The problem with mathematical models is that they tend to cause you to ignore reality in favor of the model.
There was this wonderful model for evaluating CDSes, and it said that Wall Street was made out of solid gold a few years ago.
Models also tend to encourage people to weigh decisions in favor of things that are easily measured as opposed to things that aren't easily measured. At work they're going nuts with Six Sigma, but it is all about coming up with arguments that will win elevator speeches. If you're evaluating some process you try to pick a half-dozen things at most to measure/evaluate. The problem comes that if you have a system with 500 independent variables and you pick six to optimize, then chances are you'll improve those six by slightly degrading the performance of the other 494. If you repeat that 25 times you end up with processes that aren't robust. Then when you throw in that things like "awesomeness" or "nice place to work" aren't easily measured and aren't often the top-priority of modern managers, you end up with processes that throw them out the window entirely.
Don't get me wrong - managers SHOULD know how to use models, but managers need to manage PEOPLE - not tweak formulas on spreadsheets.
Yeah, but leading an inspired R&D team is not something anybody can do. On the other hand, anybody can go down a budget and knock 10% off every line and fire the manager who misses the mark by the largest amount, then lather, rinse, repeat the next year.
A literal reading of the constitution would require a warrant to take something out of your pockets.
Now, the exception was generally made that when you're being arrested the officer is allowed to check you for weapons, check your ID, etc. That isn't unreasonable - he had to have probable cause to arrest you and checking for weapons just is being practical.
The problem is that it leads to things like - while I checked for weapons I happened to notice his gun was still hot, or that there was blood on his knife. Somehow that led to being able to fingerprint things, and then DNA, and so on. What started out as common sense will eventually turn into desurfacing the flash chips in your phone to read their contents with a STEM.
So, if you're archiving list traffic, by all means filter on the headers - that is just brain-dead simple.
However, in general if your search routine struggles with 100 emails/day * 7 years, then you need a better search routine. I think that is the biggest problem with most email clients - the search indexing is pretty lousy. If Google can index half of the Internet and retrieve results in milliseconds, then my email program should be able to sort through 200k emails in less than 10 seconds.
1. For mail that can be effectively sorted automatically, do so. I'm talking 99+% accuracy here - like filtering on list headers or something.
2. For situations where folder sorting helps with short-term projects/organization, go ahead. So, if I'm soliciting feedback on some proposal, go ahead and create a "feedback" folder that I put stuff in to go through later. Ditch the folder as soon as the task is completed. Another case of this is the high/med/low folders I create after I come back from vacation for a week or so.
3. Otherwise, just create one folder called "archive" and stuff everything into it.
The problem with organizing stuff into folders is that it costs you time for EVERYTHING you get, and it only saves you time if you go back to that particular email (and the folder helps you find it faster than a search would). What I've found is that 80% of my email is one-time-read stuff and so most of the time the effort to sort is completely wasted. Then, when I do need to go back unless it was just in the last few days I end up searching anyway, and if it is in the last few days I can usually quickly sort the archive folder. Now, folders can speed up lousy search algorithms, but the solution to that is to fix the brain-dead email software.
I do the same. At the start of this year I started using Gmail as my MUA at home, and decided to try a similar workflow at work (ugh, using Outlook). I set up an archive folder. It isn't quite the same thing, and is far more limited and much slower, but it works.
My only problem is that I tend to collect a lot of "half-done" stuff in my inbox. Periodically I tend to mass-purge it into my archive.
Now what I need is a decent search solution for outlook. Right now I'm using xobni, and it isn't bad, but it isn't great either. It is also painfully slow, but that could be the corporate IT stuff weighing it down.
My other issue is that I have a 1995ish email quota - another relic of corporate IT.
Yup, though in this particular case good luck suing somebody who actually manages to build a couple of hydrogen bombs. I suspect that their disputes will get settled in a somewhat different court.
