The "amount" of homework means little when its content is trivial, and does not do anything but repeat something that should be obvious based on what is learned in class. Application of knowledge to a trivial task just doesn't do anything other than insult the student, however the application of the same to something even slightly challenging, is both useful for remembering the material, and good thinking practice in general.
Of course, making homework less of a mindless chore and more an exercise in thinking means that there will be always some students, who will be unable to complete it because of their insufficient abilities and poor motivation. My response for that will be, SCREW THEM! They won't get much good from a shitty homework, either, and if they are going to drag everyone down into the horrors of rote memorization, there is always a short bus for them, and decent education for the rest. Treating everyone like a retard, accomplishes nothing positive.
Padio/tape -- worked well because the tape mechanism easily fits into the radio case, and devices share the same amplifier and speakers.
VCR/TV combos -- ended up being small, cheap, low-end devices because they are saving the material on separate cases and cables, and don't do much more.
Printer/fax/scanner devices -- again, low-end, and prone to "one thing failed, you are screwed" problems. Those devices are basically a fax machine with improved quality (and fragility) of components, plus a parallel or USB port to talk to the computer.
Motherboard with integrated graphics, sound and ethernet -- integrated graphics usually is low-quality, suitable for either a "dummy vga" on servers, or small office boxes that don't need a good video card. Integrated sound often isn't anything spectacular, either, and only builtin ethernet is actually a good solution suitable for all kinds of devices.
So, we have:
1. Successful combinations such as radio/tape and motherboard/ethernet -- in both cases integration did not require any compromise, both combinations preserved all functionality and reliability, and eliminated the need in additional devices (amplifiers, PCI slots).
2. Low-end devices that benefit from combination of components that are small and simple enough to combine, losing the possibility for upgrade, such as tv/vcr combo and motherboards with builtin video (minus the upgrade problem).
3. Devices that gained multiple functionality by becoming an overgrown version of another device, such as a fax machine, becoming a high-end solution for original purpose and a low-end solution for new purposes.
PDA/phone combo are usually in the second category -- usually it's either not a good phone, not a good PDA, or not a good either. However if the communications infrastructure will sufficiently improve to make it practical to provide a fast flat-rate or nearly flat-rate data service, high-end PDAs may simply acquire the phone functionality because they would already have everything necessary for a phone built in, and no compromise would be necessary. For example, if cities will implement their "full wifi coverage" projects, people would just install SIP clients on their PDAs, and voila, instant phones.
I am writing this response on the box with two Athlons, cooled by acetone thermosyphons (thermosyphon is a heat pipe without a wick -- it relies on gravity instead of surface tension).
Actually the box where Firefox is running, is a 1u server with both thermosyphons squeezed into 1.75", and I am sitting in front of an ancient P3-450 that I use as a "thin client".
And yes, Kazakhstan is a country (formerly USSR member), where the Baikonur Cosmodrome is located -- this is where Russian ISS-related flights are launched.
The only thing in there that is anywhere close to the truth is that Soyuz was assembled by hand -- though it does not mean that it was not mass-produced, considering the accepted practice in aerospace industry worldwide. For everything else I can't even begin explaining how wrong it is.
What are you talking about? Both Mir and ISS are modular, and many ISS modules were launched on Protons, just like Mir. Proton and Shuttle are very old designs, so even comparison between them looks embarrassing now. It's insane to assume that a modern heavy rocket would not be far superior to both for launching the space station modules -- if only that was the direction of development instead of shuttle patching.
However I am talking specifically about routine manned missions, and not space station construction. You need neither heavy rocket, nor a shuttle for that -- you have to reach the station, bring people and cargo, sit for a while connected to the station, and return people plus whatever little amount of cargo should be brought back to Earth. And with all improvements that were done to Shuttle design, it is suboptimal for pretty much every thing that it is applied to -- too small to launch large modules, or be useful by itself, too cumbersome to be a good temporary part of the station, too expensive to launch often, too fragile to perform emergency... pretty much emergency anything, leave alone landing... I guess, there is something it's good for, but without a different vehicle, designed for things that Shuttle does poorly, the whole space program suffers, and looks bad compared to the situation 20 years ago.
...but why, oh why, an old, simple combination of Salyut/Mir and Soyuz/Progress ships constantly visiting it, was a much more reliable, convenient, useful and cheaper than all this pretending-to-do-2001-the-space-odyssey-remake stuff?
