It does discuss how they used the existing Indiana counties that were already DST to control for non-DST changes between the sample years. The thing that gets me is that it states but does not explain that DST impacts heating/cooling. As best I can tell, the idea is that people are more likely to turn on the heat or A/C when awake. I wish that they had investigated that more.
It's also worth noting that the study says that DST does save electricity in the spring. It's the fall when DST has the negative effect overall. That suggests to me that we should end DST earlier if we want to maximize the electricity benefits.
Eventually this morphed into liberal socialism; which supported the state providing for individuals so that they had equality of opportunity as well as freedom of opportunity. Only in the US. In Europe (e.g. UK), liberal means more what libertarian does in the US: social and economic freedoms are both important. There is more discussion of this at http://en.wikipedia.org/wiki/Classical_Liberalism and on the accompanying discussion page.
The Liberal Party of Australia is actually economically laissez-faire and socially conservative--the exact opposite of the US.
Btw, I would argue that liberal socialism involves a trade off between equality of opportunity and freedom of opportunity. It limits the latter to increase the former.
It's pretty easy to "re-implement" newly developed pharmaceutical drugs as well... The differences between pharmaceuticals and software include:
1. Pharmaceuticals are only covered by patent; software is also covered by copyright. With pharmaceuticals, the only protection that a company has against another company copying their drug is the patent. That's why there is a flourishing generics market. By contrast, software binaries are covered by copyright.
2. If a company comes up with a new drug that has the same effect as an existing drug, it's not covered by the pharmaceutical patent. See Cialis and Levitra as examples of this (both are primarily billed for the same problem as Viagra). Software patents are generally applied to others with the same idea rather than the same implementation (implementation is already covered under copyright, so patenting it is somewhat redundant).
3. Pharmaceuticals require clinical trials to be sold. These clinical trials are the majority of the cost of pharmaceutical development. Software seldom has the same level of testing requirements, and when it does, software is covered by copyright. A reimplementation that evades copyright would also require new testing (in the same way that Cialis and Levitra required separate testing from Viagra).
I would be fine with a software patent that was strictly limited to a specific implementation. However, such a patent is almost entirely redundant to the existing copyright protection. It's just as expensive (if not more expensive) to reverse engineer software as it is to write it new. In writing new, you only need to understand how what you wrote works; in reverse engineering, you need to understand how what they wrote works and how what you wrote works (and some way of verifying that those are the same).
In terms of pharmaceuticals, I actually think that the correct thing would be to rename the protection from patent to copyright. The pharmaceutical copyright could work as follows:
1. Some entity registers the copyright for a drug recipe's application to a specific problem. That registration requires a clear plan for starting clinical trials within one year.
2. That application is protected by copyright throughout the clinical trial period. If the clinical trials need to be extended for further investigation, then the protection extends as well.
3. Once the clinical trials complete and the FDA signs off, the copyright protection will last for seven to ten years (about the length of time protection extends after clinical trials now). After that, the drug goes into the public domain.
Pharmaceutical copyright would better describe this than patent and this program would better serve the needs of both consumers and companies. Currently companies under tremendous pressure to bring drugs out of clinical trials as soon as possible. The clock is already ticking on the expiration of patent protection. By changing to a system where the patent clock starts when clinical trials end, it removes the fear of losing revenues. This will allow companies to more naturally balance the risk of waiting (where other drugs can enter the market and where the company needs to pay interest on the money borrowed to perform the clinical trials) with the risk of being sued (because the drug has undesirable side effects).
That would require the patent office to be able to find 12,000 people in the US each year that want to be patent examiners.
I still think that the solution is to eliminate the idea of patents being "granted" after a review by the patent office. Instead they should change to a system where patent applications are recorded and only reviewed when someone tries to enforce the patent. At that time, the defendant can be responsible for doing the prior art search and the patent office only needs to be responsible for *evaluating* the claims. It's a much stronger model.
The current patent examiners have no incentive to deny patents, as their productivity is entirely measured in patent applications processed. Obviously it is easier to approve an application than it is to find a valid reason to deny it, so approval becomes the default state.
Doing things that way would also push patent applications to be better up front. Under the current system, it makes sense to request the broadest patent that you can. What's the worst that happens? The patent is denied and you refile with more narrow claims. By removing that initial review and weakening the meaning of the patent application being accepted, the system pushes more of that burden on to the appliers. Now they want to write the best patent applications possible so that the patent application survives the more rigorous test of examination by a defendant (who is presumably knowledgeable in the field and highly incented to find reasons why the patent is invalid). Particularly since there now is no second chance for a patent application -- if it's badly written and falls over, refiling now would be too late. People already have competing implementations.
The current system is bad because it disassociates costs and benefits. The cost of the bad patents are borne by the defendants in patent infringement cases, but those defendants are not involved in the original patent application. As such, they have no opportunity to block bad patents even though they have the incentive to do so. Patent examiners benefit from quick resolution of patent applications but bear none of the costs of passing bad patents. Patent applicants actually benefit from their own bad patents.
Fixing this either requires the patent office to bear some of the cost of passing bad patents or it requires replacement of the current system with one which allows defendants to participate in the evaluation of the patent claims. Making the patent office bear the costs of bad patents would be hard to implement and the transition would be expensive. It seems much simpler to change the process so that patents are not awarded but only requested. That moves patent application processing into a realm where productivity goals make sense and allows defendants to participate in the invalidation of patents.
I've had Firefox use several hundred megabytes. In my experience, it becomes unstable between half a gig and a full gig, but it will grow that far under the default install. Part of this may be that I was using a lot of intranet applications that were basically just giant lists of drop down menus. These pages were huge. It's also worth noting that I would leave a browser window open until it crashed (I don't like giving up my current state).
It sounds like the poster wants to get a small number of correct results but instead is getting a large number of incorrect results. In other words, a large number of programmers who claim to be the best turn out not to be so good. How do you handle that? Resume screening, phone screens, and interviews are expensive processes that distract your current expert programmers from their real work. How can you minimize the amount of time that your limited resource (good programmers) wastes on identifying false positives (people who are passed by the initial resume screening process but who really aren't that good).
Also, how do you catch the false negatives? What about the programmers who are not good resume writers or especially social but who can code up a storm. How do you find these people?
