Lots of/. articles contain links to further information. Those that claim to show something of interest sometimes even contain calculations. Or at least, they used to in the old days. Possibly the standard here has sunk even further.
Regardless of that, I'm not going to believe him until he tells me what his results are based on.
1. How did you assign costs to CPU time ?
2. Where did you get your figures on the computational cost of a search, and its relation to the size of the web ?
3. Where did you get your figures on the size of the web ?
4. Can you supply a reference for the figure of 1 cent per impression for banner ads ? and do you know how much google gets for its text ads ?
Only temporary variable names are lost, unless the author used a scrambling program. Class, field and method names are preserved, because the link process for Java occurs at load time, and they are needed for this.
Re:What can be done? Nothing.
on
More On Tragedy
·
· Score: 2
One of the characteristics of resistance groups - and make no mistake that is how these people see themselves - is that if you start out with 100 revolutionaries and execute 90, when you're finished you don't have 10, or even 100, but 10,000, or even more.
1. Tell them if they're too busy to use the system, you're too busy to write it for them. Seriously, you need to reach an agreement with clients when the project starts on how much work you expect them to do.
2. You need to do a proper analysis - interview the customer's staff, write use cases. This is where XP falls down - how are you supposed to know the problem domain ?
is that all average programmers, without exception, believe they belong to the tiny minority of the truly talented.
In some respects exceptional confidence in your own abilities is a more reliable indicator of inability, because you lack the ability needed to judge your own work critically. Noone truly brilliant I have met or heard of in any field believed themselves to be any better than average.
There are also downsides to having the state supply food and housing, which vary depending on how you administer such a scheme, and generally reflect the fact that it would have to be funded from tax receipts.
If the state owns and administers housing stock on behalf of those who would otherwise be homeless, there would be a lack of ownership and care over the houses. The government would not want to pay for nice houses or maintanenace, and the tenants, not owning the capital, would not want to either. This is, in fact, what happens in Britain. Similarly for food. If people go and collect their rations every day with no control over the contents, you are not going to get haute cuisine.
On the other hand, if the state merely agrees to fund individual choices as to where they want to live or what they want to eat, a price limit would have to be agreed. This is essentially like the voucher system used in some places for paying for education. It would work OK, except that the "best" providers of food and housing would price themselves out of the scheme, by refusing to accept voucers at all (stores in Britain do this to assylum seekers who are given tokens in place of cash, and do stores in the US with food stamps), or by raising their prices above what can be payed for with them. The value in many goods comes only from their exclusiveness, and anyway in many cases supply is limited and prices must be raises if demand rises.
So, do away with vouchers and just give people the cash. This is actually pretty close to being a basic income gaurantee - something I think is quite a good idea. The issues then are twofold. Firstly the redistributive nature of the scheme is now clear. Food and housing don't appear from nowhere after all - you're taking resources from one person and giving them to another. Secondly, do you mind that people may not spend their money on what you wanted to give them ? are you going to stop them from investing it in the stock market, losing it, and then having to beg ?
The book being reviewed is about capital. Your comment is about money. Capital is productive property, and as such can consist of ideas, land, machinery and even personal skills in some views.
Money is a reliable store of value that is accepted in exchange for other good.
Both capital and money are social realities: there's nothing "imaginery" about it. If I try to pay for dinner using a turnip, patent water or enclose the center of London to graze my cows I'll soon discover that imagination is not enough.
If there was ever a reason to abolish the first amendment, it's the kiddie porn issue.
Its unclear why. Its established precedent everywhere that I know of that freedom of speach protection does not apply in cases of obscenity. To an even greater extent, you have no freedom of speech where you're revealing that a criminal act took place (in this case statutory or actual rape of sexual assault). Thus kiddie porn (viewing, knowingly distributing, or making) is already illegal.
On the ethical, rather than the legal, level, there are two arguments for making child pornography illegal. The first, and best, is that making (some of) it involves rape (statutory and actual), and sexual assault, and that these are crimes in themselves. The second, and weakest, is that being expose to such images is somehow a horrible trauma that adults and especially children who come across it accidentally will suffer harm from.
