> > A trade deficit as massive as ours is with China is never a good thing. > > If you want to reduce the deficit, stop wasting money.
Hmm, do you know what a trade deficit actually is? Only one of the things your list (the first) makes any sense to me as a way of reducing a trade deficit.
A trade deficit is what happens when your country spends more on its imports than it receives from its exports. It can do this only by either selling its assets (land, companies, central bank holdings of foreign currency etc.), borrowing money from other countries or by transfers of money from other countries (such as aid).
When you run out of things to sell or when being no longer want to lend (or give) as much to you your currency will fall - pushing up the price of imports (like DVD players) and pushing down the cost to everyone else of your exports (like your subsidised agricultural produce). That's exactly what seems to have been happening to the dollar recently.
Of course, a problem with that is then you get the developers writing absolutely shitty insecure code using RPCs and god knows what else expecting you to duplicate their craptastic architecture in a production environment. No.. sorry, I'm NOT going to open NTLM authentication through the firewall asshat. You should've thought of that in your design when you were developing your security requirements. Oh, you didn't do that phase? Sorry, it doesn't go in then.
If your software has got to this stage with basic security mistakes in it then something has gone very badly wrong in how the project has been managed.
Security needs to be thought about during every stage of development. So does usability, support, system administration, performance and whole variety of other issues. Someone needs to be involved right throughout development (and the project as a whole) who thinks about and knows about each of these areas.
If you've got some supposedly completed software with such basic security flaws in it that it can't be used then not only has poor project management wasted a whole load of developers' time but you're also unlikely to be able to fix it by blocking a few open ports. What about all of the buffer overflows, SQL injection or cross-site scripting attacks?
Specialized sysadmins who block the deployment of new and broken software long after it's possible to do anything about it aren't the answer. Properly managed development teams who have the right skills available to them (including sysdmin skills) are the answer.
I'd really like to know since I'm now seriously thinking about using amazon.co.uk next semester.
Hmmm.....I remember ordering books from Amazon in the US in about 1996 at the end of one term (to use in the next one) because it was cheaper than buying in the UK. I don't think that that was true of books published in the UK, though.
What kind of oxymoron is this? What do I get for the extra money? Not a damned thing. What asshole thinks that this has ANY value just because it is added on to a transaction?
Errr....what?
Value Added Tax is a tax on the value you've added to a product. If you buy a widget at 10UKP and sell it at 15UKP then you pay tax on 15-10 = 5 UKP. (Well, in principle; the mechanics are a bit more complicated. You charge VAT on the full price of everything you sell, pay VAT on the full price of everything you buy, take the latter from the former and pay the result to the authorities).
Take your apple example. Why is you eating an apple out of this years harvest a cost to an economy? Well...it's because you eating an apple reduces the amount of a scarce resource (this years apples) - and you eating it means that someone else can't.
Over a short period - short enough that apple growers can't grow more apples - a perfect market economy would balance demand and supply to accurately put a number on the cost of doing this. The more desperate others are to have apples the greater the cost of you taking one away from them - and, also, the higher the price would be.
Over longer periods more apples can be grown. What are the costs then? It's no longer simply because you are taking an apple away from others - more apples can be grown to supply your extra consumption. Instead the cost is now that the extra land and man-hours needed to produce the apple are resources which are no longer available to produce other products. What you'd pay, in this mythical perfectly function economy, represents the cost to others of not having those other products because as a result of the resouces which you have used.
Now go back to bandwidth. If your bandwidth use (and I'm not counting support, etc, here) is a short term spike then that's a cost because it takes bandwidth away from others. If no-one was going to use the bandwidth anyway then it doesn't cost anybody anything at all. That'd be just like you eating an apple that would have only gone rotten anyway. If you cause congestion then is more than zero: the loss to the economy of your bandwidth use is equal to the value of the network use which your use has prevented.
If it's a long term bandwidth use - so that your provider (or their provider) has the chance to increase it's capacity - then the cost of your use is the cost using up resources (engineers, energy, raw materials, etc) in providing extra capacity which could have been used to produce something else.
Obviously no real-life economy stands a hope in hell of actually pricing these things anything like close to their true costs. And no real life firm will ever think about it's pricing this way. In priciple, though, that's where the costs come from.
4. Lost jobs. If you have ever worked in industry, you'd realize that the first cost cut
is with lost jobs. "Robots don't get benefits." Any increase in costs will result in the
loss of unskilled jobs. These new jobs that you speak of will be very few and far
between... not to mention drawn from the skilled labor pool.
This is a rather simplistic analysis...
Think about WHY job losses tend to happen at times of crisis. The potential
for automation will already have been there beforehand as a result of the
steady stream of new technology and processes. Adopting a new technology
like this, however, is very disruptive within an organisation. It involves
jobs losses and maybe changes to structure, too. That makes it politically
very difficult. However, it's much easier to implement something of this
sort and there will be much less resistance if this is done when there is
already some sort of crisis in the organisation. A fall in profits is merely
the trigger and not the underlying cause of the change.