No one to rescue -- Soyuz docks with Salyut/Mir, all work is done in a relatively large station + modules, and if anything wrong happens, there is another Soyuz attached.
No giant airplane-thing to land -- a small landing capsule is the last thing you would expect to fail (not that there weren't early failures, but that was long ago).
Soyuz can sit attached to the station being actually useful, with its living space, fuel and engines, as opposed to the shuttle that mostly produces corrosive gas and stress on the flimsy station.
If anything is REALLY wrong, another Soyuz can be launched in a reasonable time, and without some insane risk, as long as the Khrunichev factory will continue making what by then can be considered mass-produced parts, as opposed to unique shuttles.
That was the state of the art two decades ago. Six Salyuts plus Mir operated like this. And there was more scientific work done than bickering and genitalia-waving between participants in those projects (bickering and waving between the countries was another story though). Can we now make something that isn't significantly worse than things that flied 20 years ago?
"Executives in $250M/year to $500M/year companies don't know, what systems their engineers are running. If the company is smaller, executives likely know more about what they manage, and if the company is larger, it's an IBM client."
...that most of "savings" from defective chips manufacturers will get from users that WON'T BOTHER TO RMA PRODUCTS AFTER A COMPLETE FAILURE. After all, if chip manufacturers will consistently produce crappy chips, customers won't have an option to get a better brand, the quality will be dropped across the board.
I can't remember seeing a public disclosure of a Linux (or any Unix) vulnerability that did not contain a description of either patch or workaround, or wasn't followed by such within hours in Bugtraq, so unless those researchers used sources unavailable to the public, the vulnerability window for a system administered by a bugtraq-reading admin would be zero, or hours per vulnerability.
Counting the time that Red Hat takes to issue their official patch for their "Enterprise" product would show the upper limit that applies to "infinitely lazy" admin that only run auto-update. This is reasonable for a home desktop system, however I doubt that anyone runs RHEL on those. I guess, even counting "infinitely lazy" admin's updates in Debian or Gentoo, the disclosure to patch time would be much less than for Red Hat.
Another issue is that disclosure does not mean exploit -- exploit could exist before, or appear after the disclosure. Many Linux vulnerabilities end up unexploited because they are published after the patch is issued, and only few are exploited before the first patch or workaround release. Famous Debian servers' compromise was a result of a known by that time kernel hole, and even though Debian project's sysadmins initially believed that it was an unknown hole, it happened to not be the case, and I guess, they have changed their security policy based on that.
On the other hand, Windows exploits commonly happen before the disclosure, there is no workaround published at the time of disclosure, and often patches are issued late, don't cover all vulnerable versions, have dangerous side effects, or are bundled with things that can be only described as "unrelated shit".
chroot doesn't affect a processes namespace, it just affects path name resolution so one can easily escape the chroot with "/..".
This can only work if you are root user in a chroot environment -- what any sane secure design avoids or limits to a small, secure part of code. And no one places setuid binaries into chroot environment, so privileges elevation can be only a result of a kernel bug -- what is not unheard of (recently patched in Linux), but is a very uncommon compared to other vulnerabilities.
that humans are capable of using (that is, they can remember and type them) is approximately the same as the number of pass phrases because phrases contain common words. If every pass phrase was replaced by an abbreviation ("Mary had a little lamb 88aapzF" -> "marhalilmb88aapzF"), there would be a pretty low number of collisions, and abbreviations would be usable as short passwords that are just as good as the phrases they were derived from. Therefore this idea produces nothing but an increased amount of typing.
If by "usable" you mean "able to misrepresent itself" then yes, it's useless. I don't necessarily buy your definition of usable.
Also, define "work". If by "work" you mean "convince others to interoperate by lying about itself", then it won't work. Guess what? I don't buy your definition of work either.
"Lying" is irrelevant to the matter -- "Trusted computing" and DRM are not persons, and have no legal meaning as concepts. They are parts of the device, that is sold as a general-purpose computer, so it's supposed to run software, and indeed that device runs some software -- one that someone distributed with an encrypted key, and provided a "source", that, by definition in the license, should be the preferred form for modifications. However since the key is not distributed, any modification results in a code that can not run on the available hardware, thus making a source without a key something other than a "preferred for modification" form.
