Except speed and size are highly correlated. Computer architectures are still designed in such a way that if you use as little RAM as possible, and have a small computation footprint that's highly localized, then your performance skyrockets. So the available 2GB RAM is irrelevant - if you actually make full use of it, you'll be running a lot slower than if you can fit everything you need to do in 4K.
Seriously, running and hosting a website is expensive. If you completely
removed all adverts from the web then many websites would simply have to
close as it is impossible to reliably host something popular without
incurring costs.
Really? Bullshit. For about $100 per year you can easily have a hosting account for a small website. Add to that the cost of a domain name, and you're set.
Just publish any content you like and see what happens.
Now, if you're saying that you can't afford to sink $150 per year on a website, and you would only be able to do it by getting paid for adviews, I'm going to have to seriously wonder about your ability to deliver even a small amount of content.
It's true that some of the mega websites have huge bills in the millions for all sorts of things, but on that end of the spectrum, the actual money's coming from investors and real products, not adviews.
If you have an idea for a website, start small and see if you can get 10 (decimal) people to pay money as a result. Then maybe try for 20. Most ideas fail that test, and the sooner your idea fails, the sooner you can try another. If it passes these hurdles, you'll be self-funding and I wish you the best of luck.
It's not a geeky solution, and it involves people touching each other, but it's very reliable.
Ah, but I'm afraid that only solves 99.99% of cases. To fully solve the problem, you also have to deal with the 3 blind-choralists-in-a-row problem, which is well known to not have an analytic solution except in trivial cases. Perhaps a numerical approach using a Sobolev auditorium might work, though.
True, but he uses a special Tektronix terminal with a custom ROM that computes on the fly the optimal TeX rendering of any HTML document while he's browsing it. Please nobody tell him what the rest of us see...
IMHO, mathematics should not be patentable AT ALL and IN ANY FORM. However, what galls me particularly here is how this whole patent business appears being settled among a bunch of ignorant monkeys.
In mathematics, the analysis of rounding operations is nontrivial, and in general making a choice about rounding before or after some other operation can sometimes be extremely clever. So while the judge made ultimately the right decision, his justification based on labelling the claims "simple mathematical operations" seems woefully inadequate, suggesting that he has no clue at all. His official argument should at least be based on a more substantive understanding of whatever the system claims to do.
This reminds me of another famous example when a bunch of non-mathematicians decided they had figured out some simple mathematical operations.
Nevermind the ACs, that's pretty impressive. Although my first inclination would be to ask how sensitive are the sim results to the amount of detail being modeled? Mathematicians and physicists come up with a lot of approximations to simplify computations and reduce complexity, do those kinds of fully detailed simulations confirm the approximate answers?
Are you now or have you ever been a member of the Terrorist Party? Never mind, we'll just google your emails.
Why intercept emails retroactively in time, when those archives are being voluntarily put on servers we can google automatically from the office? Hey Schmidt, pass me a donut, will ya?
Meh, who says you're the target? Thanks for your data, now we can track that terrorist sympathising friend of yours!
PGP what now? We just care about your regular address book, we'll ask you to give us the PGP key after we find your suspicious activity from 10 years ago. You don't have anything to hide, do you?
You're not spoiling my argument. There are no economies of scale in publishing many journals from a single website. The reason is that handling an order of magnitude more traffic requires much more expensive hardware, more performant software, and more maintenance at the same level of service.
Economies of scale on the web are based on decentralization ideas, distributed server farms, crowdsourcing, etc. A publisher wanting to serve many journals from a single website faces the need for server farms, complex administration, bespoke software, maintenance, etc. It is cheaper to have several independent micropublishers, who each can handle reduced traffic on lower end hardware and software. The overall traffic will be the same, while the individual capital expenditures for each micropublisher will be much less than the single publisher's.
If you are talking about economies of scale associated with printing and binding journals, I agree with you. But this seems to be on its way out, to be replaced with printing on demand.