I was hoping to find a finely considered argument for regulation here, of the kind Lawrence Lessig generally produces. I did not. In fact, I found a short essay practicing what it claims to condemn. It pushes a couple of predictable buttons for the general public - paedophiles and large corporations - in order to scare people into buying the "dangerous information" argument in general, and they goes on to argue that the government should regulate the internet for accuracy because it is different (no attempt to made to explain how) to media for which that would be considered a gross violation of human rights. No attempt is made to explain how to avoid the inevitable corruption and contention involved in such regulation.
I kind of agree, actually, that the public is in danger of poisoning its conciousness with inaccurate and oversimplified information, but thats been going on for years and the mass media have a lot more to answer for than the internet here. I believe our attitude to information is wrong: that we are too happy to sit back and absorb what is fed to us by people we never see and who'd motives we rarely know ? Is it surprising that some of what is fed to us is poison ? That attitude has been encourage by mass media, and in the age of the internet, we need to get over it. Someone suggested a kind of global moderation system: that seems like a good start. Ultimately, people need to accept that it is their responsibility to verify what they are told, and, if certain images disturb them, to control what they see. In my opinion, that means a change in our whole cultural attitude, starting with hwo we are taught in school.
In the end, this is just another attempt to fudge an issue of private responsibility by declaring that the government must take care of us all.
What you describe is certainly what the American patent system has come to, but it was not the intention. The patent office, until recently, researched the claims in a patent with considerable care and accuracy. In principle they still have duty to do so, and frequently reject the initial patent application with examples of possible prior art, usually from other patents. Whats changed is that patent exmaminers are now encouraged, by the way they are payed to pass everything, and the patent office has poor records of the existing prior art for many fields in which they are issuing patents, especially softwate and business models.
I'd tend to agree that civil servants are not the appropriate people to determine obiousness or novelty, but, due to the great expense of challenging a patent, nor are the courts. The existing system allows patent holders to exploit their competitors for licensing fees as long as these remain below the expenditure needed to challenge the patent.
If you're going to have a patent system, and especially if you're going to allow patents in so many fields, wouldn't the appropriate thing to do be to allow challenges by interested parties during the examination process, as most European patent offices do ?
I was trying to be provocative. Looks like it worked, huh ? Doing maths in Java is pretty unpleasant. For such things I tend to prefer functional languages, but I can see C++ might have its place if you're doing something performance critical. Java's mathematical deficiencies bother me rarely, as I spend my time writing application and server code in domains where "int" is all you really need.
Class library wise, I'm don't like the STL much - or the new Java collections, for that matter, but Java's class libraries go way beyond what the STL provides. Whether that matters to you or not, depends on what you're doing. Lack of portable threading primitives in C++ gets to me for what I prety badly, though, especially since this is the main reason the STL is either thread safe or correct, never both, in all the implementations I've tried.
I don't believe the core problem with C++ is its power. The core problem is the incoherent language design - its been accreted over a great length of time now, with each new level of features trying to make up for the known deficiencies in the last one, and with no sign of this stopping. The different features work together pretty poorly in general: try using old-style pointers with exceptions, for instance.
I frequently find myself being forced to make a decision in writing something that I know will affect any code that tries to use mine, but which I have no good basis for making. For instance: do I return a pointer, a reference or an auto-pointer ?
Expert C++ users (amongst whom I reluctantly count myself, as many others who claim that status know the language less well than I do) generally have a subset of the language, the STL and the compilers they know how to use safely, and some idea of what likely to go wrong whent they step outside that box. Many, however, don't realise that this is what they are doing. Raising these double-bind questions about the usage of C++ tends to get blank looks from may so-called experts on the language, simply because they've never used 1/3 of the language features. Thats not a sign of a language thats too powerful - its a sign of an ADA or PL/1 like agglomeration of unrelated features, in contravention of the C spirit of "worse is better".
I was trying to describe what you should do, not what gets done. My personal strategy in the kind of messed up closed source product shop situation you're describing is too make sure the first implementation is minimal function and absolutely rock solid. Then you can spend those later passes adding features to a stable base, rather than pushing the much around the bag.