What the author of a DRM scheme (let's call things by their real names -- "Trusted Computing" is a DRM scheme) intended to do, does not make him "right" when applied to the license -- from the user's point of view, DRM vendor maliciously sabotaged the hardware sold to him, and the software distributor conspired with him by using and not distributing the key, thus intentionally breaking otherwise perfectly usable and modifiable source, turning it into a meaningless text.
There already were smartasses that tried to distribute obfuscated source, source that requires proprietary and unavailable for others tools to compile, and other crap like that. None of that would work with GPL because of the single phrase that I have quoted before -- where "preferrable for modification" was mentioned. Making a "signed" version and not distributing the key far exceeds those amateurish forms of source code sabotage, so it would be unreasonable to expect that it would work.
It's your responsibility to persuade others to talk to you. How do you do that is not anyone's business but yours.
Another, far superior strategy, is to thoroughly prove wrong and humiliate the obviously wrong or stupid opponent, so people who actually are going to make decisions, will have no doubts about the true value of his arguments. I really don't care if you agree with me or not, you are neither a judge, nor an executive in a hardware or software company that may make a decision that matters in this situation. Your only effect on this is in spewing propaganda.
You're making things up, don't you? It is written where exactly?
In GPL:
The source code for a work means the preferred form of the work for
making modifications to it.
No one "prefers" to make modifications that do not produce a usable executable.
Of course it does. You can sign it, run it, sell it, bake it, drink it, do whatever you want with it. What you can't do is convincingly misrepresent it to be something else than it is. Guess what, no one ever promised you such an ability, now or in the future.
Again, "preferred for modifications". See above -- no one cares what else can be done with it. No one guarantees that software, before or after modifications, will do anything in particular, and this is absolutely unrelated to the definition of "preferrable", that means that a software user should be able to produce a modified software from it. It may work better, worse, or not work at all, but if it's guaranteed to break after modification, it's not merely something that lack of warranty can cover. If you can guarantee that modified software will not work, this is no longer a preferrable form for modification, so your argument is invalid.
BTW, I have never seen this amount of stupidity expressed by a person without clear understanding how stupid it is, and a desperate desire to get something that does not belong to him. You are either extremely dumb, or a shill for DRM pushers.
GPL'ed software is distributed without any guarantees, however this is completely unrelated to all other clauses of the license. Therefore:
1. If the binary is distributed, the distributor MUST provide the source.
2. The source is defined as preferred form for modification, that can be compiled into binary.
3. The "signed" source, if modified, does not make a usable binary, NO MATTER WHAT the recipient does, therefore it is not a source code. Lack of warranty is absolutely unrelated to this -- the code that can not possibly produce a usable binary , if modified, is not any more "preferred form" for modification than a hex dump. Actually, a hex dump of a non-"trusted" executable is a far more preferrable form for modification because its modification sometimes does produce an improved product, and a modification of "trusted" source invariably produces an unusable product.
Therefore the requirement for signing makes it impossible to distribute the source code, unless the source code includes the private key.
GPL defines the source code as the form preferred for writing and editing the software. Not as a preferred form of editing things that can not be used after editing.
The original phrase, as Khruschev said it, was "We will show you Kuzka's mother", what in Russian is a mildly rude version of "We will show you!", and definietly was meant to be applied to competition in economy. At the same time, same Khrushchev promoted a slogan "Catch up and Overtake the US", that was also used, in an abbreviated form, as trademark used for some industrial equipment.
Admitting that there is a lot of "catching up" to do was pretty far from arrogance and aggressiveness that US propaganda attributed to Communists in 60's-80's.
First of all, I have to remind all of you (and especially to the obviously mentally deficient authors of the article) that chimeras are not hybrids because they do not have a combined DNA, and therefore can not produce offsprings that inherit their traits.
Second, I really don't see what the problem is. Anyone wants to write an angry email to the Pope about the possibility of keeping human organs in glass jars? No? What about metal jars? Plastic? Quartz?
Slashdot Mirror
The "amount" of homework means little when its content is trivial, and does not do anything but repeat something that should be obvious based on what is learned in class. Application of knowledge to a trivial task just doesn't do anything other than insult the student, however the application of the same to something even slightly challenging, is both useful for remembering the material, and good thinking practice in general.