Wrong, or did you miss the whole start of this discussion where the editing
staff for a journal resigned from the publisher? The publisher does much of
the selection, editing and peer reviews.
Do you mean the editors of the Journal Of Library Administration? Do you actually know why they resigned? Because they believe that the publisher of the journal, Taylor & Francis, doesn't deserve the free copyrights that they collect from the authors who actually do the work. And by the way, the editors and staff who resigned aren't Taylor & Francis employees, they are academics employed by universities. Taylor & Francis staff aren't involved in academic aspects, for the simple reason that non-academics aren't normally qualified to judge the merits of papers, nor are they qualified to edit papers without strong supervision, as even a single incorrect comma can change the meaning of a sentence or an equation.
Imagine a local academic talking to thousands of librarians trying to
convince them to host their paper. That may be a full time job.
Luckily, academics don't need to convince librarians to host their paper, they just need to convince a journal to accept their paper, and convince their own librarian to host a list of journals that they would like to read. This is a lot of work for the former, and about 5 minutes for the latter.
Remember you are talking about small repositories that will not impact a
host's storage or bandwidth. What is a paper is only hosted on a few sites?
Let's do a quick calculation. A small repository of papers might be 1 gigabyte.
A disk with that capacity is less than $100 today. A typical electronic journal article is about 300kb on average. So you can put about 3000 papers on a 1 gig drive, which is several years worth of articles for one journal, probably its complete archive since the year it was founded. So it's almost trivial to host a complete journal archive, and what you're asking is what if some obscure journal is only hosted in a handful of places? The answer is, find someone with a spare 1 gigabyte on their drive to host a copy of it. I could probably do it from home, or you. For an obscure journal, I'd expect about 1 or 2 downloads per month or less.
If it is not hosted locally one would call the support desk where it is
hosted. What happens when thousands of people from all over the world start
relying on the local repository and it goes down causing the local support
desk to have to deal with thousands of calls and emails? The repository will
not last long.
You're thinking of something like a slashdotting? If you look at how this problem is solved on slashdot, there's always one guy who posts a cached/mirrored copy of the slashdotted site. Same thing will happen there.
In case you're actually thinking that this one repo will become the single point of regular access for everyone, that's nonsense. That's only what happens for commercial publishers _because_ they don't permit free redistribution. The whole point of a distributed system is to solve that problem. So what would actually happen is that the initial journal archive would implement HTTP 302 redirects to a mirror site which would receive half the load, and in turn it would redirect to reduce its load. In the end, you would have N mirror sites each with about 1/N of the initial load.
Why are there not free open journals that do not require copyright transfers
right now? No one is holding a gun to author's heads and forcing them to use
a paid publisher.
In some sense they are. The funding agencies have a list of preferred journals that they want to see academics publish in. These journals could be anything, but nearly all are owned by commercial publishers right now. Free/Open journals need to stick around for another 10/20 years to gain a
I think the thing you missed is that publishing is not the same as hosting.
Indeed, "publishing" is *not needed*, only hosting is. "Publishing" is the dubious service that "publishers" offer, after the papers have already been written, selected, peer reviewed, re-edited by the authors, and accepted, all for free by the academic community as part of their job.
Who decides which document goes on which computer?
Anyone who wants to copy the documents and host them. Imagine thousands of university librarians talking to local academics, then downloading the journals of interest into a local archive, accessible to everyone on campus, and even the wider public.
What happens if one or more of the small repositories goes down?
Nothing. You pick another repository somewhere else in the world which has a copy of the paper you want. Once your local repo is up again, you can use them again.
Who get the editing and peer
reviews done?
The same academics who are doing it now. It's part of their job.
These small repositories are not going to provide any customer
support.
On the contrary. If the local university library has a journal archive, and
you want support, you can just pick up the phone and call someone, and if you really want to you can walk down the street and get someone to help you *in person*.
Who uses USENET now? There might be a reason.
USENET encompasses a protocol and distribution model to share messages and documents in a decentralized and redundant way. It is used for everything, including discussions (moderated and not), trolling, and swapping pirated media of arbitrary size, by the general public. That's who uses USENET now, and I'm not suggesting dumping science journals on there as is.
