However, I propose that the starting altitude should be 0m from the surface too (how did the thing take off?).
Good luck even defining what the surface is on Jupiter and the other gas giants. You'll need some kind of platform to launch from first there, so you can postulate any altitude you want. You'll want to pick one with atmospheric density similar to that of Earth...
What's dangerous is not running a red light, what's dangerous is passing the light when cars from the other direction are already entering the crossing. So what matters is not the time between yellow and red, what matters is the time between yellow on my side and green on the other side.
Plus the phasing for other directions might be different (e.g., a dedicated cross-traffic turn phase) or there might be a pedestrian-exclusive phase. (Some jurisdictions have them, others don't.) All you really know when you see a red light is that you're not supposed to be entering the junction at that point. That's even true if it is a junction you know well; a highway engineer might've just altered the sequence for all you know for sure. Cars are dangerous (if very convenient) and so should be driven carefully.
Red light cameras don't decrease crashes. What happens when the guy 4 feet in front speeds up at the yellow, and you follow, then he slams the brakes because he changes his mind because of the camera? Oh yeah, more crashes.
Then you were driving incompetently. You shouldn't tail-gate. You should always leave enough room for you to stop if the guy in front does something strange like stamp on the brakes or swerve or something. Yes, they might be a lot to blame but you're still supposed to take care of yourself by anticipating the (immediate) future road conditions and driving so that you remain safe. Didn't you ever get taught that as part of showing you're fit to drive on the public highway?
And the worst crashes are when someone is more than a second after the red. The tickets go out to people like you describe at 0.5s after the red. But it's those seconds late (drunk, asleep, reading the morning paper) that kill, and they don't see the red light, they won't see the camera.
So, you're insisting that because cameras don't prevent all idiotic driving at an intersection, they're useless? I really don't agree, not at all. If you're behind the wheel, you should be fit to be driving safely, if not for yourself then for all your other fellow road users. That means being sober, alert and attentive. If you're not all three when driving, you're just a fucking jerkwad whose travel should be restricted to walking around the prison exercise yard.
Before you ask, I'm just as strict with myself about driving safely. Safely or not at all. No excuses. No third option. (Being a passenger when someone else is driving safely instead — bus, taxi, whatever — is a variant on "not at all".)
In the What-if it's explicitly stated that the gas tanks have been replaced with batteries and had an electric engine installed.
And also that this means that the plane won't fly for very long anyway, around 10 minutes. Not that this is a particular issue on many of the worlds of the solar system (unless you can also make the Cessna acid-proof, which would help a lot for Venus).
Let's look at your idea: you want something that can charge an electric car's battery in 10 seconds. Ok; a typical Prius battery is rated at about 4 kWh. That's roughly 15 million Joules of energy. To deliver that much energy in 10 s, you need a power supply that provides 1.5 million watts of power. At the battery voltage (~275 V), that's a current of over 5000 A, or only an order of magnitude less than a typical lightning strike.
So... you're saying that we should work on powering our cars with lightning strikes?
The mad scientist in me approves! Igor? Igor! Clear that lump of meat off the table and plug the Prius in: we've got science to do!
I suppose it depends on where the average lies....
Big population in both cases, and intelligence is a complex attribute that depends on many variables, so assume bell-curve distributions in both cases. That puts the average at about 50%, and since China has more than twice the population of the US, you'd expect just from mathematics to have "China has more people of above-average intelligence than America has people." be true. There could be some skewing, but the population-level differences between people are fairly small (much smaller than individual-level differences) and wouldn't overcome the massive difference in populations.
In other words, wave your prejudice around all you want you fool: basic math still wins.
Seriously, who really cares about that platform? It's just not got the market penetration. Now, the 360 has the penetration, but the hardware there is looking rather elderly; the visual quality of the platform is noticeably worse than its competitors and it's pretty clear it needs a platform-refresh or successor, which would be a reasonable time to ramp up support for OpenGL.
Yeah, but if the law in Germany says otherwise, then it is the law that applies. That's the difference between laws and contracts (of which a license agreement is just a small part) and it's actually impossible to have a contract to break the law; contracts must be lawful or they are simply not contracts by definition. Even if the agreement says that it is not conducted under German law, German consumers will have the right to use German law anyway. (Well, probably; I've not actually checked what the relevant law says, but there's a lot of similarity in this area across different EU members and I know that UK law is very clear on this point.)