Secondly, suggesting that an increase in the cost of using energy and
machines is going to reduce a firm's demand for labour is a little peculiar.
Surely if using more energy consuming automated machine becomes more
expensive a firm would want to *reduce* it's use of such machines and use
labour in it's place?
Finally you shouldn't forget that most economies are heavily biased against
using labour because of the tendency to levy high taxes on income for
reasons of fairness or income redistribution. In other words the signal
which the economy uses to tell firms about the relative scarcity of labour
versus energy - the relative price of these two resources - is distorted
and does not correctly represent their relative scarcity. If this were
evened up a bit through, for example, higher taxes on energy and lower taxes
on income then the economy would use its resources more effectively and
actually produce MORE, not less.
You know how stock works, right? Each share you buy is a corresponding piece of the company. Currency works roughly the same way -- each dollar you have is a share in the company that is the US economy; each pound sterling you have is a share in the UK economy.
I'm not sure if this post was serious or not....but I'll answer anyway.
This analogy seems completely wrong...with stock what you own is an entitlement to part of the future revenue stream of a company plus whatever is left after liquidation. Money doesn't entitle you to anything of the sort from the economies which use it.
Just as the stock price of a company is affected by many factors, so too the worth of a currency. One of the chief factors is national debt.
The worth of a unit of currency is whatever you can buy with it. It's not directly related to national debt (which is just another name for government debt). Government debt only affects exchange rates because it affects interest rates: more government debt => greater demand for money => higher interest rates => capital flowing in to the economy => higher demand for the currency => currency appreciation.
Exchange rates (well, freely floating ones anyway) at any particular time will depend on a lot of things but it seems that market exchange rates can't get ever further from purchasing power parity exchange rates (to be precise it has been found that market and PPP exchange rates are cointegrated). For an obvious example consider what would happen if prices in the UK stayed constant whilst inflation caused all Eurozone prices to double. How many Euros would you expect to get for a pound afterwards? Answer: twice as many.
The high performer, the UK, would end up in effect underwriting Italy's national debt.
Nah. If you lent some money to Italy's government last week then it's the Italian government's responsibility to pay you back this week just as much as it was last week. Consider, though, what would happen if the Italians suddenly started borrowing more and more. By increasing the demand for loans they would tend push Euro-wide interest rates upwards (pulling investment funds away from other investment markets in the process) rather than just their *own* interest rates as they would have done a couple of years ago. Similarly poor management in just one country which would previously just caused inflation in that one country will now cause Euro-wide inflation. The Euro states have become dependendent on each others' good behaviour for monetary stability.
That's one of the biggest reasons why I'm glad that the UK isn't in the Euro - at least not yet. There's just/too much/ national self interest around - especially when there are elections coming up (which within the EU's 15 members is most of the time).
If they had any at all they'd know that even if they removed all the direct connections with a sanctioned country connects could be made through any other country. The internet cant be policed at borders and service that claims it can do that is just lying to itself, avoid things like that is trivial.
Err....could I just ask; what are you talking about?
The data protection rules have nothing to do with the internet as such. They don't restrict 'data' in the general techy sense. The rules restrict what an organisation in the EU can do with information it has collected about individuals - names, addresses, buying history, etc. For example you aren't allowed to arbitrarily pass it on to anyone you choose.
IIRC, the EU wants to restrict EU organisations so that they can't pass data covered by these laws on to organisations outside the EU which are not subject to 'adequate' constraints on what they can do with the data.
This is not uncontentious. If by 'prosperous' you mean 'GDP per capita according to official accounts' then the US comes either second or fifth (in 1999) depending on how you do the currency conversion. Luxembourg comes first.
Either way the big problem with this discussion (and with the representation of economics in the media) is the idea that economic output is a measure of how well the economy is doing.
The purpose of the economy is not to produce as much output as possible but to satisfy people's preferences. Output is important - people typically prefer more goods to less - but its not the whole thing. If you would prefer to work 30 hours instead of 40 (or 60 or whatever) despite receiving a pay cut that adequately compensates the employer then the economy is working inefficiently if it doesn't let this happen.
Economics as an academic subject has recognised this for many decades. Why are the media, politicians and the population at large so slow to catch up?
The BSD license gives the code away for anyone to do with as they will, relying on the kindness of others to repay the gift in terms of giving code back to the community and not abusing it.
Yes..but how do you define 'abusing it'? Anyone who buts their code under a BSD licence probably knows what that implies. If it gets reused by a corporation in order to make a profit then it's not abuse - it's something the author had intentionally made possible.
This maybe isn't strictly an answer to the question, but if you're interested in the future of threading under free Unixes in general it might be worth looking at the direction FreeBSD is heading in.
They plan to use a scheduler activations based system, instead of the process-per-thread, LWP or user space thread model. I suspect it to be quite a way off before a working version is included in a FreeBSD release but it does look very promising.
