Let me present both sides - and then say I while have shot weddings - only for friends and gifted them the negatives - I write software used by pros.
This view of retaining negatives and copyrights is a stressed tradition and is shifting.
One of the reasons photographers are so expensive up front is the growing expectation that the back-end purchases will be shorted by reproduction.
But the essense of the artistic reason is much the same as a chef in a kitchen - he wants to retain the right to prepare in secret and to toss the rotton egg and the sunken souffle while keeping up a shining - everything-is-perfect mirage.
That is the world of art.
How many olp painting did picasso simply paint over because he wasn't happy with the result.
Now you with your tech-savvy computer graphics computer want to come back into the kitchen with the raw materials, and cook up whatever you like, and then present it to your friend like some kind of original. Who's responsible for the result?
It get's tricky.
Photography is a process that hardly ends when the trigger is fired - and the more professional the artist - the more this is true.
All that said - I believe the industry is quickly moving towards charging once for the copyright of an image and allowing you to use or print it as you like - with - in some cases the hope that you would see value in the print service provided under the same roof.
A Website can be cached and forwarded without copyright issues - much of the web is locally cached at some intersection in any event - and all of it is buffered and forwarded.
however - to properly service the advertisers - a cahing service would serve a frequently refreshed copy - count the copies, and then over time - request page views to correspond to the copies served.
There are many things we can create a blackbox algorithm which represents the first principles as good of better than algorithms based on a full understanding of the first principles.
Example: At a division of Honeywell, we worked on "flattening the response curve" of the crt.
to do this the engineers developed complicated models based on energy output, gamma curves, crystal variations, etc and used these models to sense the point to point deviations across the crt and create a LUT to modulate the signal for flattest response.
Using first principles they were able to slowly converge on a resonable solution.
however - using historical data of solutions to prior deviations and their corresponding LUT value - it was possible to more quickly converge the tubes than using the first principle calculations.
While it wasn't exactly a GA - it was accumulated knowledge reapplied to a problem - rather than complex equations - probably of the kind I couldn't understand.
In my experience as a not physicist - I can solve problems related to physicists using models and other calculation more akin to brute force.
These are the kind of Magazines that get their ad spreads by fax.
Yeah I know this breed.
Probably sells commodities to rednecks.
Serious printers understand what it takes to get the most from their press. and they probably don't own a fax machine or a 2mp camera - certainly not as part of the copy generating process.
Johnny come lately - anyone can buy a mac and be in the printing busines types - the kind that think design is best accomplshed by making sure no two consecutive phrases share the same font (or color)
This is your market for 3MP publication quality cameras.
1. parsing for terms 2. determining the uniqueness of each term 3. select terms which are
a. used in about half the documents
b. not used in the other half
The Historic problem with your statement is that such a filing system requires an a priori understanding of the problem space = that is - of the complete set of documents.
Before you can organize the animal kingdom - you must look at all the known animals - compare their differences with their similarilties and then establish some fitness function for determining whether a given gene expression denotes speciation or mere variation.
That's fine - if you know what the set of animals is - but it fails miserably going forward.
Say we organize all the companies using Linux in the future, but we have no idea what those companies will be, what they will use or do, where they will operate - in short we don't know much about the problem set - so we don't organize a priori - we merely heap.
The ability to serve data in real-time categories is the solution to filing.
And the fitness function is intended to subdivide the data heap into a tree such that each bifurcation presents two options with nearly equal number of sub nodes.
There are two kinds of customers (West coast and east cooast) of the west customer there are two types, medical and legal,
Etc...
Note the sub groups aare not mutually exclusive and nodes may appears more than one for example:
The West coast branch may include national transportation companies with operations on both coasts,
The legal option may include cases for medical malpractice.
It is up to the fitness function to choose the next logical question in such a way that the number of ambiguous nodes is minimized.
The curious difference here - is that there is no empirical classification. each choice creates a new top level clasiification of the remaining nodes which is optimized only for that moment in time and that particular history of choices.
If ambiguous answers are allowed - then the algorithm becomes truly non-deterministic with respect to the absolute order of next questions.
enhance that with the option to have more than two choices and you approach a completely browsable AND optimized real-time filing system.
My point is that medicine should increase its use of simple walk around data input tools.
Doctor's need to read info - fine - but Nurses generally need to collect the data Doctors use. a $120 scanner can collect a ton of accurate timestamped unforgeable data quite quickly.