Well, security updates are important, unless you plan to firewall individual systems (which is an option if you REALLY need to be running unpatched systems, but should be frowned upon and such systems should probably be limited to point-to-point VPNs across the corporate network to specific other systems). Besides, most vendors will support basic OS security patches, or at least can be talked into it.
However, all the desktop junk is a different story. You don't need to push out the latest MS Office upgrade to the server that runs your CNC mill or whatever. It probably doesn't need full-disk encryption either. Oh, and you should probably schedule those patches and not just push them out at some random time when some server is managing a pressure vessel full of explosive gases - do the updates during downtime and re-qualify the system before using it for safety-impacting operations.
At work we do provide close-to-vendor-OS images for things like this, though I'll admit in practice they aren't handled perfectly (again, the push to cut costs).
There is no reason that corporate IT can't be done well - the problem is the bottom-line mentality that aims to put the screws on any budget line item that is large, and which puts the decision-making outside of the group impacted by the decisions.
Patents document things sufficiently to mount legal battles - not so well that it makes it easier to reproduce an invention.
Claim - a method in which atomic nuclei are fused together releasing energy capable of destroying cities.
Go ahead and try to build a hydrogen bomb now...
Chemists often mine old patents looking for ways to make things, but they're usually only useful as a starting point. If somebody wants to patent some molecule they'll publish some method that creates the molecule in 0.1% yield or whatever - enough to prove by spectral observation that they made it. They don't actually publish a practical method that can be used to make it, since they don't actually want anybody to make it.
Look here for an example of this in action. The patent (#4,444,784) describes Simvastatin/Zocor - one of the more popular drugs ever taken and due to its expiry probably one of the most commonly prescribed medications around. The first method in the patent starts with 200 gallons of mold and ends up with enough compound to make about 10 pills - at one of the lower dosages. Suffice it to say this is not economical even at $5/pill.
I didn't spend enough time digging through it to figure out which of the large family of compounds described in the patent ended up being the final pill and how you'd have to make it following the patent, but this is just illustrative. The patent was really just intended to cover the molecule, and the method just proves that they indeed had made it and tested its activity.
Exclusive deals aren't about labor - they're about exploiting copyright to limit consumer choice.
The problem with ebooks isn't labor at all. Google has tons of books digitized and made available for searching at no cost to publishers at all, but publishers sue them because they don't like not being the gatekeepers.
If DC just emailed B&N a.mobi file I'm sure they'd be happy to convert it so that they can sell it.
And, ultimately, if you take anything to the extreme it often breaks down. That doesn't mean that promoting platform-neutral formats is a bad thing. We shouldn't avoid making stuff available on 99% of the hardware out there just because it won't run on a toaster.
Agreed - and I work in a corporate IT group. This sort of thing happens when you put MBAs in charge of everything - it becomes more about saving money than good operations. People blame IT usually for this sort of thing, but really this is the result of a directive to the IT manager to put cost savings above all else. The guys destroying your control systems are just following orders as a result.
If I were managing PCs across the enterprise I'd probably put them into a couple of classes: 1. Generic desktops/laptops/etc. 2. Servers 3. Systems that are primarily maintained by a vendor or some other 3rd party. 4. Systems that perform realtime operations with a safety impact, a cost impact of error/downtime > $x, etc.
Your engineering systems would probably fall into #4, unless they are fairly trivial in what they do, in which case I'd probably ask you to give serious thought to whether the costs of giving them special treatment really outweighs the reduced risk of problems. The control system for the break room coffee pot probably doesn't need mission-critical treatment.
That said, there are real benefits to EVERYBODY from standardization/etc. The problem comes when after those benefits are realized the order comes down to shave an extra 20% off each year. To me this is like every day going into your basement, finding a beam, and drilling a half-inch hole in it. Chances are you can do this for a year or more without any impact. However, eventually the house will collapse, and when it does it will be quite the thing to see as the structural failures cascade through the whole support network. In the same way when companies sabotage themselves with subtle cost-cutting across the board find that once a disaster does strike, they have no way to deal with it as EVERYBODY is short.