See
http://people.FreeBSD.org/~jasone/kse/.
You may not copy, modify, sublicense, or distribute the Program except as expressly provided under this License. Any attempt otherwise to copy, modify, sublicense or distribute the Program is void, and will automatically terminate your rights under this License.
So this is where RMS' "forgiveness" thing comes in. He is explicitly regranting KDE the license to use anything copyrighted by the Free Software Foundation. If he hadn't done that, then the integrity of the GPL would have been questioned. Granted, he could have used a different word, but that's still no excuse for the KDE people to get up in arms over word choice.
How's this argument:
Suppose I give you a program and a licence to use it where that licence is the GPL. You then break that licence and this clause is triggered. My licence to you becomes void.
You later cease to do things which would violate the licence (if you had one). Then you go to someone else who gives you a copy of the same program and a licence to use it under the GPL. You now have a new licence---one which ISN'T void because you haven't broken it.
Finally I think it's perfectly reasonable for people in the KDE project to complain when they are the subject of attacks over things which are of no practical significance whatsoever. RMS appears to be on an ego trip, plain and simple. I'm having trouble imagining that he is doing this out of regard for 'freedom' or 'free software'.
Re:Yeah C makes a poor OO language
on
Qt Going GPL
·
· Score: 1
Using plain C, structs and macros you can hack out a very inelegant version of an OO-like language, true. But why would you want to? I see - you must have way too much time on your hands and enjoy fixing mechanical mistakes that would never have appeared in a higher-level proper OO language like C++ in the first place.
The only time I've found a reason to use C for an OO program was when I was first trying out OO design. The C compiler I had was very nice and very fast on my 33MHz proccessor. The C++ compiler was VERY slow and buggy. Oh, and I didn't know C++ very well:-)
I suspect that the second of those is the primary reason for people doing it nowadays...
C versus C++ isn't much of a war, people. C is a clean procedural language, C++ is a dirty OO language.
The only thing going for C++, when you think about it, is easy compatibility with C. That's all that stopped it losing to a cleaner OO language like Eiffel.
It is a bit of a pity that open source projects have tended to ignore higher level languages and stick rigidly to C... C++ is at least a bit better, especially if you have a need to work at a low level sometimes.
Compatibility is always going to be an issue, though. There is a huge body of free C code out there which projects such as KDE draw on heavily (consider, for instance, OpenSSL, libjpeg, libpng and xdm). It's possible to write bindings for other languages, like the Qt bindings for Python, say, but that barrier is always going to be there and there's no guarantee that the bindings will always be well maintained, up to date or clean when the main project doesn't have the language as one of its priorities.
Finally, I'd far rather have OO code written as cleanly as possible in a 'dirty' OO language like C++ than have OO code written as cleanly as possible (ie, not very cleanly at all) in C. I've always found trying to write OO C to be a somewhat painful and messy experience.
Re:KDE is already GPL.
on
Qt Going GPL
·
· Score: 1
ne advantage of GPL over QPL is it's hydra-headedness.. you could kill Qt before by putting Troll out of business.. but now Qt is immortal.
Not true. From the Free Qt Foundation announcement:
Should Troll Tech ever discontinue the Qt Free Edition for any reason including, but not limited to, a buy-out of Troll Tech, a merger or bankruptcy, the latest version of the Qt Free Edition will be released under the BSD license.
Furthermore, should Troll Tech cease continued development of Qt, as assessed by a majority of the KDE Free Qt Foundation, and not release a new version at least every 12 months, the Foundation has the right to release the Qt Free Edition under the BSD License.
When QT becomes GPL, KDE won't be able to be put in debian only because of this. Before this can be done, any program has to fully under the GPL or fully under QPL.
I don't think that's true. You can quite easily have a program which uses Qt (under the QPL), an LGPL library (eg, the KDE libraries) and is itself licenced under a variety of other possible licences (eg BSD, GPL with exception caluse for linking to Qt, etc) and remain definitely legal.
Programs which main-code is qpl and using libraries under gpl are still illegal, until the program itself is gpl.
Such programs might infringe the GPL author's copyright, so there may indeed be a problem if he chose to sue. Assuming, that is, that dynamic linking produces a 'derived work' which I personally don't think is true. Since the only significant code out there under the QPL is Qt which is now dual licenced I don't think this is a problem anyway...
KDE itself is a mixture of licences. The libraries are LGPL, so you can link a BSD licenced program to the KDE libraries and Qt with no problems. Much of the rest of KDE is GPL with some portions BSD licenced. Since anything which is BSD licenced satisifies the conditions of the GPL (in fact, the BSD licence is such that I believe you can relicence code under the GPL (or a closed source licence) if you want to) that doesn't present a problem either.
There shouldn't now be any problem with Debian including all of KDE now, as far as I can see. Previously there was no problem with including KDE libraries anyway, but they didn't all the same.
