telemetry to a pile of goo at several thousands of degrees f.
Well, the probe can't get hotter than 2750 C., or the molten iron blob would boil away and the ride would be over.
There are several substances which are still solid at this temperature: carbon, tungsten, thorium and magnesium oxide, etc. Carbon is a conductor; MgO is an insulator; and ThO2 won't dissolve in molten iron. So, it might be possible to build an electronic device out of refractory substances that can function in a bath of molten iron.
SF writers always look their most foolish when trying to make up future slang or cuss-words.
OK, the producer doesn't wanna use the real f-word. Understandable. That's what [bleep] is for. Or else 'fricking' (although Austin Powers has pretty much spoiled that euphemism...).
Now how long do you think you're going to survive if you're glowing like a neon sign saying "Eat Me"
Depending on the chemicals that make them glow, other fish might quickly learn to view them as neon signs saying "I AM TOXIC" or at least "I TASTE BAD".
Our goal is not to blow up Linux [..] but for us to get a transaction fee every time it's sold. That's really our goal.
Little Darl says you can't give it away. Everybody has to pay and pay. A license here, and a license there, Salt Lake City was the place where
they said, "Hey SCO, take your lips off the crack pipe." "Hey sugar, Linux don't infringe Sys V." And the IBM lawyers go, doo-de-doo, de-doo,
doo-dee-doo-doo, countersue you...
To stop this phishing technique, browsers ought to pop up a warning dialog for URLs with a username field (especially if it contains one or more dots). Something like:
| Alert -- Actual URL is: | | Domain Path: badpeople.hackedsite.ru/hahaha | Username: www.citibank.com | Password: verify=
This would at least highlight the real site the link is pointing to.
The spammers are either doing it to get accidental clickthroughs from careless readers, or to improve their Google rank by spreading lots of links to their spamsite.
So: why not just filter "http://" from all comments? The rest of the URL still shows up, so readers can copy-n-paste it if they're/really/ interested in visiting that link; but it won't be a clickable link and won't affect the Google ranking.
Disney currently makes about $3 billion a year licensing Winnie the Pooh movies, books, toys, pajamas, breakfast cereals, etc.
A. A. Milne died in 1956, so they/really/ need all the life+50 countries to please extend their copyright term to life+70 as soon as possible, so Pooh doesn't inadvertantly go public-domain.
1. SCO gets court ruling that GPL is invalid. 2. Now nobody can ship Linux. Bill Gates sez "W00T!" 3. SCO stock falls to 0.01 since they can't ship Linux either. 4. Darl and pals quietly buy all outstanding shares. 5. Mysterious strangers who can't be tied to Microsoft now
exercise previous options to buy SCO stock at $15/share.
Darl sez "WOOT!" and retires to Ibiza.
>> So you're saying that Linux should make it easier for users to run scripts and executeables they receive in the mail?
Yes, if you want the average non-geek to use it. BUT, it also needs to make it so that those scripts and executables they get in the mail aren't capable of infecting permanent changes outside a secured sandbox.
Slashdot Mirror
I disagree. Programmers use fixed-sized buffers because it is difficult to properly allocate and manage dynamically-sized buffers in "C".
Say I want to write a function that returns multiple foo items (parsed from a file).
In Java, I just create a new foo[] array or an ArrayList and return it.
In "C", I can either: dynamically alloc() space for a returned array or linked list of foo items, and hope that all possible callers free() it at the appropriate time (which may well require reference counts, etc. and will almost inevitably lead to leaks or bad pointers) ; or I can just cram them into a fixed-size array that the caller passed me. Guess what programmers end up doing 99% of the time (and too often without checking the bounds)...
And the irony is: with Java you do have RMS's desired freedoms, the important ones anyway.
Thanks to the bytecode design and decompilers like "jad", you can almost always get a workable approximation of the original source (which has saved my butt more than once, when dealing with buggy third-party jars for which I didn't have actual source). Try doing that with C or C++.
I'm a lot more free to run Java programs than "C" on the hardware of my choice, because I generally don't have to port or recompile them; as opposed to dealing with bizarre and nonstandard, broken, or outright unavailable POSIX functionality on a dozen different platforms (Surprise, "rename()" doesn't work as expected on OS/400, you have to use "Qp0lRenameUnlink()". Surprise, the "timezone" variable doesn't work on z/OS. Surprise, Solaris lets you forget to link against -lpthread, then blows up at runtime. etc. etc. etc.)
But Oh No! Java's not "free-as-in-freebase" (which is what RMS seems to be high on) -- let's stick to "C" instead, nevermind if our software keeps crashing with buffer overruns (ultimately due to lack of garbage-collection) or from unexpected system call behavior. We can't use better technology, because it's [shudder] proprietary.
Yes, and their product demo page is especially unhelpful. There's not a single photo of the actual case, just a bunch of animated figures. I wonder if they even have the product ready, or if this is just a cheap attempt to sell pre-ordered vaporware.
I see. Do you think it should be illegal for corporations to bribe law enforcement or courtroom judges? After all, it's also their responsibility to behave correctly.
If it's not okay to bribe the people who enforce or judge the laws, why is it okay to bribe the people who write the laws?