Assume for sake of argument, that the documents are word files, and that the user is an accountant.
Analysis of the files reveals the following statistics,
the phrase" McGillicuttys Stormdoor and undertaking" is used in 10% of the documents while "Mrs Greencows Dry Cleaning" appears in 5% of the documents and both appear in only 1% of the documents.
The software can then make three piles - called McGillicutty, Mrs, Greencow, and Summaries
Within each set of documenst - it is found that the phrase 2004 appears a great deal in some docs while 2003 appears in others - mostly exclusively - so the next level of distinction reads 2003, 2004
At every level - the choice is statistically optimize to provide the very best indicator - I have guessed that customer name and year are good divisions - but I guess - statistically other metrics might be better.
It isn't hard to devise a fitness function for how well a term serves to identify a given group of documents with respect to the other documents heaps at a given point x.
Properly implemented - I think even doctors would find a portable data collector useful. (But I said medicine - meaning the whole group)
But yes - certainly nurses.
I believe the difference between fast and accurate data collection
vs.
PDA's which are slow, tedious and require comprehension - which is the opportunity for error.
Scanners require no user comprehension - when we read data - we scan the words against the words we recognize. One latin word may be closer to a different but more familiar word than we realize, and a mistake is made.
A great deal of medicine involves watching the patient react to treatments - logging temperature, meds, foods, pisses, wakes, and sleeps can lead to better medicine at lower cost - and a scanner is the tool for this. a PDA is not.
To input - patient A is asleep - 12:13 pm into a PDA is a serious task, but a quick swipe of the patients chart and a second pass at a simple vital checklist and the data is collected and timestamped.
Add a digital camera to your bag of tricks, and you can prove and record (against the patients scanned chart and clock) almost any metric empirically - click - the thermometer says 103 degrees. click - patient is eating.
A good deal of medicine is proving you haven't screwed up.
You've touched it really - these things need to solve a real 9 to 5 problem. And for doctors - probably because doctors don't bring their desk to the bedside - the PDA is the right thing.
But hospitals have invested heavily in the WIFI infrastructure to accomidate a PDA in the hostpital almost like a cellphone in the outside world.
That said - i think the real tool for medicine is a simple barcode reader - scanning patient tags, medicine bottles, blood pressure readings, and a range of tests into a portable scanner - allows the objective facts of medicine to be collected in real time and consolidated without double entry.
You mention cross checking medicines - that can be done by scanning the patients current collection of pill bottles - and the recommended new entry, the patient id and - viola - a printed drug cross check.
(For others - This is the Best Practice analysis - the rest of you morons can drown in your "Outsourcing isn't as bad as that rhetoric"
To further this discussion.
What happens when there are negative externalities inflicted into the production system.
For example in the US - we inflict the cost of healthcare, retirement, clean air, worker safety, paternal leave, and the bulk of taxation on those who Produce goods whether they are sold here or not - while we inflict importers with little more than the obligation to say a few nice words about the environment and they are free to sell into a "hands off" consumption market.
If the United State would merely shift the external social costs (tax, retirement, healthcare) to forms of taxation which impacted foriegn goods as well as domestic production - then much of the rest of these comments would bear out - we could benefit by trade, and compete in those global markets in which we excel.
The Statement is only true in the current context.
I happen to believe that soon we will share our wireless ports in a harmonic fashion - and that will permit us to get service beyond current base radius.
Add to this the possibility that personal wifi phones could also be mesh points and you have the makings of a personal beehive of connectability.
ironically - the power requirements of repeating signals - as opposed to making single long hops suggests that sharing your service with others actually saves you bandwidth and power.
(assuming random locations and equal useage)
So vowifi is a choice technology for the sharing type - and will probably be a badge of community before long.
the only requirement is a hackable model with the right form and feature set. This has happened already with the linux linksys wifi router - a hackable linux wifi phone could spark a new craze of peacedriving - wandering around looking for others who WANT you on their link.
the dynamics should look a lot link ametuer radio without the damn keying requirement.
I think the factor that most compromises education is the single generation population expansion unadjusted for infant mortality. In short the total number of kids per adult.
Since only a fraction of adults can teach - that fraction must be divided among n number of children per available teacher. n is a factor which represents the average hours of teaching available per child. You can then allocate that as you like - but you cannot wish it away. Some will get more - other less.
As a thought experimenet - I would like to consider giving one child from every family the allocation required to go all the way to the top. This means a doctor is every pot - so to speak.
