I was disappointed that the article only explained why OS X was more productive than Windows. Hell, I knew that. Watching over the shoulder of expert Windows developers is painfully slow. Yeah, you can do everything that you need to do, but it takes so many keystrokes and/or mouse events.
These are not really experts then. I always thought that this was the case as well, and on my linux box, I'm usually much faster than most developers on a window box. However, there came the expert windows user. Using the IDE, using the file explorer, that guy was extremely fast. He never touched the mouse for anything, just opened everything through the keyboard, navigated the file explorer with the keyboard, opening/creating directories, firing up an application, using the thing. Never seen anything like that. He was probably twice as fast doing the stuff on the windows box as I was on my linux box. Humbling.
It might be that the mouse handling in X can be a better experience than that of Windows, but true speed is reached on the keyboard. If you know what you're doing, you can just fire up a sequence of keystrokes that will do what needs to be done, while with the mouse you always need visual feedback to what is happening. Very tiresome and slow.
It's my experience that the keyboard shortcuts on X applications are at the very least inconsistent and too often not complete. A good GUI should be 100% usable without a mouse to help in being productive if you want to (and remove the risk of mouse related injuries).
Both Apple and Microsoft got this (mostly?) right. Maybe at one point in the development of both Gnome and KDE, the developers should just get rid of their mouse for a couple of weeks, and see how to make the thing usable without one.
Oh definitely the voters will protest. If proper care is taken to make understandable ballots, the protests will stay, but you could safely ignore them. I never understood the issue with the butterfly ballots as well. Just read the thing!
The paper receipt is however not about making things understandable, it's about making things verifiable. In a hostile environment that is a United States (presidential) election, where apparently partisan loyalty is more important than the democratic process itself, a verifiable election is a necessity, simply because the risk of tampering is too large. Voting mechanisms with independent sources of error (read: tampering) and a thorough auditing track can make the election results believable, also for the supporters of the candidate that lost.
As it is now, the outcome of a US election is as believable as that of a random third world country. Gerrymandering also doesn't help. This is not about creating a feel-good democracy,
this is about creating a solid process, backed up by fact and investigation if need be. Holding a fair and verifiable election is not rocket-science, and it's unbelievable that the US is seemingly incapable of holding one.
"The machine printed out bad receipts" (faded ink / or claims of tampering, remember we can't even get people to punch a hole properly, let alone check a receipt)
Uhm, if told at the spot the vote is still invalid until they fix the thing. If not told at the spot, the complaint doesn't count. If they can't/won't fix the thing, the precint needs to revote. If the machine prints out bad receipts regularly, one would expect at least one voter to complain.
"The receipts were confusing, people didn't understand what they were saying"
If the receipts and electronic votes add up and the name of the candidate they voted for is on the receipt and the name of the candidate they didn't vote for is not on the receipt (and it didn't say "You didn't vote for X"), you can ignore this as the vote was correct. Nothing is idiot-proof. People will complain, but that doesn't automatically make the complaint fair.
then of course there's the missing ballot box problems, or the gjust human error. What happens if you have more receipts than electronic votes? What if you have less? What happens when kids with laser printers start forging receipts?
If the stuff doesn't add up, you do an inquiry which will also check the validity of the receipts. If you're serious about this, you can do forensics on the receipts to figure out if they were printed by the machines. Remember, as you now have two fairly independent methods of tallying the vote, with wildly differing ways to tamper with, you have more opportunity to figure out what the hell went wrong, and which of the two trails is wrong. Remember, if there's a discrepency, this does not automatically make the machine wrong. You need to find out, but with such a system, you actually can.
It might add more points of failure, but these points are very different for both methods. Checks and balances. If there are discrepencies, you have a good chance of figuring out what went wrong, and which trail is tampered with. As it is currently, if it is tampered with, you're shit out of luck and will never find out. Tampering with both trails consistently to swing an election would be very difficult, and would need a conspiracy of such a high level that the country would be fucked regardless of it being found out or not.
That's the problem with lots of people in IT: their reading skills suck! I've found indeed that the quote is often used in this flawed way and even had it used against me early in a project where I suggested to use a reasonable algorithm instead of a naive one. Alas, it was ignored and of course, the application turned out to be so doggishly slow that it was unusable.