What's really wanted here is an entirely different version of the interview process. I.e. instead of having a bunch of human beings evaluate what the prospective programmer says (in speech or written communication), find a way to evaluate coding directly. Maybe something like scriptlance, where you write a specification for software and someone implements it. If the implementation passes your tests (for unit test coverage, functionality, etc.), pass the code on to a real human being to evaluate. If that works out, ask for more code (and maybe pay for the original code or maybe that's just a test).
They would have to carry some store of energy (fuel or battery) to draw on to perform the electrolysis, which is silly. Regenerative braking? A quick google suggests that battery efficiency is more like 70-85% rather than 96%. Of course, the 96% for hydrogen production may not be the comparable number to battery efficiency.
A car engine needs to spin at times when it isn't delivering power to the wheels. At those times, it may make sense to power the alternator and produce electricity. It makes sense to store that electricity. One way to store it would be to electrolyze some water into hydrogen and oxygen. It may or may not make sense to do that rather than store it in the battery (which might be full or simply less efficient). Presumably the hydrogen process is more efficient than the typical battery; otherwise, why not simply use a bigger battery with an electric engine and skip the hydrogen altogether?
Btw, your argument about the evils of property taxes could apply to practically anything. E.g. "The whole point of income is that its money that is paid to you. Having an income tax is like part of it is not your money but really the government's."
Property taxes make sense for some kinds of spending. In private insurance, the cost of the insurance is based on the value of the property being insured. Some kinds of government spending can be considered public insurance against damage to property; law enforcement and defense come to mind.
There's actually a branch of libertarianism called geolibertarianism that maintains that the only valid tax is the tax on land. Land is unique in that it can't be produced, where most everything else only has value accrue to it based on the labor that someone does to it. A quick link from Google: http://members.aol.com/_ht_a/tma68/geo-faq.htm
Um, not to mention that what's described is essentially academic indentured servitude -- I can see VERY little reason for a company not to exploit any loan recipient who's obliged to work there or suddenly get slammed with $80k of debt, and without even the entry-level worker's minimal ability to negotiate offers from multiple employers and move if conditions are seriously unfavorable, wages for new college graduates would fall through the floor.
How so? There are basically two ways to pay back the money: one, work for the company that loans it to you; two, get a job somewhere else and pay back the money. The second option is the entire current option. What's happened to make it worse? I.e. what about this proposal makes the loan worse for a student?
Note that there is some risk in loaning money to a student before graduation. It's quite possible that the student will simply flunk out. After graduation, that risk goes away. Presumably companies would be better off hiring from the pool of graduates rather than loaning to undergraduates -- assuming that the graduates are available. A valid corporate strategy would be to only hire students who performed well in school supported by loans from other companies.
Advantages to students of this program:
1. No risk of graduating, not getting a job, and still having to pay back the loan. If the company wants to get paid back, it must make a job offer.
2. Increased information about what majors are employable (currently this is entirely on the student to evaluate). If no one is willing to loan money to support a major, then clearly it is not employable. If people are approaching the student recruiting for a particular major, then it is. This kind of program makes choosing a school more like choosing a job, where the companies define the process for the student (rather than students choosing majors based on criteria that the student has to develop).
3. Strong likelihood of internship opportunities. If a company is investing in a student, then the company is likely to want to evaluate its risk as early as possible. This means that the student is likely to get internship offers from the company.
Advantages to companies:
1. Earlier participation in student educational choices.
2. More predictability in student hiring.
3. More rewards for taking risks on students. Currently a company can give a scholarship but has no way of getting a return on that.
4. Flexibility in the face of demand changes. If they don't need to hire, they lose what they've put into the student's education already but have no further obligation.
Disadvantages to companies:
1. Requires more planning as to future hiring needs.
2. Has to take risks based on projected future effectiveness. If a student flunks out or drops out, the company has to wait for the loan to be repaid and loses the employment advantages.
I'm also not convinced that this would end pure mathematics as a major. There are "companies" that are very interested in pure math majors; we call them universities. Universities could use their extra insight into student qualifications to encourage good students to enter mathematics as a field. If the student later abandons academics for a computer programming or finance job, the university can recoup that investment as a loan (currently they would have to give a scholarship to encourage the student and gain nothing if the student later chooses not to pursue academics).
Mathematics is an example of a major where there is a shortage of students. This program is designed to help such majors (by providing a funding mechanism for such majors). This program would likely be bad for majors like English, where there are high numbers of students and low numbers of employment opportunities.
I also find it unlikely that most companies are likely to only pay for engineering courses and not literature courses. In general, I think that most companies are
Greater than George Washington or Abraham Lincoln? Oooookaaaaayyy... The greatest thing about George Washington was that when they asked him to be king, he said no.
Abraham Lincoln was simply at the right place to become famous. He was President during the US's most deadly war. Remember that the South broke off based entirely on FUD about what Lincoln might do as President. Eventually he freed the slaves, but only after it became obvious that the South was not going to negotiate until they were broken. In the end, the North won because it had more soldiers and manufacturing (as well as control of the Navy). The turning point of the war was when the South invaded Gettysburg to loot shoes. Then Lincoln unleashed the butchers, Grant and Sherman, until the South was broken.
I think that there are any number of Americans who could be considered greater than Isaac Asimov (who was a prolific writer but hardly changed the world): George Washington Carver, Jonas Salk, Richard Feynman, Elijah McCoy, Albert Einstein, Sam Walton, Henry Ford. However, why the obsession with generals? George Washington is a famous moral example, but is he really someone we want people to emulate otherwise? Henry Ford turned assembly line manufacturing from low paid grunt work into a high paid profession. Jonas Salk cured polio. Elijah "The Real" McCoy defined engineering quality (and I would argue that the phrase "The Real McCoy" has had a greater positive effect than the word robotics). So on and so forth.
There are too many people attending college today simply looking for any degree. This results in over-enrollment in so called easy majors, and less funding for science and engineering. You don't see nearly as many foreign students in those programs because, for them, the job market back home requires real knowledge, not just a piece of paper. I also think that you'd find that many foreign students have their educations funded by someone who cares what major they choose. In the US, the primary sources of funding are loans (controlled by the student), grants (given by the government for any major), need based aid (given by the school for any major), and parents. Grants and need based aid could be focused on particular majors but are not.