The real question is not whether there's recourse in law, but how any ensuing judgements would be enforced. An unenforceable ruling really isn't much use.
A trademark also has a clearly delineated function: it must act to indicate that a particular product comes from a particular source. There is a "customers can obviously tell us apart" defense in trademark law that doesn't apply to patents, and a trademark that ceases to be distinctive dies forever.
Trademarks also must be non-functional (they can only be a distinctive mark that serves to identify the company, product or service) and rapidly cease to be applicable if not actively used.
This means that nothing you might ever be interested in for interoperability is trademarkable, the lifespan is only while the mark is being used for its proper purpose, and the only reason for knowingly infringing a trademark is if you are looking to be a scumbag leech or worse. (There are awkward cases around the edge over exactly what constitutes distinctiveness, but that's why there are courts to resolve these things.) I'm not sure if the layout of a store really should be trademarkable; that's sailing very close to the wind in terms of having other functionality.
I have a bunch of software I've been very willing to set free (it has already even GPL3 headers!) but I can't, because it might be publishable one day.
And so, it'll keep on being hidden...
Sounds like you need to publish more often rather than whinging about the whether it "might be publishable one day" on slashdot. There's no point in sitting on stuff so long that it becomes irrelevant, and code that's just cowering in a dark corner of your disk might as well not exist at all.
The public KSK of the root is known by all people/software that want to check dnssec signatures (the weak point since how do you securely distribute and update that one?).
The usual way with PKI is to have two identities involved in the root. One, the master, has a public key very widely known and with a very long life, and only ever used to validate the "operational key"; the master private key is kept offline in a safe somewhere. Perhaps with armed guards or something like that. The operational key is what is used to validate child domains, and as such is in use a lot more and so is more exposed. On the other hand, you can generate new ones (with only the hassle of the armed guards) without needing to update all the consumers of the keys; operationally, that's entirely practical.
I have no idea if DNSSEC is set up to work this way. It's quite possible that it isn't, with clients assuming that the identity of the agent authorizing the root zone is its own lonesome thing.
In fact the problems in the CA system we currently have directly stem from such a hierarchical trust scheme. We would be much better of going with a truly distributed system for SSL key validation.
I'm unconvinced. (I'm particularly unconvinced by the handwave-assert-jedi-mind-trick style of argument there, but that's by-the-by.) The fundamental problem is that it is very hard to work out if the assertions in a public certificate are true; all you can tell is that the information was digitally signed by someone or something. With a web of trust model, either you have non-transitive trust (which totally doesn't scale at all!) or you have transitive trust, in which case all it takes is for one person to get it wrong and the bad guys get in (and their first acts will be to seek to leverage their new trustedness to obscure how they got in). I suppose you could have someone acting as an authority that says who can be trusted to handle transitive links, but that's virtually back to the CA model except with plenty more technical complexity than before.
I hope the outcome is that our government does a re-think on being part of the WTO in the first place
Careful for what you wish for; the outcomes might not be what you expect. For example, it would also mean that the rest of the world would feel no need at all to enforce copyrights held under US law, including on a lot of Free Software, or at least not until the negotiation of a whole new set of bipartite treaties.
I wonder whether it would be legal now (if not necessarily moral) for an Antiguan citizen to do derivative works of software where the copyright holder is the FSF and change the license to a different one (e.g., a BSD variant)...
[Perl is] not failing because it's not changing, it's failing because less people are using it. The lack of it integrating shiny new features may be one of the factors contributing to this.
Maybe, but probably not. It's probably more closely linked with the culture of dodgy human-readability and the changing profile of what people want to do with programming languages. On the first point, the issue isn't that it is impossible to write readable perl — that's emphatically not true — but rather that most perl programmers don't do that; it's a culture thing. The second point is much more to do with perl not being universally present in browsers, or numpy, or rails: some niches where perl wasn't the leader have turned out to be really important (and that's not to deny perl's success in others).
You can't win them all, and the only constant is change.
TeX (and its important derivative, LaTeX) utterly changed how academic publishing in mathematics and the sciences was done. Before, most published papers were typescripts (and difficult to read; formulæ were horribly mangled by this approach) and afterwards, every paper could be read quickly and easily. Total night-and-day difference. Even the best-produced papers before (which used tools like RUNOFF) were nothing like as good; TeX was transformative software from the future.
