When a paramedic arrives at an accident and the injured person is purple, he still tries to resuscitate,
When a paramedic arrives and find a patient with a severed limb does he try to convince it to stop bleeding or get out a tourniquet. Chamberland and The League of Nations did nothing to reign in Hitler and Europe paid the price.
I think the thing you missed is that publishing is not the same as hosting. Who decides which document goes on which computer? What happens if one or more of the small repositories goes down? Who get the editing and peer reviews done? These small repositories are not going to provide any customer support.
Who uses USENET now? There might be a reason. If is is so simple and easy, why hasn't it been done by now? Perhaps it is not as simple as you think.
Both are funded by donations and grants. Both have a much bigger audience than scientific journals and therefore a much bigger pool of possible donors to draw from. Going to a donation model would in effect be a voluntary subscription fee that may or may not cover costs. .
Here are some issues; Bigger government. Conservatives would call that another service that the Government should not be involved in. Government bias. There will be many people stating that government journals will favor government positions and would loose independence. All papers are not government funded. How do you recover costs from publishing the non-funded papers?
The extra processing would add a tiny amount. The editing and the reviewing can be obtained from the same place that the present day journals obtain them: the science community, and at the same cost: $0.
I guess you have never dealt with getting papers reviewed and edited by several people. The papers have to be sent out, people reminded, multiple edits reconciled, and all this may take several iterations. Even if the actual editing and reviewing is free the administration costs money.
Instead of looking at one aspect and saying that sisnce it does not cost anything therefore the whole process costs nothing try finding where costs could come in. Several small costs add up into a much bigger cost
Search warrants stipulate what the authorities are looking for and where they can look; not the tools they can use to get the job done. Do wiretap warrants stipulate the kind of recording devices that can be used? I doubt it very much.
There is still the point at to whether the order covers the police. I might be argued that the authorities were working as an agent for Verison to gather the information.
By the way the projected Xarciv operating costs for 2013-2017 are projected to average of $826,000 per year, including indirect expenses. It is free to the users but the costs are paid for by donations. They also don't edit, peer review or produce journals. Publishing an edited, peer reviewed paper in a journal is much more that making it available on e web site.
What Does It Actually Cost To Publish a Scientific Paper?
There are several questions that may get swept up in this debate; How much does it cost to publish an externally edited scientific paper? How much does it cost to publish a peer reviewed scientific paper? How much does it cost to publish a scientific journal? How much does it cost to publish an externally edited, peer reviewed scientific paper? How much does it cost to publish an externally edited, peer reviewed scientific paper in a scientific journal?
All of these have different costs. ArXiv is an e-print repository funded by donations with an operating costs for 2013-2017 are projected to average of $826,000 per year, including indirect expenses. They are not editors, peer reviewers or journals. In effect they are the entry level in scientific paper publishing and they have significant expenses. Even if peer reviewers and editors are not paid there are still significant support staff needed to shuffle the documents around and maintain the servers, hardware cost, bandwidth costs, insurance costs, customer service costs, etc. The cost of publishing is non-zero and adding editing, peer reviews and journals adds to the cost. Someone has to pay for it and the question is whom.
ArXiv is an e-print archive. It does not appear that they do any peer reviews or editing. It would seem that hosting journals would be even more expensive. It look more and more like the "it's cheap enough for anyone to take over" crowd is way off.
There is so much talk about how cost are minuscule and any reasonably sized institute could bear the load. If that was true then why has it not been done already? I am sure most institutes would love to get rid of the costs of journal subscriptions. Perhaps it is not as easy or low cost as some people think.
It seems reasonable that a publisher would have to recover costs and make a profit. If they can not recover it from subscription the only other choice is to charge contributors. Publishers are not charities. According to this annual report Taylor & Francis' parent compant made a 27% profit in the Academic Information sector and 7% overall. Without that cash cow the company is not viable.
If the mail outlet that initially processed the packages put them into bags there is a high likelihood that there are bags that contain only marked or unmarked packages. The difference is it is a loss of one bag not ten packages. The loss of one bag can easily be explained by fluke. That is why I say this is a poorly designed experiment. Had they thought more about the failure possibilities than trying to prove a point they may have come up with a better plan.
The thing you keep missing is that you keep treating each separate box as a shipment with it's own sorting and tracking. That is not what happens. If a outlet in Germany get a number of packages that are going to the US they will put them in a bag or shipping cage. Now that 89 shipments become one to ten amalgamated items. If one of those amalgamated items get delayed there is a significant impact on the statistics.
In this situation we have some method that when 50 branded (A) boxed and 50 unbranded (B) boxes passed through it we saw (among other things) 1 B box get lost and 10 A boxes get lost. that is significantly unlikely to be by chance.,
If somewhere along the line those 100 boxes are put into ten bags and one of those bags are lost then the statistical significance is gone.
When mail is sent from point A to point B it goes through a number of hubs where the mail is either amalgamated or dispersed. Have you seen a tracking display for a parcel? You will notice that there are several stops along the way. Moving between these hubs can take several trucks which arrive and leave at different times. If the truck that happened to have the marked packages on it was delayed there would be a significant difference. Had they done a few packages a day over a few weeks I would be more inclined to believe it.