I've toyed with ideas about programs that would be more corporately focused. For example, what if student loan recipients were chosen by companies? The company would be on the hook for hiring the student after graduation. The student would be responsible for maintaining good grades in a major approved by the company (note: students would be able to pick the company that offered a major that they wanted). Students who flunk out, change majors (without a new sponsor), or who decide not to work for their sponsor have to pay the loan back. If the company cuts back staff and does not hire the student, then the company eats the loan. If the company hires the student, the company is assumed to have adjusted the student's pay appropriately. After some number of years, the student will finish the loan period and can switch companies without paying back the loan.
Another possibility would be to replace federal grants with corporate tax credits. Companies could pay for a student's tuition and mark it down as taxes paid. Obviously it would be more efficient for a company to pay tuition for a student it would like to hire than someone who is interested in an entirely different field.
A big problem with US education before college is the shortness of the school year. Why not take a page from Germany's book and switch to ten 216 day years in elementary and secondary school (the same 2160 days that come from twelve 180 day years)? Then go to a two year program that could be more general than a university degree (i.e. something like Engineering, Science, or Liberal Arts rather than Electrical Engineering, Physics, or Philosophy) and more specific than the final two years of secondary school currently are. Afterwards, students could go to the regular university with a more consistent and focused presentation. For people who aren't college inclined, they could use those two years in a trade school.
What a needless waste of bandwidth running large applications, such as office applications on the internet would be. If every application ran from the internet, there would be little bandwidth left for anyone. Sure, but if no application used the internet, you wouldn't have posted this. There are many applications where it makes sense to have some sort of centralized storage: blogs, email, newsgroups, etc. There are others where the internet is at least needed as a communication medium: email (again), Instant Messaging, bittorrent, etc. There are advantages to interconnectedness.
It's also worth noting that in many cases what they want to do with the technology is to extend the experience that currently only works on the internet to the desktop. I.e. much of the advantages of this is not running things on the internet that used to be local but in running things locally that used to be on the internet. For example, Amazon.com has a variety of tools on their website to allow third party sellers to list products. They also have a downloadable application called Amazon Seller Desktop that serves the same purpose. Making the local copy work seamlessly with the remote version is important. In that sense, what you see happening now is more a continuation of the trend from centralized computer to a decentralized mechanism.
A lot of the work that is being done now is making it so that you don't need to care whether the information is only stored remotely or also stored locally. Only stored locally becomes problematic when you want to share with others or when you want to work somewhere else. Remote storage with local caching gives you most of the goods from both worlds. An example of a lost positive is that remote security is still inferior to the pure local solution. Examples of gains from remote storage are the ability to share with others and the ability to switch workstations without making an explicit effort to transfer the information. Examples of gains from local caching are performance and the ability to work offline.
Not every choice needs to be made between extremes. Sometimes the best choice is a hybrid between extremes. Further, there may be multiple best choices depending on the exact application under consideration.
We don't really worry about that for car insurance, or flood insurance, Actually, in Pennsylvania, flood insurance is provided through a state program and car insurance includes an assigned risk pool specifically to make sure that everyone can drive. There have been proposals to make Medicaid/Medicare like that, where anyone can buy into the program for coverage. An insurer of last resort.
It's interesting that these programs already exist for flood and car insurance (which as you note are avoidable risks) but do not exist for health insurance (which while reducible, is not really avoidable).
What is the justification for putting that semantic meaning into a variable name, instead of incorporating it into class definitions? It's to take two variables that are otherwise identical and interchangeable in behavior of operations and mark out the difference in the data they hold. For example, if you have x and y coordinates, should you have separate types for x and y? Or should you just have two integer types with x and y in the name? Remember that it's perfectly reasonable to do arithmetic operations that combine x and y values (e.g. calculating distance from origin). Should you have to redefine all possible arithmetic operators? Twice? (Once for rows and once for columns -- same code in two places.) Or just use a type like Long for each and name appropriately?
In the Safe/Unsafe string example, if there is something about the constructor that could be used to make the string "safe" (e.g. verifying that it is shorter than a certain length), then you may be right, a separate class is better. However, if you are doing complicated business logic outside the constructor to establish safety (e.g. doing multiple database lookups and correlating the results), then you're better off with a naming convention. In that case, the class is not enforcing its own meaning. However, the purpose of classes is to enforce their own meaning. By contrast, a variable name does not enforce its own meaning, but one wouldn't expect it to do so. It does indicate meaning, which is what Apps Hungarian was designed to do.
There are also cases (WSDLs come to mind) where the language simply doesn't support strong typing for some of the cases (e.g. a types enum doesn't indicate that it is a type). In those cases, Hungarian notation helps compensate for a weakness in the language.
Database table names are another example. One should put Type or Status in the table name if that's what it holds. Database tables don't have typing, so the name is where you express that. Class names are another example. You are defining the class, so you name it appropriately. E.g. SafeString rather than Foo or MyClassName.
The basic idea is that giving a patent-holder a limited time monopoly to profit from their inventions encourages invention. Not exactly. The idea is that giving the patent holder a bonus for getting the patent encourages people to invest in research and development. This was primarily designed for things that were relatively easy to reverse engineer by taking apart. The idea is that if you spend a million dollars developing something that someone can reverse engineer for $20,000, the people who only reverse engineer will be more profitable than the people who do the heavy lifting of the design.
One big problem now is that they've moved from the original patent system, which required an actual implementation, to the current patent system where people can patent generic ideas. Ideas are cheap. Any idiot can come up with an idea that could work. The hard part is turning an idea into an implementation. I have no problem with people receiving patents on source code (although they would be stupid to do so, as copyright is so much better for the holder when it comes to source code). However, when someone else reverse engineers your behavior with their own source code, they should be able to use it. It's even worse when someone independently does the same thing without being aware of your product. However, with software, that happens all the time.
There are places where the patent system makes sense. For example, with pharmaceuticals, the expensive, difficult part is doing the studies to demonstrate to the FDA that the drug is safe and useful. Without some reward for that over and above the cost of manufacture, new drugs wouldn't be studied. There are still some definite tweaks that could be made. In the current system, any delay in finishing the study reduces the effective length of the patent. Why not start the patent term at the end of the study (and shorten it by the length of the typical study)? Then the pharmaceutical companies would be incented to do the study thoroughly (as they face a liability if there is a problem with the drug) without the counter incentive to do the study quickly to maximize the time that they can charge monopoly rents.
real INVESTIGATIVE reporting (which lies at the heart of the First Amendment protection of the Press) I see this stated often but established never. I think that this is you creating a right that you think should exist. Is there any evidence that anything like investigative journalism even existed in the eighteenth century?