It's really simple. You vote on the machine, the machine prints a receipt. You verify the receipt and after verification put it in a ballot box. This completes your vote. Ideally, when there's no dispute about the election results, 5% of the ballot boxes around the country are selected at random, are recounted and compared against the machine results. If there are no discrepancies, no action is taken. If there are, a full scale inquiry is made.
The important bits are that both the machine vote and the receipt never leave the election office (to avoid vote buying), the machines and receipts are handled by different persons (to make concerted efforts of tampering more difficult), and that there is always a statistically significant sample of checks (to keep vigilant). Oh, and of course that you change the law so that the burden of proof for correctness lies with the election officials, not the citizens.
I wholeheartedly agree with this. I've worked in teams that took the "premature optimization" to such heart that they refused to consider efficiency *at all* in the program. The result is a slow mess that will need a complete rewrite to make it snappy.
Next to "premature optimization is the root of all evil", there should be a second mantra that should say something equivalent to "programmers that don't care about efficiency should be taken out and shot in the foot".
And the reason this is the case in FORTRAN is simply because this is the way it is done in mathematical notation. Whenever you have an enumeration \sum, the indices are taken from {i,j,k,l}. Upper bounds are {n,N,m,M}, etc. So if you're implementing math (and in a sense CS is discrete math), it's helpful if you write it out in its native language. Calling indices countI or whatever is hackish, as it obscures the mathematical relationship that you're implementing.
Spim is an ugly word. The word Spam is cute, has connotations with waste products and Monty Python, and generally rolls of the tongue easily. Spim is just a contraction, maybe a bit of a pun, but a lousy one at that. IM-spam would've done nicely.
Evolution of a language is not dictated by the people that use loose grammar and lack spelling ability. The written language evolves mainly through the written literature, not by the flaky email of a random CEO. As most writers are fond of their language, they usually take great care in their use of grammar (spelling is left for the editors). They will use modern grammar, but will make sure it makes sense and is readable. So, if I were you, I wouldn't lose much sleep on the flaky use of language by your fellow collegemen. Although they may be semi-illiterate, the evolution of the language doesn't depend on them.
True as that may be, but when the likes of Bill Gates are referring to OSS as "communist", they're not referrer to socialist paradise, but the 'evil empire' that once was the Soviet republic.
In that sense Microsoft is damn closer to totalitarian practice (like every company) than OSS.
If you want true civil disobedience, instead of "civil disobedience", you might want to notify the police yourself everytime you open up a debugger to do your job and insist they arrest either you for breaking the law, or your employer for making you break the law;)
The 'phages' story is even more cute. I once saw a documentary about a team in a Russian hospital that harvested those phages from the sewage system of the hospital. The reasoning behind this was that to find the bacterio-phages (sp) that are most effective, you will find them where there's plenty of prey, namely in the hospital itself. If you have an outbreak of some bacteria in a hospital, you will find its enemies near the exhaust pipe. I thought that this was extremely clever.
Oh definitely. I don't disagree with the cost, it's just that there also needs to be a human infrastructure in place that will get the warning out to the important places. That infrastructure will not be maintained for the necessary time in this case, as people need to be continuously educated on what needs to be done when a tsunami hits. As we're talking about a timespan of several generations here, the final maintainers of the system and the entire population will only have a faint memory of the tsunami that wiped out their great-grandparent's neighbours. Then suddenly the system starts beeping (after possibly a couple of false alarms in the distant past). What will people do? Go out to lunch.
As always, get the army on your side. If you look at revolutions over the world, it's always either the army actively doing it, or the army passively not taking out the revolt. For recent examples, look at Georgia and the Ukrain.
I'm talking about revolutions here, not about getting rid of foreign occupants; these have a whole different dynamics, and incidentally, there are always stringent arms laws in place with a foreign occupation. That never stopped the freedom fighters/terrorists though.
A comprehensive tsunami warning system needs to be maintained. Given that in that region a tsunami hits every century or so, try to calculate the odds that the thing will be operational around 2100 when the next one comes.
You've got an interesting perspective on your ministers as being political appointees, but I'm afraid you're confusing your own government with the European Commissioners, which are indeed a category of miscreants on their own. The UK in the council of ministers is represented by the UK ministers, with Bliar leading the pack.
Small correction. Brinkhorst is minister of economic affairs. He was the one trying to push the bill through (after being told by parliament not to do this), before they tried to sneak it through via the 'fisheries' route.
He's evidently a total scumbag and the small party he's from has in my view lost all credibility as a movement that values democracy.