Perhaps slashdot (and similar high-traffic portals) could check for a .slashdot-policy.txt file, with options like:
Sites with no .slashdot-policy.txt will be treated like ones with denied ; editors will refrain from posting any front-page links to them until/unless the site owner can be contacted to set a policy.
That's true for filters that are based on the message content.
IMHO, the only reliable way to define spam is: "messages that show up at spamtrap addresses".
Instead of having the fibers run straight through the block, maybe they could twist or braid them; so that light could still seep through but would form random shadow blobs instead of clear silhouettes.
ObCynic: If ISPs stopped carrying alt.binaries.whatever, the pirates would just flood other groups. Do you propose ISP just stop carrying all non-moderated groups?
Roads are designed so the water runs off them onto
the sides (where there are usually gutters or ditches).
Every time it rains, it's washing a little more of the
radioactive dust away from the center of the road and
towards the edges.
Er...
... stumps!"
Nanotubes are much stronger than steel, so won't they
slice through the grommets; and the polyurethane uppers
of your shoes; and the meat and bones of your feet?
[ Dr. Wierd ]
"Gentlemen! Behold! My new
>;K
If the store had a sufficiently dense network of
RFID scanners, perhaps they could repeatedly poll
everything in the store, every 30 seconds or so,
to see if anything's not answering (and to alert
store security if so). They presumably have video
of the item's last known location as well...
Oh, great. One of the proponents is a bulk-emailer called "Goodmail", who wants this system because if they pay to send out spam (with the postage going to ISPs), the ISPs will have a financial incentive not to block them.
Untrue: you could "smarthost" ; continue to use the spam-friendly local ISP to connect to the net, but relay your outgoing mail through a more responsible ISP somewhere outside your town. SPEWS filterers would now see the mail coming from an IP address that's not blocked.
(You would of course have to pay for the smarthost ISP account, in addition to your corrupt local ISP account; that's why local monopolies suck.)
Well, the probe can't get hotter than 2750 C., or the molten iron blob would boil away and the ride would be over.
There are several substances which are still solid at this temperature: carbon, tungsten, thorium and magnesium oxide, etc. Carbon is a conductor; MgO is an insulator; and ThO2 won't dissolve in molten iron. So, it might be possible to build an electronic device out of refractory substances that can function in a bath of molten iron.
Yes, I remember. At least they aren't measuring time in "centons" anymore.
SF writers always look their most foolish when
... M'kay!)
trying to make up future slang or cuss-words.
OK, the producer doesn't wanna use the real f-word.
Understandable. That's what [bleep] is for. Or else
'fricking' (although Austin Powers has pretty much
spoiled that euphemism...).
(Remember, don't say f**k, just say
Now how long do you think you're going to survive if you're glowing like a neon sign saying "Eat Me"
Depending on the chemicals that make them glow, other
fish might quickly learn to view them as neon signs
saying "I AM TOXIC" or at least "I TASTE BAD".
I thought we needed /EVIL/ Billy Big-mouth Bass /laser beams/ on their foreheads.
with frickin'
Wait: wasn't there an "Ask Slashdot" post last week,
about what to do with 120 old laser pointers...
Little Darl says you can't give it away.
Everybody has to pay and pay.
A license here, and a license there,
Salt Lake City was the place where
they said,
"Hey SCO, take your lips off the crack pipe."
"Hey sugar, Linux don't infringe Sys V."
And the IBM lawyers go, doo-de-doo, de-doo,
doo-dee-doo-doo, countersue you...
[saxophone riff]
To stop this phishing technique, browsers ought to
pop up a warning dialog for URLs with a username
field (especially if it contains one or more dots).
Something like:
| Alert -- Actual URL is:
|
| Domain Path: badpeople.hackedsite.ru/hahaha
| Username: www.citibank.com
| Password: verify=
This would at least highlight the real site the
link is pointing to.
The spammers are either doing it to get accidental
/really/ interested in
clickthroughs from careless readers, or to improve
their Google rank by spreading lots of links to their
spamsite.
So: why not just filter "http://" from all comments?
The rest of the URL still shows up, so readers can
copy-n-paste it if they're
visiting that link; but it won't be a clickable link
and won't affect the Google ranking.
Solar power and nuclear power. Both could sustain our current energy needs until the sun burns out
(at a somewhat higher price, mind you...)
Disney currently makes about $3 billion a year
/really/ need
licensing Winnie the Pooh movies, books, toys,
pajamas, breakfast cereals, etc.
A. A. Milne died in 1956, so they
all the life+50 countries to please extend their
copyright term to life+70 as soon as possible,
so Pooh doesn't inadvertantly go public-domain.
Disturbing Hypothetical Business plan:
1. SCO gets court ruling that GPL is invalid.
2. Now nobody can ship Linux. Bill Gates sez "W00T!"
3. SCO stock falls to 0.01 since they can't ship Linux either.
4. Darl and pals quietly buy all outstanding shares.
5. Mysterious strangers who can't be tied to Microsoft now
exercise previous options to buy SCO stock at $15/share.
Darl sez "WOOT!" and retires to Ibiza.
>> So you're saying that Linux should make it easier for users to run scripts and executeables they receive in the mail?
Yes, if you want the average non-geek to use it. BUT, it also needs to make it so that those scripts and executables they get in the mail aren't capable of infecting permanent changes outside a secured sandbox.