The first amendment is about protecting the rights of normal individuals. A normal individual has the right to speak freely, to print (written speech) freely, to practice one's religion freely, and to peaceably assemble with others (for the redress of grievances). The free press of the first amendment is just as much (if not more) aimed at the printing of handbills as newspapers. It does not establish special investigative rights. It establishes communication rights.
Personally, I don't think that there should be any special investigative rights. It goes back to the whole "who watches the watchers" debate. How do you choose these professionals who have the right to do special investigations that the rest of us cannot? Either the right should be one that everyone can use or it isn't a right.
That's not to say that we shouldn't give the other privileges of which you speak to journalists. However, it should be clear that these are privileges, not rights. There are also some cases (e.g. source confidentiality) where it is clearer to think about the rights of the source rather than those of the journalist.
In a perfect world yes, but until voice recognition is perfected the speech input method requires one of the same things as typed input, eyes, in order to make sure it recognizes everything correctly so that you can fix *its* mistakes when the words aren't recognized correctly. Why not have it read back to you what it said? Or wait until later to fix the problems? I used to work with a guy who would just throw words onto paper and correct the spelling afterwards. Again, this might not work for everyone, but for some people, that better models how they approach problems.
For some applications, e.g. notes to oneself, errors aren't that critical. Don't bother correcting them. Jot down the idea and delete it when you have time to get back to it and process it formally.
Accenture (renamed Arthur Andersen) which did the falsified books for Enron and Worldcom. No, Accenture is the new name for Andersen Consulting. Andersen Consulting and Arthur Andersen split years before the accounting scandals (which involved Arthur Andersen).
If MS remakes Yahoo into MSN people will just jump ship. If I were Microsoft and buying Yahoo, I would be planning to remake MSN into Yahoo, not the other way around. If MSN's growth is stalling, I don't think that increasing its size will help. Microsoft already has huge advantages over the other players with their control over the desktop. They shouldn't need to buy market. Tech on the other hand is always good to have.
Now, they could be doing this to buy market share for MSN. IMO, if they are, they are doing the wrong thing. IMO, it would be far easier to redirect their internal advantages in Yahoo's favor than to redirect Yahoo's market to MSN.
I wonder what happens with Hotmail versus Yahoo! Mail? I could see that going either way. I don't normally use Hotmail, but I looked at it today and it's not immediately obvious to me which is better. The Hotmail client may be a bit more flexible and fully featured.
It just isn't much faster than typing Sure, but it's a lot safer to do while, say, driving down the road. The problem with screen output and typed input is that you have to use both eyes and hands to operate the device. By contrast, using speech input and output only requires voice and ears. Of course, there are some circumstances where the screen/type method is superior, e.g. sending emails from your blackberry during meetings. However, there are many cases where speech is superior, e.g. driving down the road (or even just walking). Viewing speech as a replacement for screen/type is over zealous. It's really more of an alternative.
It would probably help if advocates of the technology understood this. It doesn't have to be all or nothing. Two alternative solutions can add up to a more powerful solution than either would be alone.
Did they cut their pay? It sounds like they moved from a work as many hours as you want for one rate (salary) to a dual tiered rate where they charge more per hour for each hour over forty in a week. Assuming that employees were working forty-five hour weeks (on average), they would continue to get they same weekly pay now: r == (40 + t * 1.5) * (1 -.15) * r / 40 where t ~ 5 hours. I.e. it used to be that they were paid rate r (with the assumption that they would work an average of forty-five hours per week). If you divide r by forty, you get how much they used to make per hour for a forty hour week (the "base" salary). That rate has gone down 15% to (1 -.15) * r / 40. However, they now get paid time and a half for the hours they work past forty. Assuming they average 44.7 hours per week (with a minimum of forty), they'll actually get paid the same for working the same number of hours.
This may or may not be a good deal for the employees. If they used to work more than forty-five hours per week, it's a good deal. They can cut hours and still get the same pay. If they only work forty hours per week, then they would get paid less now.
The employer (IBM) doesn't like it because it makes the system harder to administrate. Managers now have to think along three axes: productivity per pay unit in the current hour; productivity per pay unit if delayed; cost if delayed. Presumably delay increases productivity per pay unit but also increases cost. There's also benefits to consider (it's actually cheaper to pay five hours of overtime for eight existing employees than to hire one additional person to work forty hours -- health care alone is often more expensive than five hours of overtime).
In a salary based system, the manager does not have to consider if it would be better to send the employee home or to pay over time. The manager can concentrate on results. If it takes the employee fifty hours, that's not the manager's problem, it's the employee's.
IANAL, it may very well be that they can get unemployment (under an implicit assumption that their previous rate of pay was based on forty hours per week even though the parameters of the story suggest that they were working overtime). However, the bare facts of the story (15% base salary cut but time and a half for overtime) does not necessarily mean a real pay cut. In practice, these employees may actually get an increase in pay (if they work forty-five or more hours per week).
Like the SSN can be used to find your entry in a database, but it should not be usable to take money from your account, for that they better know a real secret like your password or sign with your signature. That's already true. That's not the exploit under discussion. Identity theft is not about breaking an existing trust relationship between you and one of your financial associations. That's a separate class of scam (and while an SSN might help with it, other instruments are more beneficial, e.g. a credit card). Identity theft is about pretending to be you when establishing a new financial association in such a way that the benefit goes to the identity thief but the cost goes to you. The problem that arises typically goes like this:
1. The identity thief requests credit in your name, with your SSN.
2. The thief uses a variety of scams to transfer money or goods to them. The money comes from the new credit instrument.
3. The issuer of the credit now tries to collect from you. You are then responsible for proving that you did not request or use the credit instrument.
It's not that the SSN is really seen as that secret. It's just the piece that associates two entries for "John Doe" as referring to the same John Doe rather than two separate people who happen to have the same name.