Since you're the most highly rated person to pull me up for having used the word "performance" ambiguously, I'll correct myself here and ignore the others. It really should have been clear from the fact that I disagreed only with about half of what Mr Walker said, but I guess the deliberately provocative title caused a lot of people to read what they wanted to read, rather than what I meant to write. My primary purpose was not to give a lesson in project architecture but to indicate that just the same arguments can be made against C++.
There are two quite distinct kinds of activity often represented as performance enhancing. The first is simply correct architecture. When defining the computational and engineering aspects of the architecture you should, of course, make sure that you've chosen technologies, algorithms and data structures that can provide adequate performance and scalability. They're different things, incidentally, as someone else pointed out.
The second is actual optimisation. Even the best architected project will usually develop bottlenecks and other performance problems between being function complete and actually being deemed acceptable for use. This is the time when you should get the profiler out, diagnose the problem and repair it. Its with things likely to lead to these kinds of issues (for instance, the choice of a linear rather than binary search for some collection), but with no effect on interfaces between components, where I would usually advocate doing the simplest possible thing for the initial implementation and fixing it later if it proves to be a problem.
Its not "lazy" to worry about performance later. Its best practice. The first pass at a product should get the functionality right, get the bugs out, and keep the design clean. The second pass - possibly before the initial release - should involve testing for the acceptability of performance, profiling to locate the problems, and changes to the design to remove these.
In my experience Java tends to attract people who want their code to reflect what its intended to do, whereas C++ tends to attract people who like to write obscure code for "performance reasons" because it makes them feel clever. The pleasant thing about Java is that issues with the language do not get in the way of expressing what you want to do.
Now, as to your example, doing a linear search or a bubble sort is sometimes acceptable. Linear search, for instance, is OK if the data set is small, the operation rare and the data structure you've got does not support something better because it was optimised for some other case. Similarly, recalling "get" methods rather than caching the result is not usually a real performance problem: the VM optimises the fetch away to nothing.
However, what you report is disturbing. I don't consider overuse of linear search, refetching of already fetched data, and so on, to be performance problems. They're sheer carelessness, and they mess up the structure and expressiveness of the code. They reflect a lack of thought in the design. As you say, laziness. I must say, however, that I've seen this kind of thing in C and C++ much more than in Java, as their lack of class libraries tempts people to do things in stupid ways to avoid having to write the code needed to do it properly.
There's a Cambridge (UK) company called bango.net thats been doing this for a while. Its the silliest idea I've ever heard of. Have a system for translating names to numbers and then layer a system for translating numbers to names on top of it. Doh.
While its certainly a fallacy to suppose that something came from space just because it was found in space, thats not induction its a reverse implication (assuming a=>b is the same as b=>a).
For something to come from space its necessary for it to be able to survive in space, but not sufficient.
Incidentally, induction may be logically unsound, but it or something equally unsound is the only way to make correspondences between the world as we percieve it and our formal models of it.
In spite of its pretentious rehetoric and naive recommendations, this book does contain the a kernel of truth. We do have a lot of freedom about where to live, and most firms move to the kinds of places their employees like. Hence US software firms converged on the bay area, and similar things in other countries. They may have wrecked the local environment, but they and their employees stay, mostly just in order to stay near each other.
For industries that actually make stuff it was once important, and still is to some extent to be near the sources of raw materials.
Of course, the phemonon being described is neither unique to our era nor universal within it, but it is still a valid observation.
XML has some potentially useful 'meta' features. There are a lot of parsers, and other tools and standards that interoperate with it well.
From the perspective of config files, one of the more important things is that you could define a set of schema fragments that can be used to assemble the config files for particular systems.
Its certainly not necessary to use XML, but it does seem to add something over the creation of an entirely new and different syntax,
Thats deeply unrealistic. W3C recent standards are extremely complex (I have the schema spec before me: its in three parts, over around 300 pages, is very informally written, and deeply confusing). 100% compliance is very hard, when you also want 100% compatibility with buggy web pages.
Put bluntly: if you think 100% compliance and compatibility are possible, go do it. The world will beat a path to your door.
Its true that Mozilla contains dubious features for ticklist-compatibility with NS4.0 and IE4+. Thats inevitable. Anyone who's ever worked for a software product company will know that releasing version n+1 with less features than version n is suicide, even if the features were stupid in the first place. Perhaps more effort should have gone into differentating Mozilla from Communicator, so that Mozilla could just have been a good browser, but that might have raised other problems.