Of course, making homework less of a mindless chore and more an exercise in thinking means that there will be always some students, who will be unable to complete it because of their insufficient abilities and poor motivation. My response for that will be, SCREW THEM! They won't get much good from a shitty homework, either, and if they are going to drag everyone down into the horrors of rote memorization, there is always a short bus for them, and decent education for the rest. Treating everyone like a retard, accomplishes nothing positive.
This idea has a spotty record.
Padio/tape -- worked well because the tape mechanism easily fits into the radio case, and devices share the same amplifier and speakers.
VCR/TV combos -- ended up being small, cheap, low-end devices because they are saving the material on separate cases and cables, and don't do much more.
Printer/fax/scanner devices -- again, low-end, and prone to "one thing failed, you are screwed" problems. Those devices are basically a fax machine with improved quality (and fragility) of components, plus a parallel or USB port to talk to the computer.
Motherboard with integrated graphics, sound and ethernet -- integrated graphics usually is low-quality, suitable for either a "dummy vga" on servers, or small office boxes that don't need a good video card. Integrated sound often isn't anything spectacular, either, and only builtin ethernet is actually a good solution suitable for all kinds of devices.
So, we have:
1. Successful combinations such as radio/tape and motherboard/ethernet -- in both cases integration did not require any compromise, both combinations preserved all functionality and reliability, and eliminated the need in additional devices (amplifiers, PCI slots).
2. Low-end devices that benefit from combination of components that are small and simple enough to combine, losing the possibility for upgrade, such as tv/vcr combo and motherboards with builtin video (minus the upgrade problem).
3. Devices that gained multiple functionality by becoming an overgrown version of another device, such as a fax machine, becoming a high-end solution for original purpose and a low-end solution for new purposes.
PDA/phone combo are usually in the second category -- usually it's either not a good phone, not a good PDA, or not a good either. However if the communications infrastructure will sufficiently improve to make it practical to provide a fast flat-rate or nearly flat-rate data service, high-end PDAs may simply acquire the phone functionality because they would already have everything necessary for a phone built in, and no compromise would be necessary. For example, if cities will implement their "full wifi coverage" projects, people would just install SIP clients on their PDAs, and voila, instant phones.
I am writing this response on the box with two Athlons, cooled by acetone thermosyphons (thermosyphon is a heat pipe without a wick -- it relies on gravity instead of surface tension).
Actually the box where Firefox is running, is a 1u server with both thermosyphons squeezed into 1.75", and I am sitting in front of an ancient P3-450 that I use as a "thin client".
And yes, Kazakhstan is a country (formerly USSR member), where the Baikonur Cosmodrome is located -- this is where Russian ISS-related flights are launched.
Please list all Linux kernel features that prevented your users from utilizing your software properly.
The only thing in there that is anywhere close to the truth is that Soyuz was assembled by hand -- though it does not mean that it was not mass-produced, considering the accepted practice in aerospace industry worldwide. For everything else I can't even begin explaining how wrong it is.
Nice to see, how denial is alive and well in US -- "No, dirty Russians couldn't do that, it must be all propaganda!".
In 10 more years you will insist that Soyuz never flied.
What are you talking about? Both Mir and ISS are modular, and many ISS modules were launched on Protons, just like Mir. Proton and Shuttle are very old designs, so even comparison between them looks embarrassing now. It's insane to assume that a modern heavy rocket would not be far superior to both for launching the space station modules -- if only that was the direction of development instead of shuttle patching.
However I am talking specifically about routine manned missions, and not space station construction. You need neither heavy rocket, nor a shuttle for that -- you have to reach the station, bring people and cargo, sit for a while connected to the station, and return people plus whatever little amount of cargo should be brought back to Earth. And with all improvements that were done to Shuttle design, it is suboptimal for pretty much every thing that it is applied to -- too small to launch large modules, or be useful by itself, too cumbersome to be a good temporary part of the station, too expensive to launch often, too fragile to perform emergency... pretty much emergency anything, leave alone landing... I guess, there is something it's good for, but without a different vehicle, designed for things that Shuttle does poorly, the whole space program suffers, and looks bad compared to the situation 20 years ago.
And the prize for missing the point in the most spectacular manner in the whole Slashdot history goes to... Anonymous Coward.