What I am suggesting is setting up a parallel, USENET like, distribution system where newsgroups ("journals") are read only and carry papers, while the equivalent of posting a new article is completely moderated, via the usual anonymous peer review. Anyone could "subscribe" to those journals for free without posting privileges, which simply means that they would find an existing repository that carries the "journal" feed and they would download it, making it available as an identical local feed. Well, I'm not going to explain in detail how USENET works, look it up. The point is that replacing "publishers" with this system is the right thing to do.
If is is so simple and easy,
why hasn't it been done by now?
Copyrights. The copyrights on papers accepted by the for profit publishers are relinquished by their academic authors as part of the deal, and the publishers do not allow free copying and redistribution by anyone after that.
Perhaps it is not as simple as you think.
No, it is not. The only two solutions currently are 1) to create open journals which do not prevent copying and redistribution - this takes time for the journals to flourish, and 2) to "steal" back the journal articles by force and distribute them - the pirates are making good progress on this front.
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.
And once again we go round the usual arguments: The root cause of high costs is centralization, which is a symptom of the need for control by the publishers. The solution is decentralization, which means mirroring (technically) and free copying (legally).
The world is full of computers, and people and organizations willing to host for free small repositories of documents. Let everybody copy and republish scientific articles for free, and 1) there won't be a single point of access which requires beefy hardware, beefy bandwidth, beefy customer service, etc , 2) there won't be a problem with insurance due to high redundancy. This is the USENET model, and it is ideal for scientific resarch.
The Amazon is big, and policing the farms directly is tough. OTOH, a major supermarket chain is a single point of delivery and redistribution for many farmers' produce. Thus it is simpler to prevent undesirable behaviour by refusing to buy the goods there.
Of course, this only has a chance of working if a substantial fraction of all the supermarkets in Brazil refuse to accept the produce.
Sounds like a scam to me. The whole point of trademarks is to prevent others from using the mark in the course of their normal business activities without explicit authorization from the owner. By making trademarks available as gTLDs, ICANN is directly infringing on the trademark owner's rights, doubly so if they are doing so knowing that the word happens to be a trademark. So any trademark owner could sue ICANN for using their trademarks without permission.
IANAL and all that, but I don't see how ICANN can even legally offer gTLDs that happen to be trademarks. (*)(**)
(*) I would actually expect ICANN to pay someone to filter out all words that *are* trademarks, just to make sure that none of them could possibly be sold as gTLDs. I guess I'm not a CEO.
(**) Perhaps the insurance scam is designed to contractually establish ICANN's right to use the trademark for their own business purposes, selling it back to the legitimate owner? Sneaky!
Slashdot Mirror
Except speed and size are highly correlated. Computer architectures are still designed in such a way that if you use as little RAM as possible, and have a small computation footprint that's highly localized, then your performance skyrockets. So the available 2GB RAM is irrelevant - if you actually make full use of it, you'll be running a lot slower than if you can fit everything you need to do in 4K.
Really? Bullshit. For about $100 per year you can easily have a hosting account for a small website. Add to that the cost of a domain name, and you're set. Just publish any content you like and see what happens.
Now, if you're saying that you can't afford to sink $150 per year on a website, and you would only be able to do it by getting paid for adviews, I'm going to have to seriously wonder about your ability to deliver even a small amount of content.
It's true that some of the mega websites have huge bills in the millions for all sorts of things, but on that end of the spectrum, the actual money's coming from investors and real products, not adviews.
If you have an idea for a website, start small and see if you can get 10 (decimal) people to pay money as a result. Then maybe try for 20. Most ideas fail that test, and the sooner your idea fails, the sooner you can try another. If it passes these hurdles, you'll be self-funding and I wish you the best of luck.
mv resume.txt Resume\ \ \ .DOC.txt
Locally, it would be "l'effet Streisand", pronounced leffey chtrey-zond.