Claim1 - Refer to the claims of all patents that have already submitted, and add the words "on a phone" to the end. Claim2 - Refer to the claims of all patents that have already submitted, and add the words "on a computer" to the end. Claim3...n - Refer to the claims of all patents that have already submitted, and in turn add the words "on a " plus each word from every dictionary ever published in any language to the end. Claim n+1...infinity - Refer to every dictionary ever written, and combine every possible set of words used any number of times into statements and consider each as a claim.
What we really need is an open source equivalent of this. Just create a database and some FOSS software, and then everybody can go scan their old high school yearbooks and suddenly Big Brother can't do anything you can't do.
Databases owned by Government, Credit Bureaus, and Background Check companies are all coming whether you like it or not, and nobody is going to stop them.
While we're at it go ahead and create a place for people to post stuff like "person ABCD-12345 was seen at GPS coords 12.23,12,34 at timestamp..." and now everything everybody does (from your annoying neighbor to the CIA) is open for the world to see...
Yup. This isn't really anything new to IT either - just an extension of the company-buys-Schwinn and churns out $80 Walmart bikes story. To an MBA a reputation that cost a century to build is just an asset whose NPV is less than the cash you can get for pimping it on anything with two wheels.
So, all of those are good reasons that outsourcing CAN be better. They were all talked about by consultants/managers/etc publicly when it was implemented where I work.
However, I was close enough to management to find out that basically it came down to cost. Nobody really cared if service improved so much that budgets were decreased.
There are also downsides - turnover being a very big one. You're not investing in people, and you have no reason to believe that the outsource partner is either, or their idea of investment might not mean keeping a dedicated person working on things important to you.
I think the biggest driver of outsourcing is management that just doesn't want to have to get their hands dirty. You can write a check every month and dump your problems on somebody else, and as long as you have someplace to point your finger when things go south you're fine.
In general buying a warranty from a salesman is a bad move. Free ones are fine if the price is otherwise right. However, if you are buying some kind of service contract, go 3rd party as you'll pay about a third as much and be less likely to get ripped off - you went to somebody because they were a good homebuilder, not a reputable warranty company.
Oh, and always go with named exclusions, not named inclusions. It is WAY easier to find loopholes in the latter.
In any case, regulation is less necessary where there is genuine competition - say 5+ companies to choose from. Right now, many areas have one, and two isn't much better.
You're creating money out of paper - you just have to convince people that the paper is worth something.
You have millions of Americans trying to save up for retirement. They all want to retire with a million dollars in the bank. That is trillions of dollars looking for something to invest in. The real economy isn't nearly that large, so we end up creating a new economy out of paper mache.
Convince somebody to put their pension fund in a hedge fund by pointing to historic returns of 20% (which means less contributions by the employer are required - they really could care less if goes bust (it isn't THEIR pensions) as long as on paper they can show due diligence). The hedge fund makes those 20% by investing in all kinds of crazy derivatives. All the brokers involved get paid on commissions. Create a piece of paper that says it is worth a billion dollars, get somebody to buy it for $750M, and collect a few percent of that on commissions. Later it turns out the paper is worthless, but the brokers already have their pay.
Compare all that to the conventional economy. Spending $750M on a plant site means permits, EPA filings, protestors, buying land, putting up buildings, hiring people, keeping the union happy, dealing with accidents, having employes. Man, that sounds like work! And you're not going to increase its value from $750M to $1B in six months like the piece of paper promises. In fact, that plant that cost you $750M to build probably has a liquidation value of $100M if you had to sell it, assuming that by the time you do that it isn't a superfund site with negative value (the original "toxic asset").
Doing work is just, well, work. That's far too unsophisticated for the new America, where money is a cell on a spreadsheet, to be summed and multiplied at the speed of a CPU.
We don't have the money to spend on a space program or whatever - we're too busy buying invisible clothes with it.
Hate to self-reply, but here is another way of looking at it.
Your argument boils down to the law is what the government says it is.
My counterargument is that the law is what the government says it is, and can enforce.