Most companies using BSD never contribute back to the community. They hoard the software and might dribble back some trivial fix. Usually they keep the good stuff for themselves and never help the community evolve.
My point is that if a company doesn't/want/ to give back then they simply won't use a GPL system. If the BSDs were GPLd they simply *wouldn't use it*...since they are they can and, as a result, you'd expect there to be more BSD companies not releasing their changes compared to Linux. That doesn't mean that the BSDs are losing out compared to what would happen if they used the GPL.
That is one of the main reasons BSD is stagnating.
I don't see any sign of it stagnating. They are all excellent OSs with their own strengths. What makes you think they are stagnating?
Um, not wanting to sound like some GPL-fanatic, but is this really a problem?
No, probably not. I'm just expressing surprise not attacking them.
I think I don't quite see what's worth sighing about here.
I'm just sighing because there is a really excellent OS out there that seems to be getting ignored for largely hype-related reasons...especially when that OS (well, those OSs) seem to be quite careful to consider embedded systems when decisions are being made.
Or we would never, ever see the benefits of their work. The GPL isn't all bad, guys.
What makes you think that? If they hadn't intended to make their work available then they wouldn't have chosen a GPLd system. If they'd gone for a BSD system then why does that necessarily imply that they would have suddenly changed their minds on that? It's far easier to keep up to date if you feed your generally applicable changes back to the original project.
I wonder how they'll cope if they have to write a driver for hardware under NDA...I'm not convinced that a binary only kernel module wouldn't be considered part of a 'derived work'.
However, one thing to remember (and something that Linus has mentioned recently) is that for many tasks, the Linux kernel's latency is low enough that you don't need a real-time kernel. (Of course, this could change when you're talking about an embedded system for cars or industrial robots...)
And processingly audio processing, seemingly, judging by recent discussions on linux-kernel. I got the impression that Linux's latency isn't actually all that low without a few hacky patches.
I'm a little surprised that so many people of this sort go for Linux rather than one of the BSDs because of the licence. Do/you/ want to be distributing source code with your new mobile phone? FreeBSD at least always seems to be quite sensitive to the needs of companies who don't...to the point of trying very hard to keep GPL licenced code out of critical parts of the system.
[sigh] I guess it's largely a case of better marketing/more widespread attention for Linux.
MINOR? Only as minor as downloading warez copies of Windows. KDE IS ILLEGAL TO COPY. How much more major can you get than that?
WTF are you talking about?
Let's see...the KDE libraries are LGPL, so no problem there. kdebase and most of the other KDE packages are all written specifically for KDE and distributed as such by the KDE people, so it's kinda hard to argue that you aren't allowed to link them to Qt (or, to be exact, that implicit permission hasn't been given---assuming that it's even needed which is rather debateable for dynamically linked executables). Or do you really think that people writing free software for Qt would try to deny you the ability to distribute binaries?
So, what's left? Ah yes, kghostview, kdvi and a few other sundry 'kde-ized' applications. You may have an argument to make over those but they hardly constitute very much of KDE. Some people (Trolltech, for instance, hence their rejection of the GPL) would appear to dispute that dynamically linking to a GPL library necessarily produces a derived work and hence the GPL may be irrelevant anyway.
Presumably we should be "punishing" Frustrated for even considering a proprietary sales model -- and yet half the respondants here seem to be advocating one!
I've always considered Stallman's philosophy specious -- it explains how "free software" benefits society, but provides no mechanism through which society may fund the development of such software on a large scale. As such, it's a bit like a perpetual motion machine -- lovely for society to have one, yes, but providing no blueprint for its design.
Apparently I'm not alone among Slashdotters in my my skepticism of the "free software" model?
I doubt it. If you try to charge for support for your open source product then *either* you find a way of maintaining yourself as a monopolist for that support *or* you only get to charge an amount which pays to provide support (including the return on investment you need to pay your investors). If you don't set yourself up as a monopolist then you can't charge a premium and you have no way of cross-subsidising the software. Being a monopolist (or, at least, having a degree of monopoly power; being a mere oligopolist might help) isn't going to viable for everyone.
I just can't see sufficient ability or incentive for companies to use support to fund development like that. The same argument applies to an extent to proprietary add-ons.
The whole problem is a very well known problem in economics. Markets are just not able to cope well with so called 'non-rival' goods---software, books, TV, national defence, lighthouses or anything else where one person's consumption doesn't reduce the amount left for everyone else.
Once a piece of software has been produced it is most economically efficient for the software to be given away to anyone who wants a copy (well, aside from charging the cost of making the copy). This is well known---if Software Inc charges 40USD for its product and someone is willing to pay just 30$ then you could make both parties better off by giving him the software for anything between 0 and 30. Markets just can't do that, though, they don't work that way. The only efficient market price under those circumstances is 0.