Don't believe me? Try reading old journals in an academic library. Lots of good ideas mouldering away there, yet they're almost without exception difficult to read purely by virtue of their formatting...
Of course it is anti-competitive and anti-consumer. Why do you think the US carriers are so keen on it? They're consistently anti-consumer, and put a lot of effort into persuading the "regulators" (I use that word advisedly) remain sympathetic to their point of view.
Well in this case the programming failed under normal use. That is it failed to keep people out.
We can easily secure systems such that the bad guys can't ever get in. Really. It's easy to do even. What is much harder is doing this while allowing the authorized users easy access. In the limit case of security, we just disconnect the system from the network and power it down: nobody will hack it then, and it is ever so easy to get right! But this is immensely inconvenient for people who are supposed to use the system. (To be fair, there are systems that have data so valuable that at least keeping them off the net and protected by armed guards makes sense. That's a fairly extreme level of security.)
The natural tension in the security area is with ease of use. Make things too secure and you raise the difficulty of use so much that users find other insecure methods of handling the data that you don't know about (e.g., post-it notes with passwords) and so circumvent you protections. Indeed, reducing the official level of security can actually increase the amount of practical security by encouraging people to do things properly...
In the case of buildings, normal use would include extreme weather and earthquakes etc depending on the area.
Within reason. There's a trade off with how much you want to spend. That's basic economics. Spending lots to protect against extremely rare events is a bad use of resources.
Nothing is perfect, but you don't punish people who identify flaws, especially not at a so-called place of learning.
Grasshopper, you have much to learn. (Also, how much damage was done during the "identifying flaws"? There's a very fine line there.)
Are typical university CS department professors doing meaningful "research"?
Should a "typical university" have a CS department at all? Speaking as someone who works in a CS department where the academic staff have to produce research output as well as teach, it sounds like there are places which just ought to stop the pretense and to actually call themselves "Visual Basic Training Schools" or something. (Disclosure: I mostly don't teach, and instead do software engineering to turn the CS research into practical tools to support other research areas.)
Is it true that the only/easiest way to kill a cane toad is by putting it in a plastic bag (carefully, you don't want too much of their toxin on your hands) and popping it in the freezer?
False. Like any squishy animal, they die when you drive over them, and that's pretty easy. Alternatively, set fire to the area and they'll suffocate and burn. Easy as arson!
(No, I'm not suggesting you do that. I'm suggesting that your question is utterly foolish as even a moment of thought would reveal.)
Do they not see that they, being amongst the upper middle class, will be some of the first ones hurt by higher taxes and health care costs?
They might think that they're already being hurt excessively like that, and that the Dems are more likely to change things in a way that they favor. Or they might be valuing other objectives more highly where the Dems have a stronger policy advantage (in the eyes of those concerned) such as action on climate change or infrastructure investment. It doesn't matter whether you agree with them. It's their free decision, just as you're free to decide what you do, and presumably they and you have a different weighting of the importance of various objectives.
The real problem is when you've got zealots around who reject all compromise and who drag their party with them. Right now, that's especially the Tea Party (which is blocking the rest of the Reps from negotiating properly) but the risk is always present; the Dems, for all their faults, are mostly sitting on their zealots. For now. Who knows what will happen in a few years?
I'm concerned that it might push too much development resources to get X & Y working which is popular for the gaming community but not for all other Linux / ''Nix users (personal, business, enterprise...).
I'm sure that there will continue to be many developers who won't work on the gaming side of things. On the other hand, if supporting gaming means that we finally get all the niggling media support problems sorted out, that's a good thing. The APIs don't have to be the best for very advanced use, but they must at least work reliably for simpler uses across very wide sets of configurations. (It's a matter of going from 95% done to 100% done; right now things work for almost everyone, but it's a lot of work to get universality and hence a proper stable platform for applications. The last 5% is the hardest part, of course.)
Slashdot Mirror
However, I propose that the starting altitude should be 0m from the surface too (how did the thing take off?).
Good luck even defining what the surface is on Jupiter and the other gas giants. You'll need some kind of platform to launch from first there, so you can postulate any altitude you want. You'll want to pick one with atmospheric density similar to that of Earth...