The only reason the government gets to interpret the law is because it has the force to back it up, and so its ability to back up the law with force limits its ability to interpret the law.
In this case we're talking about use tax. The ability of the government to enforce the reporting of use tax is currently poor. That means that the average taxpayer could care less whether the government thinks they ought to pay the tax.
If you want to argue morality or logic and not purely Machiavellian principles, then I'd argue that the opinion of a court has no bearing on the matter at all.
The constitution is a set of statements. A law is a set of statements. Whether the two are contradictory is a matter of logic.
A Supreme Court ruling determines whether you spend the rest of your life rotting in jail for failure to comply with a law. However, it does not in itself make the law constitutional.:)
That is, truth is not a matter of interpretation, but in practice everything comes down to a combination of interpretation and power, and the supreme court does a lot of the former and has a lot of the latter.
This is no different than court rulings on scientific issues. A court can rule on whether something is or isn't true, but that has no bearing whatsoever on whether it is true. It has lots of bearing on whether you can get government funding to study it, or can teach it in a school, or whatever.
I think that a condition of government funding should be:
1. All research results are in the public domain and discoveries cannot be patented. 2. Researchers are required to publish all results (positive and negative) within n days of discovery. This includes sequences, coordinates, raw data, you name it. 3. Researchers must place their publications in the public domain. 4. Fail to do any of the above, and you go on the government funding blacklist for n years for the first offense, and permanently for subsequent offenses.
If I as a taxpayer pay for research, then I as a taxpayer should be able to read/use it without paying further for the privilege. And, the purpose of the funding is to improve the state of knowledge, not to give some PI information to horde to dribble out across 14 articles over 7 years.
Slashdot Mirror
The problem with mathematical models is that they tend to cause you to ignore reality in favor of the model.
There was this wonderful model for evaluating CDSes, and it said that Wall Street was made out of solid gold a few years ago.
Models also tend to encourage people to weigh decisions in favor of things that are easily measured as opposed to things that aren't easily measured. At work they're going nuts with Six Sigma, but it is all about coming up with arguments that will win elevator speeches. If you're evaluating some process you try to pick a half-dozen things at most to measure/evaluate. The problem comes that if you have a system with 500 independent variables and you pick six to optimize, then chances are you'll improve those six by slightly degrading the performance of the other 494. If you repeat that 25 times you end up with processes that aren't robust. Then when you throw in that things like "awesomeness" or "nice place to work" aren't easily measured and aren't often the top-priority of modern managers, you end up with processes that throw them out the window entirely.
Don't get me wrong - managers SHOULD know how to use models, but managers need to manage PEOPLE - not tweak formulas on spreadsheets.
Yeah, but leading an inspired R&D team is not something anybody can do. On the other hand, anybody can go down a budget and knock 10% off every line and fire the manager who misses the mark by the largest amount, then lather, rinse, repeat the next year.
The problem is the slippery slope.
A literal reading of the constitution would require a warrant to take something out of your pockets.
Now, the exception was generally made that when you're being arrested the officer is allowed to check you for weapons, check your ID, etc. That isn't unreasonable - he had to have probable cause to arrest you and checking for weapons just is being practical.
The problem is that it leads to things like - while I checked for weapons I happened to notice his gun was still hot, or that there was blood on his knife. Somehow that led to being able to fingerprint things, and then DNA, and so on. What started out as common sense will eventually turn into desurfacing the flash chips in your phone to read their contents with a STEM.
Uh, yes, well, that is the whole point of indexing - spend time when you don't care so that you save time when you do care.
So, if you're archiving list traffic, by all means filter on the headers - that is just brain-dead simple.
However, in general if your search routine struggles with 100 emails/day * 7 years, then you need a better search routine. I think that is the biggest problem with most email clients - the search indexing is pretty lousy. If Google can index half of the Internet and retrieve results in milliseconds, then my email program should be able to sort through 200k emails in less than 10 seconds.