If this same problem is less prevalent in Sweden, I suspect that one of two things is true. Either it is more difficult to get anonymous credit (over the phone or internet or through the mail) in Sweden than in the US, or this problem simply hasn't quite made it there yet.
If SQL is so cool, then why don't we have a different PL for each task? SQL (Structured Query Language) isn't a programming language. It's a client/server communication protocol. When you use a database, your program (the client) is calling the database (server). You need to have some protocol for that communication. SQL is more on par with FTP or HTTP than PHP.
Now, it's quite reasonable to want to abstract out that communication in your code. For many applications, you won't care about the underlying database implementation and can manage with just create, read, update, and delete operations. That's why Object Relation Mappers (ORMs) exist. It sounds like Symfony (which I haven't used) implements an ORM. A pretty common ORM for Java is Hibernate. An ORM can make things easier for many tasks. However, in practice, one occasionally finds that an ORM does not support asking the questions that actually need answered. SQL does. That's why Hibernate still allows people to write direct SQL.
Another issue is that SQL isn't always used from a programming language. It's quite common to connect to a database directly and query with SQL. It would be very inconvenient to have to write a PHP program every time I wanted to query the database. It's much easier to use SQL either on an ad hoc basis or via a script.
If ORMs are enough for you, that's great. In my experience, they aren't enough for me and I need something with the flexibility of SQL.
They basically implement something along the lines of
for each book in DB
get the reader line mapped by the book_reader table
This takes about seven lines of code and forces the program to call the database n+1 times (where n is the number of books). SQL can express the same question in a single query. If there are a small number of readers and a lot of books, this can significantly reduce the request/reply overhead. I suspect that Propel would also have trouble with questions that would require a left join. I suspect that it would require the same kind of application level mapping of something that the database already implements.
The freedom of press in the Bill of Rights refers to the freedom of people to print things (which was done with a printing press). It confers *no* special privileges to journalists. The freedom of press is a basic human right, like exercising religion, freedom of speech, freedom to assemble peacably, and freedom to petition the government. A special right for journalism would have had a separate amendment rather than being shoehorned into a list of basic human freedoms.
If bloggers want the privileges awarded to journalists by the Bill of Rights, then bloggers need to maintain the professional standards of journalists. The Bill of Rights does not offer any special privileges to journalists. Non-journalists have just as much right to speak and print as journalists.
Anyway, to get back to the point of the thread. If this blog journalist wants to be treated like a journalist, then the blog journalist should act like a journalist and not like a blogger. We agree on that. The disagreement is that this blog journalist's desire to be a journalist some how binds non-journalist bloggers to act like journalists. It's not the blogging that requires the professionalism; it's the journalism.
A press pass is a professional courtesy for a journalist. This person did not act like a professional journalist and therefore does not deserve that courtesy. That has nothing to do with blogging.
Slashdot Mirror
http://www.nber.org.nyud.net/~confer/2008/EEEs08/kotchen.pdf
I found that at the wikipedia link: http://en.wikipedia.org/wiki/Daylight_saving_time
It does discuss how they used the existing Indiana counties that were already DST to control for non-DST changes between the sample years. The thing that gets me is that it states but does not explain that DST impacts heating/cooling. As best I can tell, the idea is that people are more likely to turn on the heat or A/C when awake. I wish that they had investigated that more.
It's also worth noting that the study says that DST does save electricity in the spring. It's the fall when DST has the negative effect overall. That suggests to me that we should end DST earlier if we want to maximize the electricity benefits.
The Liberal Party of Australia is actually economically laissez-faire and socially conservative--the exact opposite of the US.
Btw, I would argue that liberal socialism involves a trade off between equality of opportunity and freedom of opportunity. It limits the latter to increase the former.
1. Pharmaceuticals are only covered by patent; software is also covered by copyright. With pharmaceuticals, the only protection that a company has against another company copying their drug is the patent. That's why there is a flourishing generics market. By contrast, software binaries are covered by copyright.
2. If a company comes up with a new drug that has the same effect as an existing drug, it's not covered by the pharmaceutical patent. See Cialis and Levitra as examples of this (both are primarily billed for the same problem as Viagra). Software patents are generally applied to others with the same idea rather than the same implementation (implementation is already covered under copyright, so patenting it is somewhat redundant).
3. Pharmaceuticals require clinical trials to be sold. These clinical trials are the majority of the cost of pharmaceutical development. Software seldom has the same level of testing requirements, and when it does, software is covered by copyright. A reimplementation that evades copyright would also require new testing (in the same way that Cialis and Levitra required separate testing from Viagra).
I would be fine with a software patent that was strictly limited to a specific implementation. However, such a patent is almost entirely redundant to the existing copyright protection. It's just as expensive (if not more expensive) to reverse engineer software as it is to write it new. In writing new, you only need to understand how what you wrote works; in reverse engineering, you need to understand how what they wrote works and how what you wrote works (and some way of verifying that those are the same).
In terms of pharmaceuticals, I actually think that the correct thing would be to rename the protection from patent to copyright. The pharmaceutical copyright could work as follows:
1. Some entity registers the copyright for a drug recipe's application to a specific problem. That registration requires a clear plan for starting clinical trials within one year.
2. That application is protected by copyright throughout the clinical trial period. If the clinical trials need to be extended for further investigation, then the protection extends as well.
3. Once the clinical trials complete and the FDA signs off, the copyright protection will last for seven to ten years (about the length of time protection extends after clinical trials now). After that, the drug goes into the public domain.
Pharmaceutical copyright would better describe this than patent and this program would better serve the needs of both consumers and companies. Currently companies under tremendous pressure to bring drugs out of clinical trials as soon as possible. The clock is already ticking on the expiration of patent protection. By changing to a system where the patent clock starts when clinical trials end, it removes the fear of losing revenues. This will allow companies to more naturally balance the risk of waiting (where other drugs can enter the market and where the company needs to pay interest on the money borrowed to perform the clinical trials) with the risk of being sued (because the drug has undesirable side effects).
That would require the patent office to be able to find 12,000 people in the US each year that want to be patent examiners.
I still think that the solution is to eliminate the idea of patents being "granted" after a review by the patent office. Instead they should change to a system where patent applications are recorded and only reviewed when someone tries to enforce the patent. At that time, the defendant can be responsible for doing the prior art search and the patent office only needs to be responsible for *evaluating* the claims. It's a much stronger model.