Well, in Holland (and also Germany), we did. We lobbied our assess off and after a lot of struggle we got our parliaments to tell the government to stop supporting software patents.
Subsequently government ignored parliament and went right ahead.
It's absolutely not clear what to do in a situation like this. We do have the right democratic authority on our side, both locally and European, but on the grand scale of things, the software patent issue is not something to make parliament force government. It's only the entire European software industry at stake, not something important as a an extra tax-relief of three tenths of a eurocent. Who now to lobby with? Those goons in government? They've made up their mind, don't care, and have a couple of years left before the next election.
So there's no thing as an areligious position in anything? You seem to, as many religious people do, confuse "belief" with "Belief". The former is the rational weighting of possibilities and probabilities and thereby coming to a positive or negative judgement on the basis of evidence and study. Many atheist, though not all, disbelieve in any deity, tooth fairy or Santa Clause because of this process. The latter is the religious form, and is brought on by an irrational process that shuns evidence, but relies on extra-sensory experiences, revelations and good old word of mouth. It's unfortunate that a random process such as this one occupies the same linguistic space as the former one. In other words, belief is not by definition irrational, religious, or indefensible.
Maybe a simply measure could already be a bit effective. If every email receiving computer would as a default behaviour scan for urls in the email, and simply (w)get the url (making sure the portion that would be rendered is obtained), a normal spam campaign would backfire on the less bandwith endowed companies. Note that this should work for existing and non-existing usernames alike. An innocent message with a url will also get hit, but unless it's send to 10^7 users, this is not going to bring the site down. See it as an inverse email-tax, or better, a url-tax on sending urls through email: every email will result in at least a hit. So if someone wants to send a message containing an url to 10^8 people of which 9^8 recipients don't exist, it better be capable of handling 10^8 times #relays hits in a fairly short time.
Of course, vigilante, microsoft won't put up with it, etc., but the idea might have merit considering that it targets the clients of the spammers, the sellers, and makes it cost money to operate a site based on bulk advertisement.
I doubt that that's the issue here. I'm running konqueror and firefox under linux here, and for neither of these the Dutch locales are installed. I usually get very confused when I read Dutch on a computer (though it's my native tongue), so I don't ever install anything Dutch specific on any computer, so I guess it's really the address that's the issue. (as far as I know, the only regionalized place on this computer is the time zone). Also 'links' brings me straight to the Dutch site.
In any case, it's not really the point how it figures out I'm in Holland. I later noticed that for many searches, the top results returned are from.nl sites. It's an interesting strategy to be very region specific, but I really don't like that it pollutes my search as well. When I for instance type "Genetic Programming" (a subject I'm interested in), the first result I get is genetic-programming.org, the main page of John Koza from Stanford. The subsequent three hits are (English) pages on.nl domains, then genetic-programming.com, and again a lot of hits within the.nl domain (all in English as this is an international research area). This is totally unacceptable, and makes the thing completely unusable. I don't think pure regional search engines are the way to go, and the fact that I can't change it really infuriates me. I've tried a couple of research subjects more and they all bring me to.nl sites, not to authorative international sites. This is really bad.
Welcome to Microsoft's regionet: where do you want to stay put today?
My experiences differ. I'm located in the Netherlands, so I only get Dutch (and Belgian) news hits. I haven't figured out yet how to circumvent this as I'm not primarily interested in that news. Even selecting 'only return results in English' (for search) will give me the Dutch news hits (but English search hits). Going to the news search window and demanding only to get results in English will still get me results only in Dutch. So it seems that my IP-address precludes me from looking at global news through MSN. Very annoying, as I hardly ever am interested in Dutch news or Dutch rehashes of international news.
Furthermore, the first news hit I get for 'Linux' is an article in Computable, "Microsoft: veiligheid van Linux is een mythe". Translated, "Microsoft: safety of Linux is a myth". Second and third news item are ok (skype and cheap linux laptops). I do sense a bit of bias here, but it might be accidental. All in all, a less than happy user has left the MSN site, probably to never return.
Slashdot Mirror
These are not really experts then. I always thought that this was the case as well, and on my linux box, I'm usually much faster than most developers on a window box. However, there came the expert windows user. Using the IDE, using the file explorer, that guy was extremely fast. He never touched the mouse for anything, just opened everything through the keyboard, navigated the file explorer with the keyboard, opening/creating directories, firing up an application, using the thing. Never seen anything like that. He was probably twice as fast doing the stuff on the windows box as I was on my linux box. Humbling.