What's dangerous is not running a red light, what's dangerous is passing the light when cars from the other direction are already entering the crossing. So what matters is not the time between yellow and red, what matters is the time between yellow on my side and green on the other side.
Plus the phasing for other directions might be different (e.g., a dedicated cross-traffic turn phase) or there might be a pedestrian-exclusive phase. (Some jurisdictions have them, others don't.) All you really know when you see a red light is that you're not supposed to be entering the junction at that point. That's even true if it is a junction you know well; a highway engineer might've just altered the sequence for all you know for sure. Cars are dangerous (if very convenient) and so should be driven carefully.
Red light cameras don't decrease crashes. What happens when the guy 4 feet in front speeds up at the yellow, and you follow, then he slams the brakes because he changes his mind because of the camera? Oh yeah, more crashes.
Then you were driving incompetently. You shouldn't tail-gate. You should always leave enough room for you to stop if the guy in front does something strange like stamp on the brakes or swerve or something. Yes, they might be a lot to blame but you're still supposed to take care of yourself by anticipating the (immediate) future road conditions and driving so that you remain safe. Didn't you ever get taught that as part of showing you're fit to drive on the public highway?
And the worst crashes are when someone is more than a second after the red. The tickets go out to people like you describe at 0.5s after the red. But it's those seconds late (drunk, asleep, reading the morning paper) that kill, and they don't see the red light, they won't see the camera.
So, you're insisting that because cameras don't prevent all idiotic driving at an intersection, they're useless? I really don't agree, not at all. If you're behind the wheel, you should be fit to be driving safely, if not for yourself then for all your other fellow road users. That means being sober, alert and attentive. If you're not all three when driving, you're just a fucking jerkwad whose travel should be restricted to walking around the prison exercise yard.
Before you ask, I'm just as strict with myself about driving safely. Safely or not at all. No excuses. No third option. (Being a passenger when someone else is driving safely instead — bus, taxi, whatever — is a variant on "not at all".)
In the What-if it's explicitly stated that the gas tanks have been replaced with batteries and had an electric engine installed.
And also that this means that the plane won't fly for very long anyway, around 10 minutes. Not that this is a particular issue on many of the worlds of the solar system (unless you can also make the Cessna acid-proof, which would help a lot for Venus).
It's just what-iffery.
Let's look at your idea: you want something that can charge an electric car's battery in 10 seconds. Ok; a typical Prius battery is rated at about 4 kWh. That's roughly 15 million Joules of energy. To deliver that much energy in 10 s, you need a power supply that provides 1.5 million watts of power. At the battery voltage (~275 V), that's a current of over 5000 A, or only an order of magnitude less than a typical lightning strike.
So... you're saying that we should work on powering our cars with lightning strikes?
The mad scientist in me approves! Igor? Igor! Clear that lump of meat off the table and plug the Prius in: we've got science to do!
I suppose it depends on where the average lies....
Big population in both cases, and intelligence is a complex attribute that depends on many variables, so assume bell-curve distributions in both cases. That puts the average at about 50%, and since China has more than twice the population of the US, you'd expect just from mathematics to have "China has more people of above-average intelligence than America has people." be true. There could be some skewing, but the population-level differences between people are fairly small (much smaller than individual-level differences) and wouldn't overcome the massive difference in populations.
In other words, wave your prejudice around all you want you fool: basic math still wins.
What you miss out on is [...] Windows Phone.
Oh no! Save me!
Seriously, who really cares about that platform? It's just not got the market penetration. Now, the 360 has the penetration, but the hardware there is looking rather elderly; the visual quality of the platform is noticeably worse than its competitors and it's pretty clear it needs a platform-refresh or successor, which would be a reasonable time to ramp up support for OpenGL.
Yeah, but if the law in Germany says otherwise, then it is the law that applies. That's the difference between laws and contracts (of which a license agreement is just a small part) and it's actually impossible to have a contract to break the law; contracts must be lawful or they are simply not contracts by definition. Even if the agreement says that it is not conducted under German law, German consumers will have the right to use German law anyway. (Well, probably; I've not actually checked what the relevant law says, but there's a lot of similarity in this area across different EU members and I know that UK law is very clear on this point.)
The real question is not whether there's recourse in law, but how any ensuing judgements would be enforced. An unenforceable ruling really isn't much use.
Yes, but this also means using Eclipse which also means crying yourself to sleep each night.