Skinning however is not such a dubious feature. Its a side effect of a brave attempt to do something which is very hard: create a cross-platform framework for user interfaces. The old Netscape front ends (one per platform) were an immense obstacle to development. They proliferated, and they tended to accumulate features that should have been elsewhere (URL completion in the Windows version). Netscape felt they were committed to supporting the Unix versions as a differentiator fromm MS, and yet could not justify the development resources to support something used by only a few percent of their users. *Thats* why NS4.0 on X is so crash-prone and slow: the FE code is buggy as hell, and pulls a lot of nasty X tricks.
Mozilla does its UI using its rendering engine precisely in order to get away from this problem. A side effect is that you can write pretty chrome, but that was not the purpose. You may feel the decision was wrong, and Galeon et al would seem to show that you might be right, but its an utterly different issue from sidebars and "What's related?".
Standard libertarian property theory, derived from John Locke, primarily, though it has earlier antecedents. Its a deontological ethics in which humans have certain "natural" rights (its not clear where these come from: God and human nature are the usual options), one of which is the right not to be coerced into labouring for others benefit. From this you can derive in order that:
1. Taking something after someone has laboured to create it is essentially the same as coercing them into labouring for you.
2. Persons can claim unclaimed natural resources by "mixing their labour with them", providing "as much and as good" is left for others.
3. People are entitled to give away or exchange their property.
4. Taking another person's property, thus defined, is never justified.
Of course, there are other ways to justify property, and this one has 2 major flaws. Firstly, its based on a unscientific claim about human nature, and subtly reasons from is to ought in the process. Secondly, no existing property was acquired purely by the mixing of labour and fair exchange. It was all stolen from someone at some point in time.
How very British. Government controlled by the acceptabilitly of its actions, rather than by the written constitution. Actually, I suppose the same thing really happens everywhere. Its just the UK government makes no pretence.
As a matter of interest, can the queen overule the Canadian governor general, or not ?
Slashdot Mirror
Lots of /. articles contain links to further information. Those that claim to show something of interest sometimes even contain calculations. Or at least, they used to in the old days. Possibly the standard here has sunk even further.
Regardless of that, I'm not going to believe him until he tells me what his results are based on.
Some questions:
1. How did you assign costs to CPU time ?
2. Where did you get your figures on the computational cost of a search, and its relation to the size of the web ?
3. Where did you get your figures on the size of the web ?
4. Can you supply a reference for the figure of 1 cent per impression for banner ads ? and do you know how much google gets for its text ads ?
Only temporary variable names are lost, unless the author used a scrambling program. Class, field and method names are preserved, because the link process for Java occurs at load time, and they are needed for this.
One of the characteristics of resistance groups - and make no mistake that is how these people see themselves - is that if you start out with 100 revolutionaries and execute 90, when you're finished you don't have 10, or even 100, but 10,000, or even more.
1. Tell them if they're too busy to use the system, you're too busy to write it for them. Seriously, you need to reach an agreement with clients when the project starts on how much work you expect them to do.
2. You need to do a proper analysis - interview the customer's staff, write use cases. This is where XP falls down - how are you supposed to know the problem domain ?
is that all average programmers, without exception, believe they belong to the tiny minority of the truly talented.
In some respects exceptional confidence in your own abilities is a more reliable indicator of inability, because you lack the ability needed to judge your own work critically. Noone truly brilliant I have met or heard of in any field believed themselves to be any better than average.
Thats what apt-get does to. If you haven't pointed it to the right place, it can't find the packages.
There are also downsides to having the state supply food and housing, which vary depending on how you administer such a scheme, and generally reflect the fact that it would have to be funded from tax receipts.
If the state owns and administers housing stock on behalf of those who would otherwise be homeless, there would be a lack of ownership and care over the houses. The government would not want to pay for nice houses or maintanenace, and the tenants, not owning the capital, would not want to either. This is, in fact, what happens in Britain. Similarly for food. If people go and collect their rations every day with no control over the contents, you are not going to get haute cuisine.