Here's what I've found:
1. For mail that can be effectively sorted automatically, do so. I'm talking 99+% accuracy here - like filtering on list headers or something.
2. For situations where folder sorting helps with short-term projects/organization, go ahead. So, if I'm soliciting feedback on some proposal, go ahead and create a "feedback" folder that I put stuff in to go through later. Ditch the folder as soon as the task is completed. Another case of this is the high/med/low folders I create after I come back from vacation for a week or so.
3. Otherwise, just create one folder called "archive" and stuff everything into it.
The problem with organizing stuff into folders is that it costs you time for EVERYTHING you get, and it only saves you time if you go back to that particular email (and the folder helps you find it faster than a search would). What I've found is that 80% of my email is one-time-read stuff and so most of the time the effort to sort is completely wasted. Then, when I do need to go back unless it was just in the last few days I end up searching anyway, and if it is in the last few days I can usually quickly sort the archive folder. Now, folders can speed up lousy search algorithms, but the solution to that is to fix the brain-dead email software.
I do the same. At the start of this year I started using Gmail as my MUA at home, and decided to try a similar workflow at work (ugh, using Outlook). I set up an archive folder. It isn't quite the same thing, and is far more limited and much slower, but it works.
My only problem is that I tend to collect a lot of "half-done" stuff in my inbox. Periodically I tend to mass-purge it into my archive.
Now what I need is a decent search solution for outlook. Right now I'm using xobni, and it isn't bad, but it isn't great either. It is also painfully slow, but that could be the corporate IT stuff weighing it down.
My other issue is that I have a 1995ish email quota - another relic of corporate IT.
Yup, though in this particular case good luck suing somebody who actually manages to build a couple of hydrogen bombs. I suspect that their disputes will get settled in a somewhat different court.
Well, security updates are important, unless you plan to firewall individual systems (which is an option if you REALLY need to be running unpatched systems, but should be frowned upon and such systems should probably be limited to point-to-point VPNs across the corporate network to specific other systems). Besides, most vendors will support basic OS security patches, or at least can be talked into it.
However, all the desktop junk is a different story. You don't need to push out the latest MS Office upgrade to the server that runs your CNC mill or whatever. It probably doesn't need full-disk encryption either. Oh, and you should probably schedule those patches and not just push them out at some random time when some server is managing a pressure vessel full of explosive gases - do the updates during downtime and re-qualify the system before using it for safety-impacting operations.
At work we do provide close-to-vendor-OS images for things like this, though I'll admit in practice they aren't handled perfectly (again, the push to cut costs).
There is no reason that corporate IT can't be done well - the problem is the bottom-line mentality that aims to put the screws on any budget line item that is large, and which puts the decision-making outside of the group impacted by the decisions.
Patents document things sufficiently to mount legal battles - not so well that it makes it easier to reproduce an invention.
Claim - a method in which atomic nuclei are fused together releasing energy capable of destroying cities.
Go ahead and try to build a hydrogen bomb now...
Chemists often mine old patents looking for ways to make things, but they're usually only useful as a starting point. If somebody wants to patent some molecule they'll publish some method that creates the molecule in 0.1% yield or whatever - enough to prove by spectral observation that they made it. They don't actually publish a practical method that can be used to make it, since they don't actually want anybody to make it.
Look here for an example of this in action. The patent (#4,444,784) describes Simvastatin/Zocor - one of the more popular drugs ever taken and due to its expiry probably one of the most commonly prescribed medications around. The first method in the patent starts with 200 gallons of mold and ends up with enough compound to make about 10 pills - at one of the lower dosages. Suffice it to say this is not economical even at $5/pill.
I didn't spend enough time digging through it to figure out which of the large family of compounds described in the patent ended up being the final pill and how you'd have to make it following the patent, but this is just illustrative. The patent was really just intended to cover the molecule, and the method just proves that they indeed had made it and tested its activity.
Exclusive deals aren't about labor - they're about exploiting copyright to limit consumer choice.