You mean, you wait to cry until after it has finished starting up in the morning and applying its updates?
A trademark also has a clearly delineated function: it must act to indicate that a particular product comes from a particular source. There is a "customers can obviously tell us apart" defense in trademark law that doesn't apply to patents, and a trademark that ceases to be distinctive dies forever.
Trademarks also must be non-functional (they can only be a distinctive mark that serves to identify the company, product or service) and rapidly cease to be applicable if not actively used.
This means that nothing you might ever be interested in for interoperability is trademarkable, the lifespan is only while the mark is being used for its proper purpose, and the only reason for knowingly infringing a trademark is if you are looking to be a scumbag leech or worse. (There are awkward cases around the edge over exactly what constitutes distinctiveness, but that's why there are courts to resolve these things.) I'm not sure if the layout of a store really should be trademarkable; that's sailing very close to the wind in terms of having other functionality.
We don't have death panels. They were lying to you.
However, we'd be very keen on death panels if they were considering the fates of politicians.
I have a bunch of software I've been very willing to set free (it has already even GPL3 headers!) but I can't, because it might be publishable one day.
And so, it'll keep on being hidden...
Sounds like you need to publish more often rather than whinging about the whether it "might be publishable one day" on slashdot. There's no point in sitting on stuff so long that it becomes irrelevant, and code that's just cowering in a dark corner of your disk might as well not exist at all.
The public KSK of the root is known by all people/software that want to check dnssec signatures (the weak point since how do you securely distribute and update that one?).
The usual way with PKI is to have two identities involved in the root. One, the master, has a public key very widely known and with a very long life, and only ever used to validate the "operational key"; the master private key is kept offline in a safe somewhere. Perhaps with armed guards or something like that. The operational key is what is used to validate child domains, and as such is in use a lot more and so is more exposed. On the other hand, you can generate new ones (with only the hassle of the armed guards) without needing to update all the consumers of the keys; operationally, that's entirely practical.
I have no idea if DNSSEC is set up to work this way. It's quite possible that it isn't, with clients assuming that the identity of the agent authorizing the root zone is its own lonesome thing.
In fact the problems in the CA system we currently have directly stem from such a hierarchical trust scheme. We would be much better of going with a truly distributed system for SSL key validation.
I'm unconvinced. (I'm particularly unconvinced by the handwave-assert-jedi-mind-trick style of argument there, but that's by-the-by.) The fundamental problem is that it is very hard to work out if the assertions in a public certificate are true; all you can tell is that the information was digitally signed by someone or something. With a web of trust model, either you have non-transitive trust (which totally doesn't scale at all!) or you have transitive trust, in which case all it takes is for one person to get it wrong and the bad guys get in (and their first acts will be to seek to leverage their new trustedness to obscure how they got in). I suppose you could have someone acting as an authority that says who can be trusted to handle transitive links, but that's virtually back to the CA model except with plenty more technical complexity than before.
I hope the outcome is that our government does a re-think on being part of the WTO in the first place
Careful for what you wish for; the outcomes might not be what you expect. For example, it would also mean that the rest of the world would feel no need at all to enforce copyrights held under US law, including on a lot of Free Software, or at least not until the negotiation of a whole new set of bipartite treaties.
I wonder whether it would be legal now (if not necessarily moral) for an Antiguan citizen to do derivative works of software where the copyright holder is the FSF and change the license to a different one (e.g., a BSD variant)...
[Perl is] not failing because it's not changing, it's failing because less people are using it. The lack of it integrating shiny new features may be one of the factors contributing to this.
Maybe, but probably not. It's probably more closely linked with the culture of dodgy human-readability and the changing profile of what people want to do with programming languages. On the first point, the issue isn't that it is impossible to write readable perl — that's emphatically not true — but rather that most perl programmers don't do that; it's a culture thing. The second point is much more to do with perl not being universally present in browsers, or numpy, or rails: some niches where perl wasn't the leader have turned out to be really important (and that's not to deny perl's success in others).
You can't win them all, and the only constant is change.
I've seen M$ shoot themselves in the foot before, I have never seen them do it with such a large canon
You've got me thinking about a modern day Friar Tuck there, sitting in an A10 and blowing Ballmer's feet off...