Ah, but I'm afraid that only solves 99.99% of cases. To fully solve the problem, you also have to deal with the 3 blind-choralists-in-a-row problem, which is well known to not have an analytic solution except in trivial cases. Perhaps a numerical approach using a Sobolev auditorium might work, though.
True, but he uses a special Tektronix terminal with a custom ROM that computes on the fly the optimal TeX rendering of any HTML document while he's browsing it. Please nobody tell him what the rest of us see...
In mathematics, the analysis of rounding operations is nontrivial, and in general making a choice about rounding before or after some other operation can sometimes be extremely clever. So while the judge made ultimately the right decision, his justification based on labelling the claims "simple mathematical operations" seems woefully inadequate, suggesting that he has no clue at all. His official argument should at least be based on a more substantive understanding of whatever the system claims to do.
This reminds me of another famous example when a bunch of non-mathematicians decided they had figured out some simple mathematical operations.
Less anuses than Goatse? Lame!
Why is this a problem? The French Foreign Legion is one of France's elite military regiments, and is full of foreign-born individuals.
It's too late by then. Security needs to be designed into a system from the start. You can't put it in within minutes of somebody wanting it.
See Microsoft, they've been trying for decades to retrofit security into their systems, and failing. You think Apple's engineers are can do better?
Nevermind the ACs, that's pretty impressive. Although my first inclination would be to ask how sensitive are the sim results to the amount of detail being modeled? Mathematicians and physicists come up with a lot of approximations to simplify computations and reduce complexity, do those kinds of fully detailed simulations confirm the approximate answers?
Apparently, he also made a lousy boss...
"It's life, Jim, but not as we know it."
"Green life? Ok, Bones, leave the ... diplomacy to me!"
Well, that's the *good* news. The *bad* news is that the patch to make this one parameter change is 12Gb.
Thank you for your efforts, we appreciate it.
Sincerely,
The FBI.
Economies of scale on the web are based on decentralization ideas, distributed server farms, crowdsourcing, etc. A publisher wanting to serve many journals from a single website faces the need for server farms, complex administration, bespoke software, maintenance, etc. It is cheaper to have several independent micropublishers, who each can handle reduced traffic on lower end hardware and software. The overall traffic will be the same, while the individual capital expenditures for each micropublisher will be much less than the single publisher's.
If you are talking about economies of scale associated with printing and binding journals, I agree with you. But this seems to be on its way out, to be replaced with printing on demand.
Do you mean the editors of the Journal Of Library Administration? Do you actually know why they resigned? Because they believe that the publisher of the journal, Taylor & Francis, doesn't deserve the free copyrights that they collect from the authors who actually do the work. And by the way, the editors and staff who resigned aren't Taylor & Francis employees, they are academics employed by universities. Taylor & Francis staff aren't involved in academic aspects, for the simple reason that non-academics aren't normally qualified to judge the merits of papers, nor are they qualified to edit papers without strong supervision, as even a single incorrect comma can change the meaning of a sentence or an equation.
Luckily, academics don't need to convince librarians to host their paper, they just need to convince a journal to accept their paper, and convince their own librarian to host a list of journals that they would like to read. This is a lot of work for the former, and about 5 minutes for the latter.
Let's do a quick calculation. A small repository of papers might be 1 gigabyte. A disk with that capacity is less than $100 today. A typical electronic journal article is about 300kb on average. So you can put about 3000 papers on a 1 gig drive, which is several years worth of articles for one journal, probably its complete archive since the year it was founded. So it's almost trivial to host a complete journal archive, and what you're asking is what if some obscure journal is only hosted in a handful of places? The answer is, find someone with a spare 1 gigabyte on their drive to host a copy of it. I could probably do it from home, or you. For an obscure journal, I'd expect about 1 or 2 downloads per month or less.
You're thinking of something like a slashdotting? If you look at how this problem is solved on slashdot, there's always one guy who posts a cached/mirrored copy of the slashdotted site. Same thing will happen there.