At the same time, though, charging the efficient market price fails to provide the correct degree incentive to produce software. The software should be produced if the sum total value to every potential user exceeds the sum total cost. Once developed it should be distribution for free. Again, markets don't work like that, hence they just can't provide the right incentives.
There is no just other than to just 'do the best we can'. Free software is a very good thing where it exists but it can never replace commercial software completely. We just have to accept that---short of government funding of exactly the right amount of the right things (yeah right, like that'd happen; and of course, it has ill effects elsewhere in the economy thanks to the taxes needed)---the software market will always be imperfect.
In short both forms of software have a place. It's unreasonable and unfeasible for free software to take over completely; it's inefficient for commercial software to take over completely.
Forget the dogma. You need to pay your developers---close the source (to non-customers, at least).
He also seems to confuse GPL with LGPL: "If the GPL effectively protected a GPLed library from being used to develop proprietary software, we would allow relicensing Qt under the GPL. But, as I have said, it is not our belief that using a library is making a derivative work. "
The LGPL license was developed to give non GPL compatible software access to GPL libraries. If a library is GPL:ed there is no way of linking (dynamic or static) with that library. E.g. the cygwin library is GPL:ed in Windows, so you can only create GPL:ed software with Cygwin.
He hasn't confused the GPL and the LGPL, he seem to merely be disputing that linking (I presume he means dynamically linking) produces a derived work. If that is correct then you can dynamically link against a GPL library and not be subject to any of the restrictions of the GPL.
In other words: If you believe that linking in that way doesn't include any GPLed code in your product (or that it includes it in a way which is 'fair use') then the GPL is irrelevant to the final product.
Slashdot Mirror
>
> If you want to reduce the deficit, stop wasting money.
Hmm, do you know what a trade deficit actually is? Only one of the things your list (the first) makes any sense to me as a way of reducing a trade deficit.
A trade deficit is what happens when your country spends more on its imports than it receives from its exports. It can do this only by either selling its assets (land, companies, central bank holdings of foreign currency etc.), borrowing money from other countries or by transfers of money from other countries (such as aid).
When you run out of things to sell or when being no longer want to lend (or give) as much to you your currency will fall - pushing up the price of imports (like DVD players) and pushing down the cost to everyone else of your exports (like your subsidised agricultural produce).
That's exactly what seems to have been happening to the dollar recently.
If your software has got to this stage with basic security mistakes in it then something has gone very badly wrong in how the project has been managed.
Security needs to be thought about during every stage of development. So does usability, support, system administration, performance and whole variety of other issues. Someone needs to be involved right throughout development (and the project as a whole) who thinks about and knows about each of these areas.
If you've got some supposedly completed software with such basic security flaws in it that it can't be used then not only has poor project management wasted a whole load of developers' time but you're also unlikely to be able to fix it by blocking a few open ports. What about all of the buffer overflows, SQL injection or cross-site scripting attacks?
Specialized sysadmins who block the deployment of new and broken software long after it's possible to do anything about it aren't the answer. Properly managed development teams who have the right skills available to them (including sysdmin skills) are the answer.
Hmmm.....I remember ordering books from Amazon in the US in about 1996 at the end of one term (to use in the next one) because it was cheaper than buying in the UK. I don't think that that was true of books published in the UK, though.
Errr....what?
Value Added Tax is a tax on the value you've added to a product. If you buy a widget at 10UKP and sell it at 15UKP then you pay tax on 15-10 = 5 UKP. (Well, in principle; the mechanics are a bit more complicated. You charge VAT on the full price of everything you sell, pay VAT on the full price of everything you buy, take the latter from the former and pay the result to the authorities).
Take your apple example. Why is you eating an apple out of this years harvest a cost to an economy? Well...it's because you eating an apple reduces the amount of a scarce resource (this years apples) - and you eating it means that someone else can't.
Over a short period - short enough that apple growers can't grow more apples - a perfect market economy would balance demand and supply to accurately put a number on the cost of doing this. The more desperate others are to have apples the greater the cost of you taking one away from them - and, also, the higher the price would be.
Over longer periods more apples can be grown. What are the costs then? It's no longer simply because you are taking an apple away from others - more apples can be grown to supply your extra consumption. Instead the cost is now that the extra land and man-hours needed to produce the apple are resources which are no longer available to produce other products. What you'd pay, in this mythical perfectly function economy, represents the cost to others of not having those other products because as a result of the resouces which you have used.
Now go back to bandwidth. If your bandwidth use (and I'm not counting support, etc, here) is a short term spike then that's a cost because it takes bandwidth away from others. If no-one was going to use the bandwidth anyway then it doesn't cost anybody anything at all. That'd be just like you eating an apple that would have only gone rotten anyway. If you cause congestion then is more than zero: the loss to the economy of your bandwidth use is equal to the value of the network use which your use has prevented.
If it's a long term bandwidth use - so that your provider (or their provider) has the chance to increase it's capacity - then the cost of your use is the cost using up resources (engineers, energy, raw materials, etc) in providing extra capacity which could have been used to produce something else.