On the other hand, if the state merely agrees to fund individual choices as to where they want to live or what they want to eat, a price limit would have to be agreed. This is essentially like the voucher system used in some places for paying for education. It would work OK, except that the "best" providers of food and housing would price themselves out of the scheme, by refusing to accept voucers at all (stores in Britain do this to assylum seekers who are given tokens in place of cash, and do stores in the US with food stamps), or by raising their prices above what can be payed for with them. The value in many goods comes only from their exclusiveness, and anyway in many cases supply is limited and prices must be raises if demand rises.
So, do away with vouchers and just give people the cash. This is actually pretty close to being a basic income gaurantee - something I think is quite a good idea. The issues then are twofold. Firstly the redistributive nature of the scheme is now clear. Food and housing don't appear from nowhere after all - you're taking resources from one person and giving them to another. Secondly, do you mind that people may not spend their money on what you wanted to give them ? are you going to stop them from investing it in the stock market, losing it, and then having to beg ?
The book being reviewed is about capital. Your comment is about money. Capital is productive property, and as such can consist of ideas, land, machinery and even personal skills in some views.
Money is a reliable store of value that is accepted in exchange for other good.
Both capital and money are social realities: there's nothing "imaginery" about it. If I try to pay for dinner using a turnip, patent water or enclose the center of London to graze my cows I'll soon discover that imagination is not enough.
If there was ever a reason to abolish the first amendment, it's the kiddie porn issue.
Its unclear why. Its established precedent everywhere that I know of that freedom of speach protection does not apply in cases of obscenity. To an even greater extent, you have no freedom of speech where you're revealing that a criminal act took place (in this case statutory or actual rape of sexual assault). Thus kiddie porn (viewing, knowingly distributing, or making) is already illegal.
On the ethical, rather than the legal, level, there are two arguments for making child pornography illegal. The first, and best, is that making (some of) it involves rape (statutory and actual), and sexual assault, and that these are crimes in themselves. The second, and weakest, is that being expose to such images is somehow a horrible trauma that adults and especially children who come across it accidentally will suffer harm from.
I was hoping to find a finely considered argument for regulation here, of the kind Lawrence Lessig generally produces. I did not. In fact, I found a short essay practicing what it claims to condemn. It pushes a couple of predictable buttons for the general public - paedophiles and large corporations - in order to scare people into buying the "dangerous information" argument in general, and they goes on to argue that the government should regulate the internet for accuracy because it is different (no attempt to made to explain how) to media for which that would be considered a gross violation of human rights. No attempt is made to explain how to avoid the inevitable corruption and contention involved in such regulation.
I kind of agree, actually, that the public is in danger of poisoning its conciousness with inaccurate and oversimplified information, but thats been going on for years and the mass media have a lot more to answer for than the internet here. I believe our attitude to information is wrong: that we are too happy to sit back and absorb what is fed to us by people we never see and who'd motives we rarely know ? Is it surprising that some of what is fed to us is poison ? That attitude has been encourage by mass media, and in the age of the internet, we need to get over it. Someone suggested a kind of global moderation system: that seems like a good start. Ultimately, people need to accept that it is their responsibility to verify what they are told, and, if certain images disturb them, to control what they see. In my opinion, that means a change in our whole cultural attitude, starting with hwo we are taught in school.
In the end, this is just another attempt to fudge an issue of private responsibility by declaring that the government must take care of us all.
What you describe is certainly what the American patent system has come to, but it was not the intention. The patent office, until recently, researched the claims in a patent with considerable care and accuracy. In principle they still have duty to do so, and frequently reject the initial patent application with examples of possible prior art, usually from other patents. Whats changed is that patent exmaminers are now encouraged, by the way they are payed to pass everything, and the patent office has poor records of the existing prior art for many fields in which they are issuing patents, especially softwate and business models.
I'd tend to agree that civil servants are not the appropriate people to determine obiousness or novelty, but, due to the great expense of challenging a patent, nor are the courts. The existing system allows patent holders to exploit their competitors for licensing fees as long as these remain below the expenditure needed to challenge the patent.
If you're going to have a patent system, and especially if you're going to allow patents in so many fields, wouldn't the appropriate thing to do be to allow challenges by interested parties during the examination process, as most European patent offices do ?