...but why, oh why, an old, simple combination of Salyut/Mir and Soyuz/Progress ships constantly visiting it, was a much more reliable, convenient, useful and cheaper than all this pretending-to-do-2001-the-space-odyssey-remake stuff?
No one to rescue -- Soyuz docks with Salyut/Mir, all work is done in a relatively large station + modules, and if anything wrong happens, there is another Soyuz attached.
No giant airplane-thing to land -- a small landing capsule is the last thing you would expect to fail (not that there weren't early failures, but that was long ago).
Soyuz can sit attached to the station being actually useful, with its living space, fuel and engines, as opposed to the shuttle that mostly produces corrosive gas and stress on the flimsy station.
If anything is REALLY wrong, another Soyuz can be launched in a reasonable time, and without some insane risk, as long as the Khrunichev factory will continue making what by then can be considered mass-produced parts, as opposed to unique shuttles.
That was the state of the art two decades ago. Six Salyuts plus Mir operated like this. And there was more scientific work done than bickering and genitalia-waving between participants in those projects (bickering and waving between the countries was another story though). Can we now make something that isn't significantly worse than things that flied 20 years ago?
"Executives in $250M/year to $500M/year companies don't know, what systems their engineers are running. If the company is smaller, executives likely know more about what they manage, and if the company is larger, it's an IBM client."
Warmongers -- secret police -- monopolists -- spammers -- malware writers.
...that most of "savings" from defective chips manufacturers will get from users that WON'T BOTHER TO RMA PRODUCTS AFTER A COMPLETE FAILURE. After all, if chip manufacturers will consistently produce crappy chips, customers won't have an option to get a better brand, the quality will be dropped across the board.
I can't remember seeing a public disclosure of a Linux (or any Unix) vulnerability that did not contain a description of either patch or workaround, or wasn't followed by such within hours in Bugtraq, so unless those researchers used sources unavailable to the public, the vulnerability window for a system administered by a bugtraq-reading admin would be zero, or hours per vulnerability.
Counting the time that Red Hat takes to issue their official patch for their "Enterprise" product would show the upper limit that applies to "infinitely lazy" admin that only run auto-update. This is reasonable for a home desktop system, however I doubt that anyone runs RHEL on those. I guess, even counting "infinitely lazy" admin's updates in Debian or Gentoo, the disclosure to patch time would be much less than for Red Hat.
Another issue is that disclosure does not mean exploit -- exploit could exist before, or appear after the disclosure. Many Linux vulnerabilities end up unexploited because they are published after the patch is issued, and only few are exploited before the first patch or workaround release. Famous Debian servers' compromise was a result of a known by that time kernel hole, and even though Debian project's sysadmins initially believed that it was an unknown hole, it happened to not be the case, and I guess, they have changed their security policy based on that.
On the other hand, Windows exploits commonly happen before the disclosure, there is no workaround published at the time of disclosure, and often patches are issued late, don't cover all vulnerable versions, have dangerous side effects, or are bundled with things that can be only described as "unrelated shit".
chroot doesn't affect a processes namespace, it just affects path name resolution so one can easily escape the chroot with "/..".
This can only work if you are root user in a chroot environment -- what any sane secure design avoids or limits to a small, secure part of code. And no one places setuid binaries into chroot environment, so privileges elevation can be only a result of a kernel bug -- what is not unheard of (recently patched in Linux), but is a very uncommon compared to other vulnerabilities.
281016rbelong2us ?
that humans are capable of using (that is, they can remember and type them) is approximately the same as the number of pass phrases because phrases contain common words. If every pass phrase was replaced by an abbreviation ("Mary had a little lamb 88aapzF" -> "marhalilmb88aapzF"), there would be a pretty low number of collisions, and abbreviations would be usable as short passwords that are just as good as the phrases they were derived from. Therefore this idea produces nothing but an increased amount of typing.
If by "usable" you mean "able to misrepresent itself" then yes, it's useless. I don't necessarily buy your definition of usable.
Also, define "work". If by "work" you mean "convince others to interoperate by lying about itself", then it won't work. Guess what? I don't buy your definition of work either.
"Lying" is irrelevant to the matter -- "Trusted computing" and DRM are not persons, and have no legal meaning as concepts. They are parts of the device, that is sold as a general-purpose computer, so it's supposed to run software, and indeed that device runs some software -- one that someone distributed with an encrypted key, and provided a "source", that, by definition in the license, should be the preferred form for modifications. However since the key is not distributed, any modification results in a code that can not run on the available hardware, thus making a source without a key something other than a "preferred for modification" form.