Obviously no real-life economy stands a hope in hell of actually pricing these things anything like close to their true costs. And no real life firm will ever think about it's pricing this way. In priciple, though, that's where the costs come from.
This is a rather simplistic analysis...
Think about WHY job losses tend to happen at times of crisis. The potential for automation will already have been there beforehand as a result of the steady stream of new technology and processes. Adopting a new technology like this, however, is very disruptive within an organisation. It involves jobs losses and maybe changes to structure, too. That makes it politically very difficult. However, it's much easier to implement something of this sort and there will be much less resistance if this is done when there is already some sort of crisis in the organisation. A fall in profits is merely the trigger and not the underlying cause of the change.
Secondly, suggesting that an increase in the cost of using energy and machines is going to reduce a firm's demand for labour is a little peculiar. Surely if using more energy consuming automated machine becomes more expensive a firm would want to *reduce* it's use of such machines and use labour in it's place?
Finally you shouldn't forget that most economies are heavily biased against using labour because of the tendency to levy high taxes on income for reasons of fairness or income redistribution. In other words the signal which the economy uses to tell firms about the relative scarcity of labour versus energy - the relative price of these two resources - is distorted and does not correctly represent their relative scarcity. If this were evened up a bit through, for example, higher taxes on energy and lower taxes on income then the economy would use its resources more effectively and actually produce MORE, not less.
You know how stock works, right? Each share you buy is a corresponding piece of the company. Currency works roughly the same way -- each dollar you have is a share in the company that is the US economy; each pound sterling you have is a share in the UK economy.
I'm not sure if this post was serious or not....but I'll answer anyway.
This analogy seems completely wrong...with stock what you own is an entitlement to part of the future revenue stream of a company plus whatever is left after liquidation. Money doesn't entitle you to anything of the sort from the economies which use it.
Just as the stock price of a company is affected by many factors, so too the worth of a currency. One of the chief factors is national debt.
The worth of a unit of currency is whatever you can buy with it. It's not directly related to national debt (which is just another name for government debt). Government debt only affects exchange rates because it affects interest rates: more government debt => greater demand for money => higher interest rates => capital flowing in to the economy => higher demand for the currency => currency appreciation.
Exchange rates (well, freely floating ones anyway) at any particular time will depend on a lot of things but it seems that market exchange rates can't get ever further from purchasing power parity exchange rates (to be precise it has been found that market and PPP exchange rates are cointegrated). For an obvious example consider what would happen if prices in the UK stayed constant whilst inflation caused all Eurozone prices to double. How many Euros would you expect to get for a pound afterwards? Answer: twice as many.
The high performer, the UK, would end up in effect underwriting Italy's national debt.
Nah. If you lent some money to Italy's government last week then it's the Italian government's responsibility to pay you back this week just as much as it was last week. Consider, though, what would happen if the Italians suddenly started borrowing more and more. By increasing the demand for loans they would tend push Euro-wide interest rates upwards (pulling investment funds away from other investment markets in the process) rather than just their *own* interest rates as they would have done a couple of years ago. Similarly poor management in just one country which would previously just caused inflation in that one country will now cause Euro-wide inflation. The Euro states have become dependendent on each others' good behaviour for monetary stability.
That's one of the biggest reasons why I'm glad that the UK isn't in the Euro - at least not yet. There's just /too much/ national self interest around - especially when there are elections coming up (which within the EU's 15 members is most of the time).
Err....could I just ask; what are you talking about?
The data protection rules have nothing to do with the internet as such. They don't restrict 'data' in the general techy sense. The rules restrict what an organisation in the EU can do with information it has collected about individuals - names, addresses, buying history, etc. For example you aren't allowed to arbitrarily pass it on to anyone you choose.
IIRC, the EU wants to restrict EU organisations so that they can't pass data covered by these laws on to organisations outside the EU which are not subject to 'adequate' constraints on what they can do with the data.
This is not uncontentious. If by 'prosperous' you mean 'GDP per capita according to official accounts' then the US comes either second or fifth (in 1999) depending on how you do the currency conversion. Luxembourg comes first.
Either way the big problem with this discussion (and with the representation of economics in the media) is the idea that economic output is a measure of how well the economy is doing.
The purpose of the economy is not to produce as much output as possible but to satisfy people's preferences. Output is important - people typically prefer more goods to less - but its not the whole thing. If you would prefer to work 30 hours instead of 40 (or 60 or whatever) despite receiving a pay cut that adequately compensates the employer then the economy is working inefficiently if it doesn't let this happen.
Economics as an academic subject has recognised this for many decades. Why are the media, politicians and the population at large so slow to catch up?
And last time /I/ checked there was a free (but limited, IIRC) version for students.
Rather a lot of them. In a (Free|Net|Open)BSD specific format and package manager which I, personally, prefer.
You can also install and use RPM and Linux binaries, if you wish.