TeX (and its important derivative, LaTeX) utterly changed how academic publishing in mathematics and the sciences was done. Before, most published papers were typescripts (and difficult to read; formulæ were horribly mangled by this approach) and afterwards, every paper could be read quickly and easily. Total night-and-day difference. Even the best-produced papers before (which used tools like RUNOFF) were nothing like as good; TeX was transformative software from the future.
Don't believe me? Try reading old journals in an academic library. Lots of good ideas mouldering away there, yet they're almost without exception difficult to read purely by virtue of their formatting...
Of course it is anti-competitive and anti-consumer. Why do you think the US carriers are so keen on it? They're consistently anti-consumer, and put a lot of effort into persuading the "regulators" (I use that word advisedly) remain sympathetic to their point of view.
Well in this case the programming failed under normal use. That is it failed to keep people out.
We can easily secure systems such that the bad guys can't ever get in. Really. It's easy to do even. What is much harder is doing this while allowing the authorized users easy access. In the limit case of security, we just disconnect the system from the network and power it down: nobody will hack it then, and it is ever so easy to get right! But this is immensely inconvenient for people who are supposed to use the system. (To be fair, there are systems that have data so valuable that at least keeping them off the net and protected by armed guards makes sense. That's a fairly extreme level of security.)
The natural tension in the security area is with ease of use. Make things too secure and you raise the difficulty of use so much that users find other insecure methods of handling the data that you don't know about (e.g., post-it notes with passwords) and so circumvent you protections. Indeed, reducing the official level of security can actually increase the amount of practical security by encouraging people to do things properly...
In the case of buildings, normal use would include extreme weather and earthquakes etc depending on the area.
Within reason. There's a trade off with how much you want to spend. That's basic economics. Spending lots to protect against extremely rare events is a bad use of resources.
Nothing is perfect, but you don't punish people who identify flaws, especially not at a so-called place of learning.
Grasshopper, you have much to learn. (Also, how much damage was done during the "identifying flaws"? There's a very fine line there.)
Are typical university CS department professors doing meaningful "research"?
Should a "typical university" have a CS department at all? Speaking as someone who works in a CS department where the academic staff have to produce research output as well as teach, it sounds like there are places which just ought to stop the pretense and to actually call themselves "Visual Basic Training Schools" or something. (Disclosure: I mostly don't teach, and instead do software engineering to turn the CS research into practical tools to support other research areas.)
Is it true that the only/easiest way to kill a cane toad is by putting it in a plastic bag (carefully, you don't want too much of their toxin on your hands) and popping it in the freezer?
False. Like any squishy animal, they die when you drive over them, and that's pretty easy. Alternatively, set fire to the area and they'll suffocate and burn. Easy as arson!
(No, I'm not suggesting you do that. I'm suggesting that your question is utterly foolish as even a moment of thought would reveal.)
Do they not see that they, being amongst the upper middle class, will be some of the first ones hurt by higher taxes and health care costs?
They might think that they're already being hurt excessively like that, and that the Dems are more likely to change things in a way that they favor. Or they might be valuing other objectives more highly where the Dems have a stronger policy advantage (in the eyes of those concerned) such as action on climate change or infrastructure investment. It doesn't matter whether you agree with them. It's their free decision, just as you're free to decide what you do, and presumably they and you have a different weighting of the importance of various objectives.
The real problem is when you've got zealots around who reject all compromise and who drag their party with them. Right now, that's especially the Tea Party (which is blocking the rest of the Reps from negotiating properly) but the risk is always present; the Dems, for all their faults, are mostly sitting on their zealots. For now. Who knows what will happen in a few years?
Republicans with their friendly corporations can hire better programmers and put together equal or better code.
Or at least they can hire more expensive programmers. Or outsource to India; they might as well stick to their true principles after all...
I'm concerned that it might push too much development resources to get X & Y working which is popular for the gaming community but not for all other Linux / ''Nix users (personal, business, enterprise...).
I'm sure that there will continue to be many developers who won't work on the gaming side of things. On the other hand, if supporting gaming means that we finally get all the niggling media support problems sorted out, that's a good thing. The APIs don't have to be the best for very advanced use, but they must at least work reliably for simpler uses across very wide sets of configurations. (It's a matter of going from 95% done to 100% done; right now things work for almost everyone, but it's a lot of work to get universality and hence a proper stable platform for applications. The last 5% is the hardest part, of course.)