Almost forgot. Just because they were handed to Deutsch Post at the same time does not mean they were processed at the same time. It takes time to process packages and it is possible that the truck was full part way through the batch. Without looking at the tracking data all we can do is guess.
I can't seem to find the source of your quote. What document is it referencing?
My point is that if there are other possibilities then the hypotheses is not proven.
Even if many of the packages arrived at the same time there could be other groups of packages that missed the cut off time for the hub. I would like to see the distribution of the delays and losses. There may be a pattern.
They sent 86 packages at the same time. This shipment is broken up along the way so the N is somewhere between 1 an 86. Had they sent the packages on different days then the N would be more accurate.
The step they did not take was to try to explain why the significant difference happened. They did not look at the tracking data to see how or where the packages were lost. They did not repeat the experiment to confirm that this one multipart shipment was not a fluke.
It could be argued that the N is actually somewhere between 2 and 178 as at different points in the route some packages are grouped together and a issue local issue could effect more than one package. I would have been much more convinced had they sent a packages of each type per day over a period of weeks so that a single anomaly, such as the loss of one mail bag, could not skew the statistics.
This whole study is based on a false assumption; that packages sent on the same day will arrive at the same time. A package sent before the cut off time will arrive on a different day than a package sent after the cut off time. A package that will not fit on a full plane will wait for the next plane, A package that will not fit on a full truck will wait for the next truck. Were one set of packages mostly delivered on Saturday and the others Monday? There is two days difference. Are they talking calendar days or business days?
No medical study that has not been peer reviewed is taken as proof so why should this one be. There is an old saying "lies, damn lies and statistics".
Only if both batches made the cut off time for that day. One group of packages could have gone out one day and the other group may have gone out the next day.
Actually two packages sent as little as half an hour apart could arrive on different days. If one package made the cut off time and the other one did not one package would be on tonight's truck and the other one would wait for the next one. I did a test on the USPS site and, depending on the service requested the cut off times varied from 2:15PM to 4:30PM.
The problem with this "study" is that it makes a lot of assumptions on what is happening without even looking at the tracking data.
This is only one sample and not repeated, therefore not proof.
What I would like to see is the distribution of the missing parcels. I would check the tracking on the missing packages and see if they were on the same truck. If they were on the same truck did that truck have an accident?
Slashdot Mirror
When a paramedic arrives at an accident and the injured person is purple, he still tries to resuscitate,
When a paramedic arrives and find a patient with a severed limb does he try to convince it to stop bleeding or get out a tourniquet. Chamberland and The League of Nations did nothing to reign in Hitler and Europe paid the price.
I think the thing you missed is that publishing is not the same as hosting. Who decides which document goes on which computer? What happens if one or more of the small repositories goes down? Who get the editing and peer reviews done? These small repositories are not going to provide any customer support.
Who uses USENET now? There might be a reason. If is is so simple and easy, why hasn't it been done by now? Perhaps it is not as simple as you think.
Both are funded by donations and grants. Both have a much bigger audience than scientific journals and therefore a much bigger pool of possible donors to draw from. Going to a donation model would in effect be a voluntary subscription fee that may or may not cover costs. .
Here are some issues;
Bigger government. Conservatives would call that another service that the Government should not be involved in.
Government bias. There will be many people stating that government journals will favor government positions and would loose independence.
All papers are not government funded. How do you recover costs from publishing the non-funded papers?
The extra processing would add a tiny amount. The editing and the reviewing can be obtained from the same place that the present day journals obtain them: the science community, and at the same cost: $0.
I guess you have never dealt with getting papers reviewed and edited by several people. The papers have to be sent out, people reminded, multiple edits reconciled, and all this may take several iterations. Even if the actual editing and reviewing is free the administration costs money.
Instead of looking at one aspect and saying that sisnce it does not cost anything therefore the whole process costs nothing try finding where costs could come in. Several small costs add up into a much bigger cost
Search warrants stipulate what the authorities are looking for and where they can look; not the tools they can use to get the job done. Do wiretap warrants stipulate the kind of recording devices that can be used? I doubt it very much.
There is still the point at to whether the order covers the police. I might be argued that the authorities were working as an agent for Verison to gather the information.
By the way the projected Xarciv operating costs for 2013-2017 are projected to average of $826,000 per year, including indirect expenses. It is free to the users but the costs are paid for by donations. They also don't edit, peer review or produce journals. Publishing an edited, peer reviewed paper in a journal is much more that making it available on e web site.
What Does It Actually Cost To Publish a Scientific Paper?
There are several questions that may get swept up in this debate;
How much does it cost to publish an externally edited scientific paper?
How much does it cost to publish a peer reviewed scientific paper?
How much does it cost to publish a scientific journal?
How much does it cost to publish an externally edited, peer reviewed scientific paper?
How much does it cost to publish an externally edited, peer reviewed scientific paper in a scientific journal?