I would wonder how many commercial apps will run w/o a hitch (ala Oracle, Sybase, Star Office, Netscape servers).
Well, Sybase and Star Office run just fine. As does, for instance, the Linux netscape binaries. I haven't tried Oracle or any Netscape servers.
But then I suppose the argument would be that glibc and binary compatibility is the great equalizer???
I have glibc on my FreeBSD machine - in /compat/linux/lib. Linux apps work just fine with it.
Yes..but how do you define 'abusing it'? Anyone who buts their code under a BSD licence probably knows what that implies. If it gets reused by a corporation in order to make a profit then it's not abuse - it's something the author had intentionally made possible.
This maybe isn't strictly an answer to the question, but if you're interested in the future of threading under free Unixes in general it might be worth looking at the direction FreeBSD is heading in. They plan to use a scheduler activations based system, instead of the process-per-thread, LWP or user space thread model. I suspect it to be quite a way off before a working version is included in a FreeBSD release but it does look very promising. See http://people.FreeBSD.org/~jasone/kse/.
So this is where RMS' "forgiveness" thing comes in. He is explicitly regranting KDE the license to use anything copyrighted by the Free Software Foundation. If he hadn't done that, then the integrity of the GPL would have been questioned. Granted, he could have used a different word, but that's still no excuse for the KDE people to get up in arms over word choice.
How's this argument:
Suppose I give you a program and a licence to use it where that licence is the GPL. You then break that licence and this clause is triggered. My licence to you becomes void.
You later cease to do things which would violate the licence (if you had one). Then you go to someone else who gives you a copy of the same program and a licence to use it under the GPL. You now have a new licence---one which ISN'T void because you haven't broken it.
Finally I think it's perfectly reasonable for people in the KDE project to complain when they are the subject of attacks over things which are of no practical significance whatsoever. RMS appears to be on an ego trip, plain and simple. I'm having trouble imagining that he is doing this out of regard for 'freedom' or 'free software'.
The only time I've found a reason to use C for an OO program was when I was first trying out OO design. The C compiler I had was very nice and very fast on my 33MHz proccessor. The C++ compiler was VERY slow and buggy. Oh, and I didn't know C++ very well :-)
I suspect that the second of those is the primary reason for people doing it nowadays...
The only thing going for C++, when you think about it, is easy compatibility with C. That's all that stopped it losing to a cleaner OO language like Eiffel.
It is a bit of a pity that open source projects have tended to ignore higher level languages and stick rigidly to C... C++ is at least a bit better, especially if you have a need to work at a low level sometimes.
Compatibility is always going to be an issue, though. There is a huge body of free C code out there which projects such as KDE draw on heavily (consider, for instance, OpenSSL, libjpeg, libpng and xdm). It's possible to write bindings for other languages, like the Qt bindings for Python, say, but that barrier is always going to be there and there's no guarantee that the bindings will always be well maintained, up to date or clean when the main project doesn't have the language as one of its priorities.
Finally, I'd far rather have OO code written as cleanly as possible in a 'dirty' OO language like C++ than have OO code written as cleanly as possible (ie, not very cleanly at all) in C. I've always found trying to write OO C to be a somewhat painful and messy experience.Not true. From the Free Qt Foundation announcement:
I don't think that's true. You can quite easily have a program which uses Qt (under the QPL), an LGPL library (eg, the KDE libraries) and is itself licenced under a variety of other possible licences (eg BSD, GPL with exception caluse for linking to Qt, etc) and remain definitely legal.
Programs which main-code is qpl and using libraries under gpl are still illegal, until the program itself is gpl.
Such programs might infringe the GPL author's copyright, so there may indeed be a problem if he chose to sue. Assuming, that is, that dynamic linking produces a 'derived work' which I personally don't think is true. Since the only significant code out there under the QPL is Qt which is now dual licenced I don't think this is a problem anyway...
KDE itself is a mixture of licences. The libraries are LGPL, so you can link a BSD licenced program to the KDE libraries and Qt with no problems. Much of the rest of KDE is GPL with some portions BSD licenced. Since anything which is BSD licenced satisifies the conditions of the GPL (in fact, the BSD licence is such that I believe you can relicence code under the GPL (or a closed source licence) if you want to) that doesn't present a problem either.
There shouldn't now be any problem with Debian including all of KDE now, as far as I can see. Previously there was no problem with including KDE libraries anyway, but they didn't all the same.
My point is that if a company doesn't /want/ to give back then they simply won't use a GPL system. If the BSDs were GPLd they simply *wouldn't use it*...since they are they can and, as a result, you'd expect there to be more BSD companies not releasing their changes compared to Linux. That doesn't mean that the BSDs are losing out compared to what would happen if they used the GPL.
That is one of the main reasons BSD is stagnating.
I don't see any sign of it stagnating. They are all excellent OSs with their own strengths. What makes you think they are stagnating?