I was trying to be provocative. Looks like it worked, huh ? Doing maths in Java is pretty unpleasant. For such things I tend to prefer functional languages, but I can see C++ might have its place if you're doing something performance critical. Java's mathematical deficiencies bother me rarely, as I spend my time writing application and server code in domains where "int" is all you really need.
Class library wise, I'm don't like the STL much - or the new Java collections, for that matter, but Java's class libraries go way beyond what the STL provides. Whether that matters to you or not, depends on what you're doing. Lack of portable threading primitives in C++ gets to me for what I prety badly, though, especially since this is the main reason the STL is either thread safe or correct, never both, in all the implementations I've tried.
I don't believe the core problem with C++ is its power. The core problem is the incoherent language design - its been accreted over a great length of time now, with each new level of features trying to make up for the known deficiencies in the last one, and with no sign of this stopping. The different features work together pretty poorly in general: try using old-style pointers with exceptions, for instance.
I frequently find myself being forced to make a decision in writing something that I know will affect any code that tries to use mine, but which I have no good basis for making. For instance: do I return a pointer, a reference or an auto-pointer ?
Expert C++ users (amongst whom I reluctantly count myself, as many others who claim that status know the language less well than I do) generally have a subset of the language, the STL and the compilers they know how to use safely, and some idea of what likely to go wrong whent they step outside that box. Many, however, don't realise that this is what they are doing. Raising these double-bind questions about the usage of C++ tends to get blank looks from may so-called experts on the language, simply because they've never used 1/3 of the language features. Thats not a sign of a language thats too powerful - its a sign of an ADA or PL/1 like agglomeration of unrelated features, in contravention of the C spirit of "worse is better".
I was trying to describe what you should do, not what gets done. My personal strategy in the kind of messed up closed source product shop situation you're describing is too make sure the first implementation is minimal function and absolutely rock solid. Then you can spend those later passes adding features to a stable base, rather than pushing the much around the bag.
Since you're the most highly rated person to pull me up for having used the word "performance" ambiguously, I'll correct myself here and ignore the others. It really should have been clear from the fact that I disagreed only with about half of what Mr Walker said, but I guess the deliberately provocative title caused a lot of people to read what they wanted to read, rather than what I meant to write. My primary purpose was not to give a lesson in project architecture but to indicate that just the same arguments can be made against C++.
There are two quite distinct kinds of activity often represented as performance enhancing. The first is simply correct architecture. When defining the computational and engineering aspects of the architecture you should, of course, make sure that you've chosen technologies, algorithms and data structures that can provide adequate performance and scalability. They're different things, incidentally, as someone else pointed out.
The second is actual optimisation. Even the best architected project will usually develop bottlenecks and other performance problems between being function complete and actually being deemed acceptable for use. This is the time when you should get the profiler out, diagnose the problem and repair it. Its with things likely to lead to these kinds of issues (for instance, the choice of a linear rather than binary search for some collection), but with no effect on interfaces between components, where I would usually advocate doing the simplest possible thing for the initial implementation and fixing it later if it proves to be a problem.
Its not "lazy" to worry about performance later. Its best practice. The first pass at a product should get the functionality right, get the bugs out, and keep the design clean. The second pass - possibly before the initial release - should involve testing for the acceptability of performance, profiling to locate the problems, and changes to the design to remove these.
In my experience Java tends to attract people who want their code to reflect what its intended to do, whereas C++ tends to attract people who like to write obscure code for "performance reasons" because it makes them feel clever. The pleasant thing about Java is that issues with the language do not get in the way of expressing what you want to do.
Now, as to your example, doing a linear search or a bubble sort is sometimes acceptable. Linear search, for instance, is OK if the data set is small, the operation rare and the data structure you've got does not support something better because it was optimised for some other case. Similarly, recalling "get" methods rather than caching the result is not usually a real performance problem: the VM optimises the fetch away to nothing.
However, what you report is disturbing. I don't consider overuse of linear search, refetching of already fetched data, and so on, to be performance problems. They're sheer carelessness, and they mess up the structure and expressiveness of the code. They reflect a lack of thought in the design. As you say, laziness. I must say, however, that I've seen this kind of thing in C and C++ much more than in Java, as their lack of class libraries tempts people to do things in stupid ways to avoid having to write the code needed to do it properly.