It is very fair. it is immune to historical prejudices - without being counterracial.
And it avoids the trap of welfare which is to perpetuate poverty by encouraging professionals to avoid children (by taxing them into stress) while encouraging overutilized parents to add more kids to a situation in which the kids are already self-parenting.
Social policy should strive to provide one parent for every child.
Sorry - I don't mean to get into a pissing match here.
It really doesn't matter to me - but I believe there are - yes consumer - experiences which now seem the domain of WIKA digital - but like the betamax - which seemed a magic wetdream when it was first built - is not a subpar $68 dollar lost leader in grocery stores. The HDTV recorder - which seems at present to be a wetdream - must surely be considered mainstream in the very near future.
Consumers tolerate the current level - but they do want better technologies.
Look at digital cameras - 5 years ago - 2MP for $5,000 (polaroid) seemed like a fun idea - today 2MP sony is $179. And $5k will get you 11MP or more.
11MP is the equivelent of a 35mm camera - that is the resolution owned by a majority of photobuffs going in.
Why would you expect 35mm owners not to expect their digital cameras to match film?
underestimating consumer desire got the APS standard into trouble.
Bare bones experience - sure - you can move resampling to an MMU
You could as you say - build a proper Image Processing Unit as a seperate chip but the economics of general purpose computers suggests that a lot will be done by the processor.
But what happens when you want a better experience - most computers for example cannot reproduce a simple slide show with a smooth fade from one image to the next in full screen resolution - this is a CPU bound problem - and very typical consumer expectation.
I think you underestimate what it takes to handle high quality images.
Images which are multi-purpose must also be of a far higher resolution than current screens - so the goal of merely showing images on a device is insuffecient.
I'm sure that you have earned the distinction - however - may I suggest fundamentally - that poor kids who get pHDs are noise in the data for a sound reason.
That education is an empirical luxury.
That poor societies are limited in the amount of life-hours they can devote to education - both for teacher hours and for student hours.
That the distribution of life-hours for education worldwide is closely correlated with the wealth of a society relative to its trading partners.
Therefore: pHDS exists only where there is sufficient economic advantage given to a select group of people with respect to another group which are made to do the heavy lifting.
Under Malthusian economics - one group can be rich - only at the expense of another. And because education is an expression of relative wealth at either a personal level or an onclave level - those who are educated ARE in fact educated at the expense of the others.
This is but one reason which suggests the educated have a debt to honor with repect to the world at large - but I ramble.
None of this is to suggest having wealth or a pHD is morally wrong - but there is a degree of moral responsability attached to priveledge. In short Ann Rand and Emerson both had it wrong - we are not independant agents - Thomas More had it right - no man is an island. (I suggest)
Slashdot Mirror
Let me present both sides - and then say I while have shot weddings - only for friends and gifted them the negatives - I write software used by pros.
This view of retaining negatives and copyrights is a stressed tradition and is shifting.
One of the reasons photographers are so expensive up front is the growing expectation that the back-end purchases will be shorted by reproduction.
But the essense of the artistic reason is much the same as a chef in a kitchen - he wants to retain the right to prepare in secret and to toss the rotton egg and the sunken souffle while keeping up a shining - everything-is-perfect mirage.
That is the world of art.
How many olp painting did picasso simply paint over because he wasn't happy with the result.
Now you with your tech-savvy computer graphics computer want to come back into the kitchen with the raw materials, and cook up whatever you like, and then present it to your friend like some kind of original. Who's responsible for the result?
It get's tricky.
Photography is a process that hardly ends when the trigger is fired - and the more professional the artist - the more this is true.
All that said - I believe the industry is quickly moving towards charging once for the copyright of an image and allowing you to use or print it as you like - with - in some cases the hope that you would see value in the print service provided under the same roof.
AIK
A Website can be cached and forwarded without copyright issues - much of the web is locally cached at some intersection in any event - and all of it is buffered and forwarded.
however - to properly service the advertisers - a cahing service would serve a frequently refreshed copy - count the copies, and then over time - request page views to correspond to the copies served.
AIK
Can't beat Free
Check Panera Bread . com foor location - most metro areas are served - show Starbucks the Cyber door!
This comment resounds with me.
There are many things we can create a blackbox algorithm which represents the first principles as good of better than algorithms based on a full understanding of the first principles.