The problem with ebooks isn't labor at all. Google has tons of books digitized and made available for searching at no cost to publishers at all, but publishers sue them because they don't like not being the gatekeepers.
If DC just emailed B&N a .mobi file I'm sure they'd be happy to convert it so that they can sell it.
And, ultimately, if you take anything to the extreme it often breaks down. That doesn't mean that promoting platform-neutral formats is a bad thing. We shouldn't avoid making stuff available on 99% of the hardware out there just because it won't run on a toaster.
Agreed - and I work in a corporate IT group. This sort of thing happens when you put MBAs in charge of everything - it becomes more about saving money than good operations. People blame IT usually for this sort of thing, but really this is the result of a directive to the IT manager to put cost savings above all else. The guys destroying your control systems are just following orders as a result.
If I were managing PCs across the enterprise I'd probably put them into a couple of classes:
1. Generic desktops/laptops/etc.
2. Servers
3. Systems that are primarily maintained by a vendor or some other 3rd party.
4. Systems that perform realtime operations with a safety impact, a cost impact of error/downtime > $x, etc.
Your engineering systems would probably fall into #4, unless they are fairly trivial in what they do, in which case I'd probably ask you to give serious thought to whether the costs of giving them special treatment really outweighs the reduced risk of problems. The control system for the break room coffee pot probably doesn't need mission-critical treatment.
That said, there are real benefits to EVERYBODY from standardization/etc. The problem comes when after those benefits are realized the order comes down to shave an extra 20% off each year. To me this is like every day going into your basement, finding a beam, and drilling a half-inch hole in it. Chances are you can do this for a year or more without any impact. However, eventually the house will collapse, and when it does it will be quite the thing to see as the structural failures cascade through the whole support network. In the same way when companies sabotage themselves with subtle cost-cutting across the board find that once a disaster does strike, they have no way to deal with it as EVERYBODY is short.
Well, considering that it was originally invented for ChromeOS I think the concept was that the browser already was the machine...
I think I'd like to submit the following patent:
Claim1 - Refer to the claims of all patents that have already submitted, and add the words "on a phone" to the end.
Claim2 - Refer to the claims of all patents that have already submitted, and add the words "on a computer" to the end.
Claim3...n - Refer to the claims of all patents that have already submitted, and in turn add the words "on a " plus each word from every dictionary ever published in any language to the end.
Claim n+1...infinity - Refer to every dictionary ever written, and combine every possible set of words used any number of times into statements and consider each as a claim.
:)
Well, with guids, and APIs, and that isn't a walled garden...
What we really need is an open source equivalent of this. Just create a database and some FOSS software, and then everybody can go scan their old high school yearbooks and suddenly Big Brother can't do anything you can't do.
Databases owned by Government, Credit Bureaus, and Background Check companies are all coming whether you like it or not, and nobody is going to stop them.
While we're at it go ahead and create a place for people to post stuff like "person ABCD-12345 was seen at GPS coords 12.23,12,34 at timestamp..." and now everything everybody does (from your annoying neighbor to the CIA) is open for the world to see...
Yup. This isn't really anything new to IT either - just an extension of the company-buys-Schwinn and churns out $80 Walmart bikes story. To an MBA a reputation that cost a century to build is just an asset whose NPV is less than the cash you can get for pimping it on anything with two wheels.
So, all of those are good reasons that outsourcing CAN be better. They were all talked about by consultants/managers/etc publicly when it was implemented where I work.
However, I was close enough to management to find out that basically it came down to cost. Nobody really cared if service improved so much that budgets were decreased.
There are also downsides - turnover being a very big one. You're not investing in people, and you have no reason to believe that the outsource partner is either, or their idea of investment might not mean keeping a dedicated person working on things important to you.
I think the biggest driver of outsourcing is management that just doesn't want to have to get their hands dirty. You can write a check every month and dump your problems on somebody else, and as long as you have someplace to point your finger when things go south you're fine.
In general buying a warranty from a salesman is a bad move. Free ones are fine if the price is otherwise right. However, if you are buying some kind of service contract, go 3rd party as you'll pay about a third as much and be less likely to get ripped off - you went to somebody because they were a good homebuilder, not a reputable warranty company.