There's a Cambridge (UK) company called bango.net thats been doing this for a while. Its the silliest idea I've ever heard of. Have a system for translating names to numbers and then layer a system for translating numbers to names on top of it. Doh.
While its certainly a fallacy to suppose that something came from space just because it was found in space, thats not induction its a reverse implication (assuming a=>b is the same as b=>a).
For something to come from space its necessary for it to be able to survive in space, but not sufficient.
Incidentally, induction may be logically unsound, but it or something equally unsound is the only way to make correspondences between the world as we percieve it and our formal models of it.
In spite of its pretentious rehetoric and naive recommendations, this book does contain the a kernel of truth. We do have a lot of freedom about where to live, and most firms move to the kinds of places their employees like. Hence US software firms converged on the bay area, and similar things in other countries. They may have wrecked the local environment, but they and their employees stay, mostly just in order to stay near each other.
For industries that actually make stuff it was once important, and still is to some extent to be near the sources of raw materials.
Of course, the phemonon being described is neither unique to our era nor universal within it, but it is still a valid observation.
XML has some potentially useful 'meta' features. There are a lot of parsers, and other tools and standards that interoperate with it well.
From the perspective of config files, one of the more important things is that you could define a set of schema fragments that can be used to assemble the config files for particular systems.
Its certainly not necessary to use XML, but it does seem to add something over the creation of an entirely new and different syntax,
Thats deeply unrealistic. W3C recent standards are extremely complex (I have the schema spec before me: its in three parts, over around 300 pages, is very informally written, and deeply confusing). 100% compliance is very hard, when you also want 100% compatibility with buggy web pages.
Put bluntly: if you think 100% compliance and compatibility are possible, go do it. The world will beat a path to your door.
Its hard to delete the IE icon, among others. If you do it insistently enough, eventually it cooperates.
Its true that Mozilla contains dubious features for ticklist-compatibility with NS4.0 and IE4+. Thats inevitable. Anyone who's ever worked for a software product company will know that releasing version n+1 with less features than version n is suicide, even if the features were stupid in the first place. Perhaps more effort should have gone into differentating Mozilla from Communicator, so that Mozilla could just have been a good browser, but that might have raised other problems.
Skinning however is not such a dubious feature. Its a side effect of a brave attempt to do something which is very hard: create a cross-platform framework for user interfaces. The old Netscape front ends (one per platform) were an immense obstacle to development. They proliferated, and they tended to accumulate features that should have been elsewhere (URL completion in the Windows version). Netscape felt they were committed to supporting the Unix versions as a differentiator fromm MS, and yet could not justify the development resources to support something used by only a few percent of their users. *Thats* why NS4.0 on X is so crash-prone and slow: the FE code is buggy as hell, and pulls a lot of nasty X tricks.
Mozilla does its UI using its rendering engine precisely in order to get away from this problem. A side effect is that you can write pretty chrome, but that was not the purpose. You may feel the decision was wrong, and Galeon et al would seem to show that you might be right, but its an utterly different issue from sidebars and "What's related?".
Standard libertarian property theory, derived from John Locke, primarily, though it has earlier antecedents. Its a deontological ethics in which humans have certain "natural" rights (its not clear where these come from: God and human nature are the usual options), one of which is the right not to be coerced into labouring for others benefit. From this you can derive in order that:
1. Taking something after someone has laboured to create it is essentially the same as coercing them into labouring for you.
2. Persons can claim unclaimed natural resources by "mixing their labour with them", providing "as much and as good" is left for others.
3. People are entitled to give away or exchange their property.
4. Taking another person's property, thus defined, is never justified.
Of course, there are other ways to justify property, and this one has 2 major flaws. Firstly, its based on a unscientific claim about human nature, and subtly reasons from is to ought in the process. Secondly, no existing property was acquired purely by the mixing of labour and fair exchange. It was all stolen from someone at some point in time.
How very British. Government controlled by the acceptabilitly of its actions, rather than by the written constitution. Actually, I suppose the same thing really happens everywhere. Its just the UK government makes no pretence.
As a matter of interest, can the queen overule the Canadian governor general, or not ?