The current patent examiners have no incentive to deny patents, as their productivity is entirely measured in patent applications processed. Obviously it is easier to approve an application than it is to find a valid reason to deny it, so approval becomes the default state.
Doing things that way would also push patent applications to be better up front. Under the current system, it makes sense to request the broadest patent that you can. What's the worst that happens? The patent is denied and you refile with more narrow claims. By removing that initial review and weakening the meaning of the patent application being accepted, the system pushes more of that burden on to the appliers. Now they want to write the best patent applications possible so that the patent application survives the more rigorous test of examination by a defendant (who is presumably knowledgeable in the field and highly incented to find reasons why the patent is invalid). Particularly since there now is no second chance for a patent application -- if it's badly written and falls over, refiling now would be too late. People already have competing implementations.
The current system is bad because it disassociates costs and benefits. The cost of the bad patents are borne by the defendants in patent infringement cases, but those defendants are not involved in the original patent application. As such, they have no opportunity to block bad patents even though they have the incentive to do so. Patent examiners benefit from quick resolution of patent applications but bear none of the costs of passing bad patents. Patent applicants actually benefit from their own bad patents.
Fixing this either requires the patent office to bear some of the cost of passing bad patents or it requires replacement of the current system with one which allows defendants to participate in the evaluation of the patent claims. Making the patent office bear the costs of bad patents would be hard to implement and the transition would be expensive. It seems much simpler to change the process so that patents are not awarded but only requested. That moves patent application processing into a realm where productivity goals make sense and allows defendants to participate in the invalidation of patents.
I've had Firefox use several hundred megabytes. In my experience, it becomes unstable between half a gig and a full gig, but it will grow that far under the default install. Part of this may be that I was using a lot of intranet applications that were basically just giant lists of drop down menus. These pages were huge. It's also worth noting that I would leave a browser window open until it crashed (I don't like giving up my current state).
It sounds like the poster wants to get a small number of correct results but instead is getting a large number of incorrect results. In other words, a large number of programmers who claim to be the best turn out not to be so good. How do you handle that? Resume screening, phone screens, and interviews are expensive processes that distract your current expert programmers from their real work. How can you minimize the amount of time that your limited resource (good programmers) wastes on identifying false positives (people who are passed by the initial resume screening process but who really aren't that good).
Also, how do you catch the false negatives? What about the programmers who are not good resume writers or especially social but who can code up a storm. How do you find these people?
What's really wanted here is an entirely different version of the interview process. I.e. instead of having a bunch of human beings evaluate what the prospective programmer says (in speech or written communication), find a way to evaluate coding directly. Maybe something like scriptlance, where you write a specification for software and someone implements it. If the implementation passes your tests (for unit test coverage, functionality, etc.), pass the code on to a real human being to evaluate. If that works out, ask for more code (and maybe pay for the original code or maybe that's just a test).
A car engine needs to spin at times when it isn't delivering power to the wheels. At those times, it may make sense to power the alternator and produce electricity. It makes sense to store that electricity. One way to store it would be to electrolyze some water into hydrogen and oxygen. It may or may not make sense to do that rather than store it in the battery (which might be full or simply less efficient). Presumably the hydrogen process is more efficient than the typical battery; otherwise, why not simply use a bigger battery with an electric engine and skip the hydrogen altogether?
Btw, your argument about the evils of property taxes could apply to practically anything. E.g. "The whole point of income is that its money that is paid to you. Having an income tax is like part of it is not your money but really the government's."
Property taxes make sense for some kinds of spending. In private insurance, the cost of the insurance is based on the value of the property being insured. Some kinds of government spending can be considered public insurance against damage to property; law enforcement and defense come to mind.
There's actually a branch of libertarianism called geolibertarianism that maintains that the only valid tax is the tax on land. Land is unique in that it can't be produced, where most everything else only has value accrue to it based on the labor that someone does to it. A quick link from Google: http://members.aol.com/_ht_a/tma68/geo-faq.htm
Um, not to mention that what's described is essentially academic indentured servitude -- I can see VERY little reason for a company not to exploit any loan recipient who's obliged to work there or suddenly get slammed with $80k of debt, and without even the entry-level worker's minimal ability to negotiate offers from multiple employers and move if conditions are seriously unfavorable, wages for new college graduates would fall through the floor.
How so? There are basically two ways to pay back the money: one, work for the company that loans it to you; two, get a job somewhere else and pay back the money. The second option is the entire current option. What's happened to make it worse? I.e. what about this proposal makes the loan worse for a student?
Note that there is some risk in loaning money to a student before graduation. It's quite possible that the student will simply flunk out. After graduation, that risk goes away. Presumably companies would be better off hiring from the pool of graduates rather than loaning to undergraduates -- assuming that the graduates are available. A valid corporate strategy would be to only hire students who performed well in school supported by loans from other companies.
Advantages to students of this program:
1. No risk of graduating, not getting a job, and still having to pay back the loan. If the company wants to get paid back, it must make a job offer.
2. Increased information about what majors are employable (currently this is entirely on the student to evaluate). If no one is willing to loan money to support a major, then clearly it is not employable. If people are approaching the student recruiting for a particular major, then it is. This kind of program makes choosing a school more like choosing a job, where the companies define the process for the student (rather than students choosing majors based on criteria that the student has to develop).
3. Strong likelihood of internship opportunities. If a company is investing in a student, then the company is likely to want to evaluate its risk as early as possible. This means that the student is likely to get internship offers from the company.
Advantages to companies:
1. Earlier participation in student educational choices.
2. More predictability in student hiring.
3. More rewards for taking risks on students. Currently a company can give a scholarship but has no way of getting a return on that.
4. Flexibility in the face of demand changes. If they don't need to hire, they lose what they've put into the student's education already but have no further obligation.
Disadvantages to companies:
1. Requires more planning as to future hiring needs.
2. Has to take risks based on projected future effectiveness. If a student flunks out or drops out, the company has to wait for the loan to be repaid and loses the employment advantages.
I'm also not convinced that this would end pure mathematics as a major. There are "companies" that are very interested in pure math majors; we call them universities. Universities could use their extra insight into student qualifications to encourage good students to enter mathematics as a field. If the student later abandons academics for a computer programming or finance job, the university can recoup that investment as a loan (currently they would have to give a scholarship to encourage the student and gain nothing if the student later chooses not to pursue academics).