Oh, and always go with named exclusions, not named inclusions. It is WAY easier to find loopholes in the latter.
Odd, not sure what happened to that post...
In any case, regulation is less necessary where there is genuine competition - say 5+ companies to choose from. Right now, many areas have one, and two isn't much better.
A few hundred shuttle flights? Maybe if you count flying the parts from factory to factory...
They're up to 135, with two catastrophic failures. On average they lose about a finger per flight or so it seems.
Yes, but financial risks pay off VERY quickly.
You're creating money out of paper - you just have to convince people that the paper is worth something.
You have millions of Americans trying to save up for retirement. They all want to retire with a million dollars in the bank. That is trillions of dollars looking for something to invest in. The real economy isn't nearly that large, so we end up creating a new economy out of paper mache.
Convince somebody to put their pension fund in a hedge fund by pointing to historic returns of 20% (which means less contributions by the employer are required - they really could care less if goes bust (it isn't THEIR pensions) as long as on paper they can show due diligence). The hedge fund makes those 20% by investing in all kinds of crazy derivatives. All the brokers involved get paid on commissions. Create a piece of paper that says it is worth a billion dollars, get somebody to buy it for $750M, and collect a few percent of that on commissions. Later it turns out the paper is worthless, but the brokers already have their pay.
Compare all that to the conventional economy. Spending $750M on a plant site means permits, EPA filings, protestors, buying land, putting up buildings, hiring people, keeping the union happy, dealing with accidents, having employes. Man, that sounds like work! And you're not going to increase its value from $750M to $1B in six months like the piece of paper promises. In fact, that plant that cost you $750M to build probably has a liquidation value of $100M if you had to sell it, assuming that by the time you do that it isn't a superfund site with negative value (the original "toxic asset").
Doing work is just, well, work. That's far too unsophisticated for the new America, where money is a cell on a spreadsheet, to be summed and multiplied at the speed of a CPU.
We don't have the money to spend on a space program or whatever - we're too busy buying invisible clothes with it.
That, and we're spending it on lawyers.
Hate to self-reply, but here is another way of looking at it.
Your argument boils down to the law is what the government says it is.
My counterargument is that the law is what the government says it is, and can enforce.
The only reason the government gets to interpret the law is because it has the force to back it up, and so its ability to back up the law with force limits its ability to interpret the law.
In this case we're talking about use tax. The ability of the government to enforce the reporting of use tax is currently poor. That means that the average taxpayer could care less whether the government thinks they ought to pay the tax.
If you want to argue morality or logic and not purely Machiavellian principles, then I'd argue that the opinion of a court has no bearing on the matter at all.
The constitution is a set of statements. A law is a set of statements. Whether the two are contradictory is a matter of logic.
A Supreme Court ruling determines whether you spend the rest of your life rotting in jail for failure to comply with a law. However, it does not in itself make the law constitutional. :)
That is, truth is not a matter of interpretation, but in practice everything comes down to a combination of interpretation and power, and the supreme court does a lot of the former and has a lot of the latter.
This is no different than court rulings on scientific issues. A court can rule on whether something is or isn't true, but that has no bearing whatsoever on whether it is true. It has lots of bearing on whether you can get government funding to study it, or can teach it in a school, or whatever.
I think that a condition of government funding should be:
1. All research results are in the public domain and discoveries cannot be patented.
2. Researchers are required to publish all results (positive and negative) within n days of discovery. This includes sequences, coordinates, raw data, you name it.
3. Researchers must place their publications in the public domain.
4. Fail to do any of the above, and you go on the government funding blacklist for n years for the first offense, and permanently for subsequent offenses.
If I as a taxpayer pay for research, then I as a taxpayer should be able to read/use it without paying further for the privilege. And, the purpose of the funding is to improve the state of knowledge, not to give some PI information to horde to dribble out across 14 articles over 7 years.