What the author of a DRM scheme (let's call things by their real names -- "Trusted Computing" is a DRM scheme) intended to do, does not make him "right" when applied to the license -- from the user's point of view, DRM vendor maliciously sabotaged the hardware sold to him, and the software distributor conspired with him by using and not distributing the key, thus intentionally breaking otherwise perfectly usable and modifiable source, turning it into a meaningless text.
There already were smartasses that tried to distribute obfuscated source, source that requires proprietary and unavailable for others tools to compile, and other crap like that. None of that would work with GPL because of the single phrase that I have quoted before -- where "preferrable for modification" was mentioned. Making a "signed" version and not distributing the key far exceeds those amateurish forms of source code sabotage, so it would be unreasonable to expect that it would work.
It's your responsibility to persuade others to talk to you. How do you do that is not anyone's business but yours.
Another, far superior strategy, is to thoroughly prove wrong and humiliate the obviously wrong or stupid opponent, so people who actually are going to make decisions, will have no doubts about the true value of his arguments. I really don't care if you agree with me or not, you are neither a judge, nor an executive in a hardware or software company that may make a decision that matters in this situation. Your only effect on this is in spewing propaganda.
You're making things up, don't you? It is written where exactly?
In GPL:
The source code for a work means the preferred form of the work for making modifications to it.
No one "prefers" to make modifications that do not produce a usable executable.
Of course it does. You can sign it, run it, sell it, bake it, drink it, do whatever you want with it. What you can't do is convincingly misrepresent it to be something else than it is. Guess what, no one ever promised you such an ability, now or in the future.
Again, "preferred for modifications". See above -- no one cares what else can be done with it. No one guarantees that software, before or after modifications, will do anything in particular, and this is absolutely unrelated to the definition of "preferrable", that means that a software user should be able to produce a modified software from it. It may work better, worse, or not work at all, but if it's guaranteed to break after modification, it's not merely something that lack of warranty can cover. If you can guarantee that modified software will not work, this is no longer a preferrable form for modification, so your argument is invalid.
BTW, I have never seen this amount of stupidity expressed by a person without clear understanding how stupid it is, and a desperate desire to get something that does not belong to him. You are either extremely dumb, or a shill for DRM pushers.
No.
GPL'ed software is distributed without any guarantees, however this is completely unrelated to all other clauses of the license. Therefore:
1. If the binary is distributed, the distributor MUST provide the source.
2. The source is defined as preferred form for modification, that can be compiled into binary.
3. The "signed" source, if modified, does not make a usable binary, NO MATTER WHAT the recipient does, therefore it is not a source code. Lack of warranty is absolutely unrelated to this -- the code that can not possibly produce a usable binary , if modified, is not any more "preferred form" for modification than a hex dump. Actually, a hex dump of a non-"trusted" executable is a far more preferrable form for modification because its modification sometimes does produce an improved product, and a modification of "trusted" source invariably produces an unusable product.
Therefore the requirement for signing makes it impossible to distribute the source code, unless the source code includes the private key.
GPL defines the source code as the form preferred for writing and editing the software. Not as a preferred form of editing things that can not be used after editing.
"We will bury you" is a horrible mistranslation.
The original phrase, as Khruschev said it, was "We will show you Kuzka's mother", what in Russian is a mildly rude version of "We will show you!", and definietly was meant to be applied to competition in economy. At the same time, same Khrushchev promoted a slogan "Catch up and Overtake the US", that was also used, in an abbreviated form, as trademark used for some industrial equipment.
Admitting that there is a lot of "catching up" to do was pretty far from arrogance and aggressiveness that US propaganda attributed to Communists in 60's-80's.
I call it "voice telnet".
And it sucks.
ffmpeg is open source, works on all platforms, and supports formats that are already popular, so many players will understand them.
First of all, I have to remind all of you (and especially to the obviously mentally deficient authors of the article) that chimeras are not hybrids because they do not have a combined DNA, and therefore can not produce offsprings that inherit their traits.
Second, I really don't see what the problem is. Anyone wants to write an angry email to the Pope about the possibility of keeping human organs in glass jars? No? What about metal jars? Plastic? Quartz?