It might be that the mouse handling in X can be a better experience than that of Windows, but true speed is reached on the keyboard. If you know what you're doing, you can just fire up a sequence of keystrokes that will do what needs to be done, while with the mouse you always need visual feedback to what is happening. Very tiresome and slow.
It's my experience that the keyboard shortcuts on X applications are at the very least inconsistent and too often not complete. A good GUI should be 100% usable without a mouse to help in being productive if you want to (and remove the risk of mouse related injuries). Both Apple and Microsoft got this (mostly?) right. Maybe at one point in the development of both Gnome and KDE, the developers should just get rid of their mouse for a couple of weeks, and see how to make the thing usable without one.
The paper receipt is however not about making things understandable, it's about making things verifiable. In a hostile environment that is a United States (presidential) election, where apparently partisan loyalty is more important than the democratic process itself, a verifiable election is a necessity, simply because the risk of tampering is too large. Voting mechanisms with independent sources of error (read: tampering) and a thorough auditing track can make the election results believable, also for the supporters of the candidate that lost.
As it is now, the outcome of a US election is as believable as that of a random third world country. Gerrymandering also doesn't help. This is not about creating a feel-good democracy, this is about creating a solid process, backed up by fact and investigation if need be. Holding a fair and verifiable election is not rocket-science, and it's unbelievable that the US is seemingly incapable of holding one.
Uhm, if told at the spot the vote is still invalid until they fix the thing. If not told at the spot, the complaint doesn't count. If they can't/won't fix the thing, the precint needs to revote. If the machine prints out bad receipts regularly, one would expect at least one voter to complain.
"The receipts were confusing, people didn't understand what they were saying"
If the receipts and electronic votes add up and the name of the candidate they voted for is on the receipt and the name of the candidate they didn't vote for is not on the receipt (and it didn't say "You didn't vote for X"), you can ignore this as the vote was correct. Nothing is idiot-proof. People will complain, but that doesn't automatically make the complaint fair.
then of course there's the missing ballot box problems, or the gjust human error. What happens if you have more receipts than electronic votes? What if you have less? What happens when kids with laser printers start forging receipts?
If the stuff doesn't add up, you do an inquiry which will also check the validity of the receipts. If you're serious about this, you can do forensics on the receipts to figure out if they were printed by the machines. Remember, as you now have two fairly independent methods of tallying the vote, with wildly differing ways to tamper with, you have more opportunity to figure out what the hell went wrong, and which of the two trails is wrong. Remember, if there's a discrepency, this does not automatically make the machine wrong. You need to find out, but with such a system, you actually can.
It might add more points of failure, but these points are very different for both methods. Checks and balances. If there are discrepencies, you have a good chance of figuring out what went wrong, and which trail is tampered with. As it is currently, if it is tampered with, you're shit out of luck and will never find out. Tampering with both trails consistently to swing an election would be very difficult, and would need a conspiracy of such a high level that the country would be fucked regardless of it being found out or not.
That's the problem with lots of people in IT: their reading skills suck! I've found indeed that the quote is often used in this flawed way and even had it used against me early in a project where I suggested to use a reasonable algorithm instead of a naive one. Alas, it was ignored and of course, the application turned out to be so doggishly slow that it was unusable.
The important bits are that both the machine vote and the receipt never leave the election office (to avoid vote buying), the machines and receipts are handled by different persons (to make concerted efforts of tampering more difficult), and that there is always a statistically significant sample of checks (to keep vigilant). Oh, and of course that you change the law so that the burden of proof for correctness lies with the election officials, not the citizens.
Maybe #define BLACK white and get run over at the next zebra crossing?
Next to "premature optimization is the root of all evil", there should be a second mantra that should say something equivalent to "programmers that don't care about efficiency should be taken out and shot in the foot".
And the reason this is the case in FORTRAN is simply because this is the way it is done in mathematical notation. Whenever you have an enumeration \sum, the indices are taken from {i,j,k,l}. Upper bounds are {n,N,m,M}, etc. So if you're implementing math (and in a sense CS is discrete math), it's helpful if you write it out in its native language. Calling indices countI or whatever is hackish, as it obscures the mathematical relationship that you're implementing.
Spim is an ugly word. The word Spam is cute, has connotations with waste products and Monty Python, and generally rolls of the tongue easily. Spim is just a contraction, maybe a bit of a pun, but a lousy one at that. IM-spam would've done nicely.