Mathematics is an example of a major where there is a shortage of students. This program is designed to help such majors (by providing a funding mechanism for such majors). This program would likely be bad for majors like English, where there are high numbers of students and low numbers of employment opportunities.
I also find it unlikely that most companies are likely to only pay for engineering courses and not literature courses. In general, I think that most companies are
Abraham Lincoln was simply at the right place to become famous. He was President during the US's most deadly war. Remember that the South broke off based entirely on FUD about what Lincoln might do as President. Eventually he freed the slaves, but only after it became obvious that the South was not going to negotiate until they were broken. In the end, the North won because it had more soldiers and manufacturing (as well as control of the Navy). The turning point of the war was when the South invaded Gettysburg to loot shoes. Then Lincoln unleashed the butchers, Grant and Sherman, until the South was broken.
I think that there are any number of Americans who could be considered greater than Isaac Asimov (who was a prolific writer but hardly changed the world): George Washington Carver, Jonas Salk, Richard Feynman, Elijah McCoy, Albert Einstein, Sam Walton, Henry Ford. However, why the obsession with generals? George Washington is a famous moral example, but is he really someone we want people to emulate otherwise? Henry Ford turned assembly line manufacturing from low paid grunt work into a high paid profession. Jonas Salk cured polio. Elijah "The Real" McCoy defined engineering quality (and I would argue that the phrase "The Real McCoy" has had a greater positive effect than the word robotics). So on and so forth.
I've toyed with ideas about programs that would be more corporately focused. For example, what if student loan recipients were chosen by companies? The company would be on the hook for hiring the student after graduation. The student would be responsible for maintaining good grades in a major approved by the company (note: students would be able to pick the company that offered a major that they wanted). Students who flunk out, change majors (without a new sponsor), or who decide not to work for their sponsor have to pay the loan back. If the company cuts back staff and does not hire the student, then the company eats the loan. If the company hires the student, the company is assumed to have adjusted the student's pay appropriately. After some number of years, the student will finish the loan period and can switch companies without paying back the loan.
Another possibility would be to replace federal grants with corporate tax credits. Companies could pay for a student's tuition and mark it down as taxes paid. Obviously it would be more efficient for a company to pay tuition for a student it would like to hire than someone who is interested in an entirely different field.
A big problem with US education before college is the shortness of the school year. Why not take a page from Germany's book and switch to ten 216 day years in elementary and secondary school (the same 2160 days that come from twelve 180 day years)? Then go to a two year program that could be more general than a university degree (i.e. something like Engineering, Science, or Liberal Arts rather than Electrical Engineering, Physics, or Philosophy) and more specific than the final two years of secondary school currently are. Afterwards, students could go to the regular university with a more consistent and focused presentation. For people who aren't college inclined, they could use those two years in a trade school.
It's also worth noting that in many cases what they want to do with the technology is to extend the experience that currently only works on the internet to the desktop. I.e. much of the advantages of this is not running things on the internet that used to be local but in running things locally that used to be on the internet. For example, Amazon.com has a variety of tools on their website to allow third party sellers to list products. They also have a downloadable application called Amazon Seller Desktop that serves the same purpose. Making the local copy work seamlessly with the remote version is important. In that sense, what you see happening now is more a continuation of the trend from centralized computer to a decentralized mechanism.
A lot of the work that is being done now is making it so that you don't need to care whether the information is only stored remotely or also stored locally. Only stored locally becomes problematic when you want to share with others or when you want to work somewhere else. Remote storage with local caching gives you most of the goods from both worlds. An example of a lost positive is that remote security is still inferior to the pure local solution. Examples of gains from remote storage are the ability to share with others and the ability to switch workstations without making an explicit effort to transfer the information. Examples of gains from local caching are performance and the ability to work offline.
Not every choice needs to be made between extremes. Sometimes the best choice is a hybrid between extremes. Further, there may be multiple best choices depending on the exact application under consideration.
It's interesting that these programs already exist for flood and car insurance (which as you note are avoidable risks) but do not exist for health insurance (which while reducible, is not really avoidable).
In the Safe/Unsafe string example, if there is something about the constructor that could be used to make the string "safe" (e.g. verifying that it is shorter than a certain length), then you may be right, a separate class is better. However, if you are doing complicated business logic outside the constructor to establish safety (e.g. doing multiple database lookups and correlating the results), then you're better off with a naming convention. In that case, the class is not enforcing its own meaning. However, the purpose of classes is to enforce their own meaning. By contrast, a variable name does not enforce its own meaning, but one wouldn't expect it to do so. It does indicate meaning, which is what Apps Hungarian was designed to do.
There are also cases (WSDLs come to mind) where the language simply doesn't support strong typing for some of the cases (e.g. a types enum doesn't indicate that it is a type). In those cases, Hungarian notation helps compensate for a weakness in the language.
Database table names are another example. One should put Type or Status in the table name if that's what it holds. Database tables don't have typing, so the name is where you express that. Class names are another example. You are defining the class, so you name it appropriately. E.g. SafeString rather than Foo or MyClassName.
One big problem now is that they've moved from the original patent system, which required an actual implementation, to the current patent system where people can patent generic ideas. Ideas are cheap. Any idiot can come up with an idea that could work. The hard part is turning an idea into an implementation. I have no problem with people receiving patents on source code (although they would be stupid to do so, as copyright is so much better for the holder when it comes to source code). However, when someone else reverse engineers your behavior with their own source code, they should be able to use it. It's even worse when someone independently does the same thing without being aware of your product. However, with software, that happens all the time.
There are places where the patent system makes sense. For example, with pharmaceuticals, the expensive, difficult part is doing the studies to demonstrate to the FDA that the drug is safe and useful. Without some reward for that over and above the cost of manufacture, new drugs wouldn't be studied. There are still some definite tweaks that could be made. In the current system, any delay in finishing the study reduces the effective length of the patent. Why not start the patent term at the end of the study (and shorten it by the length of the typical study)? Then the pharmaceutical companies would be incented to do the study thoroughly (as they face a liability if there is a problem with the drug) without the counter incentive to do the study quickly to maximize the time that they can charge monopoly rents.