No, probably not. I'm just expressing surprise not attacking them.
I think I don't quite see what's worth sighing about here.
I'm just sighing because there is a really excellent OS out there that seems to be getting ignored for largely hype-related reasons...especially when that OS (well, those OSs) seem to be quite careful to consider embedded systems when decisions are being made.
What makes you think that? If they hadn't intended to make their work available then they wouldn't have chosen a GPLd system. If they'd gone for a BSD system then why does that necessarily imply that they would have suddenly changed their minds on that? It's far easier to keep up to date if you feed your generally applicable changes back to the original project.
I wonder how they'll cope if they have to write a driver for hardware under NDA...I'm not convinced that a binary only kernel module wouldn't be considered part of a 'derived work'.
And processingly audio processing, seemingly, judging by recent discussions on linux-kernel. I got the impression that Linux's latency isn't actually all that low without a few hacky patches.
I'm a little surprised that so many people of this sort go for Linux rather than one of the BSDs because of the licence. Do /you/ want to be distributing source code with your new mobile phone? FreeBSD at least always seems to be quite sensitive to the needs of companies who don't...to the point of trying very hard to keep GPL licenced code out of critical parts of the system.
[sigh] I guess it's largely a case of better marketing/more widespread attention for Linux.
WTF are you talking about?
Let's see...the KDE libraries are LGPL, so no problem there. kdebase and most of the other KDE packages are all written specifically for KDE and distributed as such by the KDE people, so it's kinda hard to argue that you aren't allowed to link them to Qt (or, to be exact, that implicit permission hasn't been given---assuming that it's even needed which is rather debateable for dynamically linked executables). Or do you really think that people writing free software for Qt would try to deny you the ability to distribute binaries?
So, what's left? Ah yes, kghostview, kdvi and a few other sundry 'kde-ized' applications. You may have an argument to make over those but they hardly constitute very much of KDE. Some people (Trolltech, for instance, hence their rejection of the GPL) would appear to dispute that dynamically linking to a GPL library necessarily produces a derived work and hence the GPL may be irrelevant anyway.
I've always considered Stallman's philosophy specious -- it explains how "free software" benefits society, but provides no mechanism through which society may fund the development of such software on a large scale. As such, it's a bit like a perpetual motion machine -- lovely for society to have one, yes, but providing no blueprint for its design.
Apparently I'm not alone among Slashdotters in my my skepticism of the "free software" model?
I doubt it. If you try to charge for support for your open source product then *either* you find a way of maintaining yourself as a monopolist for that support *or* you only get to charge an amount which pays to provide support (including the return on investment you need to pay your investors). If you don't set yourself up as a monopolist then you can't charge a premium and you have no way of cross-subsidising the software. Being a monopolist (or, at least, having a degree of monopoly power; being a mere oligopolist might help) isn't going to viable for everyone.
I just can't see sufficient ability or incentive for companies to use support to fund development like that. The same argument applies to an extent to proprietary add-ons.
The whole problem is a very well known problem in economics. Markets are just not able to cope well with so called 'non-rival' goods---software, books, TV, national defence, lighthouses or anything else where one person's consumption doesn't reduce the amount left for everyone else.
Once a piece of software has been produced it is most economically efficient for the software to be given away to anyone who wants a copy (well, aside from charging the cost of making the copy). This is well known---if Software Inc charges 40USD for its product and someone is willing to pay just 30$ then you could make both parties better off by giving him the software for anything between 0 and 30. Markets just can't do that, though, they don't work that way. The only efficient market price under those circumstances is 0.
At the same time, though, charging the efficient market price fails to provide the correct degree incentive to produce software. The software should be produced if the sum total value to every potential user exceeds the sum total cost. Once developed it should be distribution for free. Again, markets don't work like that, hence they just can't provide the right incentives.
There is no just other than to just 'do the best we can'. Free software is a very good thing where it exists but it can never replace commercial software completely. We just have to accept that---short of government funding of exactly the right amount of the right things (yeah right, like that'd happen; and of course, it has ill effects elsewhere in the economy thanks to the taxes needed)---the software market will always be imperfect.
In short both forms of software have a place. It's unreasonable and unfeasible for free software to take over completely; it's inefficient for commercial software to take over completely.
Forget the dogma. You need to pay your developers---close the source (to non-customers, at least).
The LGPL license was developed to give non GPL compatible software access to GPL libraries. If a library is GPL:ed there is no way of linking (dynamic or static) with that library. E.g. the cygwin library is GPL:ed in Windows, so you can only create GPL:ed software with Cygwin.
He hasn't confused the GPL and the LGPL, he seem to merely be disputing that linking (I presume he means dynamically linking) produces a derived work. If that is correct then you can dynamically link against a GPL library and not be subject to any of the restrictions of the GPL.
In other words: If you believe that linking in that way doesn't include any GPLed code in your product (or that it includes it in a way which is 'fair use') then the GPL is irrelevant to the final product.