Example:
At a division of Honeywell, we worked on "flattening the response curve" of the crt.
to do this the engineers developed complicated models based on energy output, gamma curves, crystal variations, etc and used these models to sense the point to point deviations across the crt and create a LUT to modulate the signal for flattest response.
Using first principles they were able to slowly converge on a resonable solution.
however - using historical data of solutions to prior deviations and their corresponding LUT value - it was possible to more quickly converge the tubes than using the first principle calculations.
While it wasn't exactly a GA - it was accumulated knowledge reapplied to a problem - rather than complex equations - probably of the kind I couldn't understand.
In my experience as a not physicist - I can solve problems related to physicists using models and other calculation more akin to brute force.
Saved about 8 years and 100K tuition - cool.
AIK
heavy or fast.
fast is good.
Actually this flywheel ups is already being used or explored.
You don't need a great deal of weight just to convert voltage.
I've seen jobsite - transformers - which connect to a 12 car battery - turn a motor - which in turn turns a generator to power - say a skill saw.
reverse this - raise the speed - add some weight for 5 minutes of safe shutdown capabiity and you approach 98% effeciency.
Since you need the spinning wheel anyway - you can build this power supply with minimal extra parts.
AIK
You made me think of something.
Since most computer need a harddrive.
and power can be stored in something very much like a harddrive - ie a spinning disk,
and because spinning is generally a very very clean way to transfer power,
it might not be a stretch to combine the functions of a hard drive, power supply and ups into a single spinning mass.
This is an improvement because it eliminates lead acid batteries in the environment.
creates clean electricity for the computer,
Generators can be highly effecient (99%)
AIK
My coowore picked up that camera for closer to 700
(since he had lenses for canon - no lens.
That puts the digital SLR surprising close to the high end consumer stock.
Very appealling.
AIK
fixed lenses also solve a major problem in digital cameras
dust collects on the imager.
Think about it.
Film is constantly fresh out of the can.
The CCD on the other hand - just sits there and wahtever parks on its surface stays there and creates obstructions until its cleaned.
Cleaning risks the entire camera - since imagers are also easy to crack.
So keeping the chamber sealed with a fixed lens - and attachments to alter range - has an unanticiated benefit.
AIK
These are the kind of Magazines that get their ad spreads by fax.
Yeah I know this breed.
Probably sells commodities to rednecks.
Serious printers understand what it takes to get the most from their press. and they probably don't own a fax machine or a 2mp camera - certainly not as part of the copy generating process.
Johnny come lately - anyone can buy a mac and be in the printing busines types - the kind that think design is best accomplshed by making sure no two consecutive phrases share the same font (or color)
This is your market for 3MP publication quality cameras.
The good news is - its a big market.
AIK
Its done by
1. parsing for terms
2. determining the uniqueness of each term
3. select terms which are
a. used in about half the documents
b. not used in the other half
Use that term to define the half
divide the heap and run 1, 2, 3 on both halves.
AIk
The Historic problem with your statement is that such a filing system requires an a priori understanding of the problem space = that is - of the complete set of documents.
...
Before you can organize the animal kingdom - you must look at all the known animals - compare their differences with their similarilties and then establish some fitness function for determining whether a given gene expression denotes speciation or mere variation.
That's fine - if you know what the set of animals is - but it fails miserably going forward.
Say we organize all the companies using Linux in the future, but we have no idea what those companies will be, what they will use or do, where they will operate - in short we don't know much about the problem set - so we don't organize a priori - we merely heap.
The ability to serve data in real-time categories is the solution to filing.
And the fitness function is intended to subdivide the data heap into a tree such that each bifurcation presents two options with nearly equal number of sub nodes.
There are two kinds of customers (West coast and east cooast)
of the west customer there are two types, medical and legal,
Etc
Note the sub groups aare not mutually exclusive and nodes may appears more than one for example:
The West coast branch may include national transportation companies with operations on both coasts,
The legal option may include cases for medical malpractice.
It is up to the fitness function to choose the next logical question in such a way that the number of ambiguous nodes is minimized.
The curious difference here - is that there is no empirical classification. each choice creates a new top level clasiification of the remaining nodes which is optimized only for that moment in time and that particular history of choices.
If ambiguous answers are allowed - then the algorithm becomes truly non-deterministic with respect to the absolute order of next questions.
enhance that with the option to have more than two choices and you approach a completely browsable AND optimized real-time filing system.
AIK
My point is that medicine should increase its use of simple walk around data input tools.