Evolution of a language is not dictated by the people that use loose grammar and lack spelling ability. The written language evolves mainly through the written literature, not by the flaky email of a random CEO. As most writers are fond of their language, they usually take great care in their use of grammar (spelling is left for the editors). They will use modern grammar, but will make sure it makes sense and is readable. So, if I were you, I wouldn't lose much sleep on the flaky use of language by your fellow collegemen. Although they may be semi-illiterate, the evolution of the language doesn't depend on them.
I'm sure the Chinese who put together your $50 DVD-player in sixteen hour shifts every day for a wage they can barely survive on beg to differ.
In that sense Microsoft is damn closer to totalitarian practice (like every company) than OSS.
My computer model predicts that it will rain heavily ten years from next Friday around 2 PM GMT. Unfortunately, it doesn't say where :(
If you want true civil disobedience, instead of "civil disobedience", you might want to notify the police yourself everytime you open up a debugger to do your job and insist they arrest either you for breaking the law, or your employer for making you break the law ;)
The 'phages' story is even more cute. I once saw a documentary about a team in a Russian hospital that harvested those phages from the sewage system of the hospital. The reasoning behind this was that to find the bacterio-phages (sp) that are most effective, you will find them where there's plenty of prey, namely in the hospital itself. If you have an outbreak of some bacteria in a hospital, you will find its enemies near the exhaust pipe. I thought that this was extremely clever.
Oh definitely. I don't disagree with the cost, it's just that there also needs to be a human infrastructure in place that will get the warning out to the important places. That infrastructure will not be maintained for the necessary time in this case, as people need to be continuously educated on what needs to be done when a tsunami hits. As we're talking about a timespan of several generations here, the final maintainers of the system and the entire population will only have a faint memory of the tsunami that wiped out their great-grandparent's neighbours. Then suddenly the system starts beeping (after possibly a couple of false alarms in the distant past). What will people do? Go out to lunch.
I'm talking about revolutions here, not about getting rid of foreign occupants; these have a whole different dynamics, and incidentally, there are always stringent arms laws in place with a foreign occupation. That never stopped the freedom fighters/terrorists though.
A comprehensive tsunami warning system needs to be maintained. Given that in that region a tsunami hits every century or so, try to calculate the odds that the thing will be operational around 2100 when the next one comes.
You've got an interesting perspective on your ministers as being political appointees, but I'm afraid you're confusing your own government with the European Commissioners, which are indeed a category of miscreants on their own. The UK in the council of ministers is represented by the UK ministers, with Bliar leading the pack.
He's evidently a total scumbag and the small party he's from has in my view lost all credibility as a movement that values democracy.
Subsequently government ignored parliament and went right ahead.
It's absolutely not clear what to do in a situation like this. We do have the right democratic authority on our side, both locally and European, but on the grand scale of things, the software patent issue is not something to make parliament force government. It's only the entire European software industry at stake, not something important as a an extra tax-relief of three tenths of a eurocent. Who now to lobby with? Those goons in government? They've made up their mind, don't care, and have a couple of years left before the next election.
Of course, vigilante, microsoft won't put up with it, etc., but the idea might have merit considering that it targets the clients of the spammers, the sellers, and makes it cost money to operate a site based on bulk advertisement.
In any case, it's not really the point how it figures out I'm in Holland. I later noticed that for many searches, the top results returned are from .nl sites. It's an interesting strategy to be very region specific, but I really don't like that it pollutes my search as well. When I for instance type "Genetic Programming" (a subject I'm interested in), the first result I get is genetic-programming.org, the main page of John Koza from Stanford. The subsequent three hits are (English) pages on .nl domains, then genetic-programming.com, and again a lot of hits within the .nl domain (all in English as this is an international research area). This is totally unacceptable, and makes the thing completely unusable. I don't think pure regional search engines are the way to go, and the fact that I can't change it really infuriates me. I've tried a couple of research subjects more and they all bring me to .nl sites, not to authorative international sites. This is really bad.
Welcome to Microsoft's regionet: where do you want to stay put today?
Furthermore, the first news hit I get for 'Linux' is an article in Computable, "Microsoft: veiligheid van Linux is een mythe". Translated, "Microsoft: safety of Linux is a myth". Second and third news item are ok (skype and cheap linux laptops). I do sense a bit of bias here, but it might be accidental. All in all, a less than happy user has left the MSN site, probably to never return.