In case you're actually thinking that this one repo will become the single point of regular access for everyone, that's nonsense. That's only what happens for commercial publishers _because_ they don't permit free redistribution. The whole point of a distributed system is to solve that problem. So what would actually happen is that the initial journal archive would implement HTTP 302 redirects to a mirror site which would receive half the load, and in turn it would redirect to reduce its load. In the end, you would have N mirror sites each with about 1/N of the initial load.
In some sense they are. The funding agencies have a list of preferred journals that they want to see academics publish in. These journals could be anything, but nearly all are owned by commercial publishers right now. Free/Open journals need to stick around for another 10/20 years to gain a
Indeed, "publishing" is *not needed*, only hosting is. "Publishing" is the dubious service that "publishers" offer, after the papers have already been written, selected, peer reviewed, re-edited by the authors, and accepted, all for free by the academic community as part of their job.
Anyone who wants to copy the documents and host them. Imagine thousands of university librarians talking to local academics, then downloading the journals of interest into a local archive, accessible to everyone on campus, and even the wider public.
Nothing. You pick another repository somewhere else in the world which has a copy of the paper you want. Once your local repo is up again, you can use them again.
The same academics who are doing it now. It's part of their job.
On the contrary. If the local university library has a journal archive, and you want support, you can just pick up the phone and call someone, and if you really want to you can walk down the street and get someone to help you *in person*.
USENET encompasses a protocol and distribution model to share messages and documents in a decentralized and redundant way. It is used for everything, including discussions (moderated and not), trolling, and swapping pirated media of arbitrary size, by the general public. That's who uses USENET now, and I'm not suggesting dumping science journals on there as is.
What I am suggesting is setting up a parallel, USENET like, distribution system where newsgroups ("journals") are read only and carry papers, while the equivalent of posting a new article is completely moderated, via the usual anonymous peer review. Anyone could "subscribe" to those journals for free without posting privileges, which simply means that they would find an existing repository that carries the "journal" feed and they would download it, making it available as an identical local feed. Well, I'm not going to explain in detail how USENET works, look it up. The point is that replacing "publishers" with this system is the right thing to do.
Copyrights. The copyrights on papers accepted by the for profit publishers are relinquished by their academic authors as part of the deal, and the publishers do not allow free copying and redistribution by anyone after that.
No, it is not. The only two solutions currently are 1) to create open journals which do not prevent copying and redistribution - this takes time for the journals to flourish, and 2) to "steal" back the journal articles by force and distribute them - the pirates are making good progress on this front.
And once again we go round the usual arguments: The root cause of high costs is centralization, which is a symptom of the need for control by the publishers. The solution is decentralization, which means mirroring (technically) and free copying (legally).
The world is full of computers, and people and organizations willing to host for free small repositories of documents. Let everybody copy and republish scientific articles for free, and 1) there won't be a single point of access which requires beefy hardware, beefy bandwidth, beefy customer service, etc , 2) there won't be a problem with insurance due to high redundancy. This is the USENET model, and it is ideal for scientific resarch.
The Amazon is big, and policing the farms directly is tough. OTOH, a major supermarket chain is a single point of delivery and redistribution for many farmers' produce. Thus it is simpler to prevent undesirable behaviour by refusing to buy the goods there. Of course, this only has a chance of working if a substantial fraction of all the supermarkets in Brazil refuse to accept the produce.
What?! Zombies can read? Shit, I was going to hide behind the entrance of my favourite restaurant!
Non sequitur! A lot of criminals don't wear shoes, why should law abiding citizens dress like criminals?
IANAL and all that, but I don't see how ICANN can even legally offer gTLDs that happen to be trademarks. (*)(**)
(*) I would actually expect ICANN to pay someone to filter out all words that *are* trademarks, just to make sure that none of them could possibly be sold as gTLDs. I guess I'm not a CEO.
(**) Perhaps the insurance scam is designed to contractually establish ICANN's right to use the trademark for their own business purposes, selling it back to the legitimate owner? Sneaky!
AND it's upside down to confuse the Aussies! What a jerk!