The first amendment is about protecting the rights of normal individuals. A normal individual has the right to speak freely, to print (written speech) freely, to practice one's religion freely, and to peaceably assemble with others (for the redress of grievances). The free press of the first amendment is just as much (if not more) aimed at the printing of handbills as newspapers. It does not establish special investigative rights. It establishes communication rights.
Personally, I don't think that there should be any special investigative rights. It goes back to the whole "who watches the watchers" debate. How do you choose these professionals who have the right to do special investigations that the rest of us cannot? Either the right should be one that everyone can use or it isn't a right.
That's not to say that we shouldn't give the other privileges of which you speak to journalists. However, it should be clear that these are privileges, not rights. There are also some cases (e.g. source confidentiality) where it is clearer to think about the rights of the source rather than those of the journalist.
For some applications, e.g. notes to oneself, errors aren't that critical. Don't bother correcting them. Jot down the idea and delete it when you have time to get back to it and process it formally.
Now, they could be doing this to buy market share for MSN. IMO, if they are, they are doing the wrong thing. IMO, it would be far easier to redirect their internal advantages in Yahoo's favor than to redirect Yahoo's market to MSN.
I wonder what happens with Hotmail versus Yahoo! Mail? I could see that going either way. I don't normally use Hotmail, but I looked at it today and it's not immediately obvious to me which is better. The Hotmail client may be a bit more flexible and fully featured.
It would probably help if advocates of the technology understood this. It doesn't have to be all or nothing. Two alternative solutions can add up to a more powerful solution than either would be alone.
Did they cut their pay? It sounds like they moved from a work as many hours as you want for one rate (salary) to a dual tiered rate where they charge more per hour for each hour over forty in a week. Assuming that employees were working forty-five hour weeks (on average), they would continue to get they same weekly pay now: r == (40 + t * 1.5) * (1 - .15) * r / 40 where t ~ 5 hours. I.e. it used to be that they were paid rate r (with the assumption that they would work an average of forty-five hours per week). If you divide r by forty, you get how much they used to make per hour for a forty hour week (the "base" salary). That rate has gone down 15% to (1 - .15) * r / 40. However, they now get paid time and a half for the hours they work past forty. Assuming they average 44.7 hours per week (with a minimum of forty), they'll actually get paid the same for working the same number of hours.
This may or may not be a good deal for the employees. If they used to work more than forty-five hours per week, it's a good deal. They can cut hours and still get the same pay. If they only work forty hours per week, then they would get paid less now.
The employer (IBM) doesn't like it because it makes the system harder to administrate. Managers now have to think along three axes: productivity per pay unit in the current hour; productivity per pay unit if delayed; cost if delayed. Presumably delay increases productivity per pay unit but also increases cost. There's also benefits to consider (it's actually cheaper to pay five hours of overtime for eight existing employees than to hire one additional person to work forty hours -- health care alone is often more expensive than five hours of overtime).
In a salary based system, the manager does not have to consider if it would be better to send the employee home or to pay over time. The manager can concentrate on results. If it takes the employee fifty hours, that's not the manager's problem, it's the employee's.
IANAL, it may very well be that they can get unemployment (under an implicit assumption that their previous rate of pay was based on forty hours per week even though the parameters of the story suggest that they were working overtime). However, the bare facts of the story (15% base salary cut but time and a half for overtime) does not necessarily mean a real pay cut. In practice, these employees may actually get an increase in pay (if they work forty-five or more hours per week).
1. The identity thief requests credit in your name, with your SSN.
2. The thief uses a variety of scams to transfer money or goods to them. The money comes from the new credit instrument.
3. The issuer of the credit now tries to collect from you. You are then responsible for proving that you did not request or use the credit instrument.
It's not that the SSN is really seen as that secret. It's just the piece that associates two entries for "John Doe" as referring to the same John Doe rather than two separate people who happen to have the same name.
If this same problem is less prevalent in Sweden, I suspect that one of two things is true. Either it is more difficult to get anonymous credit (over the phone or internet or through the mail) in Sweden than in the US, or this problem simply hasn't quite made it there yet.
Now, it's quite reasonable to want to abstract out that communication in your code. For many applications, you won't care about the underlying database implementation and can manage with just create, read, update, and delete operations. That's why Object Relation Mappers (ORMs) exist. It sounds like Symfony (which I haven't used) implements an ORM. A pretty common ORM for Java is Hibernate. An ORM can make things easier for many tasks. However, in practice, one occasionally finds that an ORM does not support asking the questions that actually need answered. SQL does. That's why Hibernate still allows people to write direct SQL.
Another issue is that SQL isn't always used from a programming language. It's quite common to connect to a database directly and query with SQL. It would be very inconvenient to have to write a PHP program every time I wanted to query the database. It's much easier to use SQL either on an ad hoc basis or via a script.
If ORMs are enough for you, that's great. In my experience, they aren't enough for me and I need something with the flexibility of SQL.
You can see an example of a query that's simpler in SQL than in Propel at http://propel.phpdb.org/trac/wiki/Users/Documentation/1.3/ManyToManyRelationships
They basically implement something along the lines of
for each book in DB
get the reader line mapped by the book_reader table
This takes about seven lines of code and forces the program to call the database n+1 times (where n is the number of books). SQL can express the same question in a single query. If there are a small number of readers and a lot of books, this can significantly reduce the request/reply overhead. I suspect that Propel would also have trouble with questions that would require a left join. I suspect that it would require the same kind of application level mapping of something that the database already implements.
The freedom of press in the Bill of Rights refers to the freedom of people to print things (which was done with a printing press). It confers *no* special privileges to journalists. The freedom of press is a basic human right, like exercising religion, freedom of speech, freedom to assemble peacably, and freedom to petition the government. A special right for journalism would have had a separate amendment rather than being shoehorned into a list of basic human freedoms.
Anyway, to get back to the point of the thread. If this blog journalist wants to be treated like a journalist, then the blog journalist should act like a journalist and not like a blogger. We agree on that. The disagreement is that this blog journalist's desire to be a journalist some how binds non-journalist bloggers to act like journalists. It's not the blogging that requires the professionalism; it's the journalism.
A press pass is a professional courtesy for a journalist. This person did not act like a professional journalist and therefore does not deserve that courtesy. That has nothing to do with blogging.