Doctor's need to read info - fine - but Nurses generally need to collect the data Doctors use. a $120 scanner can collect a ton of accurate timestamped unforgeable data quite quickly.
AIK
Assume for sake of argument, that the documents are word files, and that the user is an accountant.
Analysis of the files reveals the following statistics,
the phrase" McGillicuttys Stormdoor and undertaking" is used in 10% of the documents while "Mrs Greencows Dry Cleaning" appears in 5% of the documents and both appear in only 1% of the documents.
The software can then make three piles - called McGillicutty, Mrs, Greencow, and Summaries
Within each set of documenst - it is found that the phrase 2004 appears a great deal in some docs while 2003 appears in others - mostly exclusively - so the next level of distinction reads 2003, 2004
At every level - the choice is statistically optimize to provide the very best indicator - I have guessed that customer name and year are good divisions - but I guess - statistically other metrics might be better.
It isn't hard to devise a fitness function for how well a term serves to identify a given group of documents with respect to the other documents heaps at a given point x.
AIK
No - you can determine the smartest way to organize any given pile of data.
.
You organize it based on its contents such that the number of choices required to find the data is minimal and optimal.
For this you need dynamic "classification"
classification identifies words as either unique or ubiquitous. Words which are neither uique or ubiquitous can be used for classification.
by finding the words best suited to sub-divide the given files - you create order on the fly.
This system allows nonexclusive groupings to form by shared content.
If you know what you are looking for, a few multiple choices can locate any document.
- note - for pictures and non-documents this is restricted to filenames, dates, and foldernames etc . .
AIK
Properly implemented - I think even doctors would find a portable data collector useful. (But I said medicine - meaning the whole group)
But yes - certainly nurses.
I believe the difference between fast and accurate data collection
vs.
PDA's which are slow, tedious and require comprehension - which is the opportunity for error.
Scanners require no user comprehension - when we read data - we scan the words against the words we recognize. One latin word may be closer to a different but more familiar word than we realize, and a mistake is made.
A great deal of medicine involves watching the patient react to treatments - logging temperature, meds, foods, pisses, wakes, and sleeps can lead to better medicine at lower cost - and a scanner is the tool for this. a PDA is not.
To input - patient A is asleep - 12:13 pm into a PDA is a serious task, but a quick swipe of the patients chart and a second pass at a simple vital checklist and the data is collected and timestamped.
Add a digital camera to your bag of tricks, and you can prove and record (against the patients scanned chart and clock) almost any metric empirically - click - the thermometer says 103 degrees. click - patient is eating.
A good deal of medicine is proving you haven't screwed up.
AIK
You've touched it really - these things need to solve a real 9 to 5 problem. And for doctors - probably because doctors don't bring their desk to the bedside - the PDA is the right thing.
But hospitals have invested heavily in the WIFI infrastructure to accomidate a PDA in the hostpital almost like a cellphone in the outside world.
That said - i think the real tool for medicine is a simple barcode reader - scanning patient tags, medicine bottles, blood pressure readings, and a range of tests into a portable scanner - allows the objective facts of medicine to be collected in real time and consolidated without double entry.
You mention cross checking medicines - that can be done by scanning the patients current collection of pill bottles - and the recommended new entry, the patient id and - viola - a printed drug cross check.
AIK
She got a free base?
There is a plan to harness wind energy using inexpensive kites.
That and higher effeciency computers (Small LCD Screens?) could provide free electricity and access to information.
Why does every computer user have to have a full screen full color experience.
A monocolor non-lit LCD screen is adequate - even if it isn't executive quality.
AIK
(For others - This is the Best Practice analysis - the rest of you morons can drown in your "Outsourcing isn't as bad as that rhetoric"
To further this discussion.
What happens when there are negative externalities inflicted into the production system.
For example in the US - we inflict the cost of healthcare, retirement, clean air, worker safety, paternal leave, and the bulk of taxation on those who Produce goods whether they are sold here or not - while we inflict importers with little more than the obligation to say a few nice words about the environment and they are free to sell into a "hands off" consumption market.
If the United State would merely shift the external social costs (tax, retirement, healthcare) to forms of taxation which impacted foriegn goods as well as domestic production - then much of the rest of these comments would bear out - we could benefit by trade, and compete in those global markets in which we excel.
AIK
The Statement is only true in the current context.
I happen to believe that soon we will share our wireless ports in a harmonic fashion - and that will permit us to get service beyond current base radius.