All of these have different costs. ArXiv is an e-print repository funded by donations with an operating costs for 2013-2017 are projected to average of $826,000 per year, including indirect expenses. They are not editors, peer reviewers or journals. In effect they are the entry level in scientific paper publishing and they have significant expenses. Even if peer reviewers and editors are not paid there are still significant support staff needed to shuffle the documents around and maintain the servers, hardware cost, bandwidth costs, insurance costs, customer service costs, etc. The cost of publishing is non-zero and adding editing, peer reviews and journals adds to the cost. Someone has to pay for it and the question is whom.
There are costs to publishing journals. That money had to come from somewhere. Where would you suggest?
ArXiv is an e-print archive. It does not appear that they do any peer reviews or editing. It would seem that hosting journals would be even more expensive. It look more and more like the "it's cheap enough for anyone to take over" crowd is way off.
There is so much talk about how cost are minuscule and any reasonably sized institute could bear the load. If that was true then why has it not been done already? I am sure most institutes would love to get rid of the costs of journal subscriptions. Perhaps it is not as easy or low cost as some people think.
It seems reasonable that a publisher would have to recover costs and make a profit. If they can not recover it from subscription the only other choice is to charge contributors. Publishers are not charities. According to this annual report Taylor & Francis' parent compant made a 27% profit in the Academic Information sector and 7% overall. Without that cash cow the company is not viable.
If the mail outlet that initially processed the packages put them into bags there is a high likelihood that there are bags that contain only marked or unmarked packages. The difference is it is a loss of one bag not ten packages. The loss of one bag can easily be explained by fluke. That is why I say this is a poorly designed experiment. Had they thought more about the failure possibilities than trying to prove a point they may have come up with a better plan.
The thing you keep missing is that you keep treating each separate box as a shipment with it's own sorting and tracking. That is not what happens. If a outlet in Germany get a number of packages that are going to the US they will put them in a bag or shipping cage. Now that 89 shipments become one to ten amalgamated items. If one of those amalgamated items get delayed there is a significant impact on the statistics.
In this situation we have some method that when 50 branded (A) boxed and 50 unbranded (B) boxes passed through it we saw (among other things) 1 B box get lost and 10 A boxes get lost. that is significantly unlikely to be by chance.,
If somewhere along the line those 100 boxes are put into ten bags and one of those bags are lost then the statistical significance is gone.
When mail is sent from point A to point B it goes through a number of hubs where the mail is either amalgamated or dispersed. Have you seen a tracking display for a parcel? You will notice that there are several stops along the way. Moving between these hubs can take several trucks which arrive and leave at different times. If the truck that happened to have the marked packages on it was delayed there would be a significant difference. Had they done a few packages a day over a few weeks I would be more inclined to believe it.
Almost forgot. Just because they were handed to Deutsch Post at the same time does not mean they were processed at the same time. It takes time to process packages and it is possible that the truck was full part way through the batch. Without looking at the tracking data all we can do is guess.
Did they arrive in the US at the same time? Lets see the tracking data.
I never completely believe the executive summary of any study and have found flaws in many studies.
I can't seem to find the source of your quote. What document is it referencing?
My point is that if there are other possibilities then the hypotheses is not proven.
Even if many of the packages arrived at the same time there could be other groups of packages that missed the cut off time for the hub. I would like to see the distribution of the delays and losses. There may be a pattern.
They sent 86 packages at the same time. This shipment is broken up along the way so the N is somewhere between 1 an 86. Had they sent the packages on different days then the N would be more accurate.
Have you ever heard the saying "once is an anomaly, twice is a coincidence, three times is a pattern"?
It is very likely if they were sent in two batches; one before the cut off time and one after.
Not if they were sent in two batches that were sent before and after the cut off time.
The step they did not take was to try to explain why the significant difference happened. They did not look at the tracking data to see how or where the packages were lost. They did not repeat the experiment to confirm that this one multipart shipment was not a fluke.
It could be argued that the N is actually somewhere between 2 and 178 as at different points in the route some packages are grouped together and a issue local issue could effect more than one package. I would have been much more convinced had they sent a packages of each type per day over a period of weeks so that a single anomaly, such as the loss of one mail bag, could not skew the statistics.
This whole study is based on a false assumption; that packages sent on the same day will arrive at the same time. A package sent before the cut off time will arrive on a different day than a package sent after the cut off time. A package that will not fit on a full plane will wait for the next plane, A package that will not fit on a full truck will wait for the next truck. Were one set of packages mostly delivered on Saturday and the others Monday? There is two days difference. Are they talking calendar days or business days?
No medical study that has not been peer reviewed is taken as proof so why should this one be. There is an old saying "lies, damn lies and statistics".
Only if both batches made the cut off time for that day. One group of packages could have gone out one day and the other group may have gone out the next day.
Actually two packages sent as little as half an hour apart could arrive on different days. If one package made the cut off time and the other one did not one package would be on tonight's truck and the other one would wait for the next one. I did a test on the USPS site and, depending on the service requested the cut off times varied from 2:15PM to 4:30PM.
The problem with this "study" is that it makes a lot of assumptions on what is happening without even looking at the tracking data.
This is only one sample and not repeated, therefore not proof.
What I would like to see is the distribution of the missing parcels. I would check the tracking on the missing packages and see if they were on the same truck. If they were on the same truck did that truck have an accident?