Add to this the possibility that personal wifi phones could also be mesh points and you have the makings of a personal beehive of connectability.
ironically - the power requirements of repeating signals - as opposed to making single long hops suggests that sharing your service with others actually saves you bandwidth and power.
(assuming random locations and equal useage)
So vowifi is a choice technology for the sharing type - and will probably be a badge of community before long.
the only requirement is a hackable model with the right form and feature set. This has happened already with the linux linksys wifi router - a hackable linux wifi phone could spark a new craze of peacedriving - wandering around looking for others who WANT you on their link.
the dynamics should look a lot link ametuer radio without the damn keying requirement.
AIK
I think the factor that most compromises education is the single generation population expansion unadjusted for infant mortality. In short the total number of kids per adult.
Since only a fraction of adults can teach - that fraction must be divided among n number of children per available teacher. n is a factor which represents the average hours of teaching available per child. You can then allocate that as you like - but you cannot wish it away. Some will get more - other less.
As a thought experimenet - I would like to consider giving one child from every family the allocation required to go all the way to the top. This means a doctor is every pot - so to speak.
It is very fair. it is immune to historical prejudices - without being counterracial.
And it avoids the trap of welfare which is to perpetuate poverty by encouraging professionals to avoid children (by taxing them into stress) while encouraging overutilized parents to add more kids to a situation in which the kids are already self-parenting.
Social policy should strive to provide one parent for every child.
AIK
Sorry - I don't mean to get into a pissing match here.
It really doesn't matter to me - but I believe there are - yes consumer - experiences which now seem the domain of WIKA digital - but like the betamax - which seemed a magic wetdream when it was first built - is not a subpar $68 dollar lost leader in grocery stores. The HDTV recorder - which seems at present to be a wetdream - must surely be considered mainstream in the very near future.
Consumers tolerate the current level - but they do want better technologies.
Look at digital cameras - 5 years ago - 2MP for $5,000 (polaroid) seemed like a fun idea - today 2MP sony is $179. And $5k will get you 11MP or more.
11MP is the equivelent of a 35mm camera - that is the resolution owned by a majority of photobuffs going in.
Why would you expect 35mm owners not to expect their digital cameras to match film?
underestimating consumer desire got the APS standard into trouble.
AIK
That may seem true without being true.
Bare bones experience - sure - you can move resampling to an MMU
You could as you say - build a proper Image Processing Unit as a seperate chip but the economics of general purpose computers suggests that a lot will be done by the processor.
But what happens when you want a better experience - most computers for example cannot reproduce a simple slide show with a smooth fade from one image to the next in full screen resolution - this is a CPU bound problem - and very typical consumer expectation.
I think you underestimate what it takes to handle high quality images.
Images which are multi-purpose must also be of a far higher resolution than current screens - so the goal of merely showing images on a device is insuffecient.
AIK
The hottest consumer electronic product right now may well be the digital camera.
Given that it takes a lot more cycles to process an image - than say - a spreadsheet - I think the consumer very much wants processing speed.
The next hot item will be likely be digital videos actually going somewhere - other than the shoebox
This will be even more cycle intensive.
The market responds to technology in fits and starts - but the analysis says many consumer products are still throttled by the speed on consumer PC's
People buying HDTV today will want HDTV camcorders tomorrow.
A computer that can render HDTV signals from disk will be a benchmark standard in only a few years.
The idea that spreadsheet surfers are satisfied with clock speed is quiet impressive - and passe.
AIK
I'm sure that you have earned the distinction - however - may I suggest fundamentally - that poor kids who get pHDs are noise in the data for a sound reason.
That education is an empirical luxury.
That poor societies are limited in the amount of life-hours they can devote to education - both for teacher hours and for student hours.
That the distribution of life-hours for education worldwide is closely correlated with the wealth of a society relative to its trading partners.
Therefore:
pHDS exists only where there is sufficient economic advantage given to a select group of people with respect to another group which are made to do the heavy lifting.
Under Malthusian economics - one group can be rich - only at the expense of another. And because education is an expression of relative wealth at either a personal level or an onclave level - those who are educated ARE in fact educated at the expense of the others.
This is but one reason which suggests the educated have a debt to honor with repect to the world at large - but I ramble.
None of this is to suggest having wealth or a pHD is morally wrong - but there is a degree of moral responsability attached to priveledge. In short Ann Rand and Emerson both had it wrong - we are not independant agents - Thomas More had it right - no man is an island. (I suggest)
AIK