*Most* people are terrible at critical reading. Just terrible.
Personally I don't find this too surprising in such a commercialised society, where people are trained to watch commercials, follow trends, and basically make less decisions for themselves.
Society in many places, as an entity, doesn't exactly want people to think critically, because then they might not respond as well to marketing, and they might spend less money.
However, it is interesting to note that it was running Linux about a month ago.
The Netcraft link you posted does claim that it was running Linux in September, but it also claims that the web server was still IIS. Perhaps this is me mis-interpreting the way that Netcraft presents its data, but how does one run IIS on a Linux platform? Unless it was something like Apache temporarily configured to report that it was Microsoft-IIS, but I can't see the point in doing that.
The IP address is also radically different, so my guess is that the domain was temporarily redirected to another hosting service in September that had a server configured quite differently, and possibly wasn't reporting its state as accurately.
The real people who advanced freedom and the cause of free software were the folks over at the GNU project and (at the time) the CSRG. Usually I agree with the people who roll their eyes when Stallman goes on about GNU/blah but this time I can see his point...
I'm pretty sure this is just Time Magazine, or at least my own impressions of it. It tells people what they want to hear and reflects what a majority of people of the time tend to be thinking, but it's not the greatest source of active journalistic investigation. Opinions may vary.
Election-system vendors and election officials generally focus on effective defense against outside attackers, usually characterized as hackers. Meanwhile, many public interest groups have focused on the possibility of election officials corrupting the results.
Personally I think the second part of this paragraph is the most important. One of the huge problems with any use of information technology as a fundamental part of an election process is trust. Above anything else, an election system should be trusted by as much of the population as is possible, and to be fully trustworthy, the election process has to be fully visible and understood by as many people as possible.
It's quite easy for most people to understand a manual election. It's as simple as voters making a mark on a paper ballot, putting it in a secure box, and then having the votes counted afterwards. Any concerned groups from nearly any cross-section of society can examine the process, provide observers, and make sure it's being done properly.
Wrapping up the selection, verification and counting process inside computers reduces the amount of people who can understand what's happening by orders of magnitude. It doesn't really matter if the voting system is open source, well designed and administered, or whatever. It's always going to exclude the majority of the population from being able to fully understand how it works, and to trust that it's working properly.
It's quite possible that IT systems can help with elections, and they already are in some places, but I don't personally think they should be used at the expense of a manual process, and I don't think they should be depended on for anything other than an early indication of the result. Voting machines, when used, should always provide voter-verified paper trails that are always deposited in a secure box in a voter-verified way using a fully visible and voter-controlled process. Manual recounts should be mandatory if there's any reasonable doubt of the outcome by anyone.
I'm not sure about the "never expected to fly this long" part - I can't find any references right now, but I'm pretty sure each orbiter was specified to last for at least 100 missions.
Well I can't vouch for its authenticity, but someone who edited Wikipedia claims that "each shuttle was designed for a projected lifespan of 100 launches or 10 years' operational life". I couldn't say whether this is because some components would deteriate too much after 10 years, or after 100 launches, but I guess part of the problem is that the shuttles haven't been launched anywhere near as frequently as the original plans projected, for a variety of reasons. (Budget cuts, accidents, etc). According to that projection, Columbia was more than twice the age of its orignally projected lifetime when it disintegrated on re-entry. (This isn't necessarily a definitive metric, though, given that there have only been two fatal accidents and Challenger was only a few years old when it exploded.)
Anyway, I can only imagine the stream of four letter words that Mullane would have yelled if the space walk, which he trained and dedicated years of his life, was canceled because a computer program couldn't handle January 1st...
It's a fair enough comment and I hope I didn't come across as condescending in any way, because I agree that it's a bit silly having a bug like this in a system being used today. I think more than anything, the problems are likely to be related to stretching the programme so far beyond what was intended. What remains is based on some very legacy technology, but probably very critical, given how much I understand the Shuttle's flight ability relies on its computer systems.
...which is more than many software development processes would reveal. Chances are that this known restriction is on a check-list which every shuttle mission has to be checked against, and the list would exist precisely because the software development and verification process is so solid and conservative.
As a professional software developer, I have heard on countless occasions about how the Space Shuttle software development process is so incredible, and how all other developers should try to live up their high standards.
Opinions vary, but I don't think I'd ever recommend working to the same standards, unless the customer actually had good reason to require it. (NASA does.) Even aside from your own code, doing it properly would require an extensive understanding of any and every third party library and system the code interacts with, which could add orders of magnitude to the development time and cost, even if it's open source and open hardware. I don't like hacks and yucky untested code any more than most people, but at some point it can just make sense to avoid extensive and pedantic formal development processes in favour of just getting it to work.
A lot of development processes (perhaps most) wouldn't have stopped the shuttle launching, even if this were reported as a bug. Chances are that it'd be forgotten about (if not fixed straight away), and someone would stumble on it again accidentally. Many bugs aren't even reported until someone's stumbled on them at least once. This is fine in most situations. Once it becomes a problem again, you can go and look it up, quickly find out everything that's known about it from before, apply any known workarounds, and spend time to fix it if necessary. The point, though, is that many systems wouldn't be sure to keep you informed about the restriction in a way that actively prevents someone stumbling on it later.
I still agree that it seems a little strange that this problem wasn't fixed ages ago. Realistically, though, the Shuttle was never expected to fly this long. It sounds a lot like a compromise that was made in the earlier days when computers were more limited, probably even moreso for the restricted range of systems that are certified to work under such conditions. Any update is likely to be very expensive and time consuming, simply because the software development and verification process is so solid and reliable.
I would say that requiring a reboot every year on December 31 is a pretty huge error. In this case, it is forcing NASA to launch earlier than they otherwise would wish.
From the article you quoted, it sounds more like they dropped a spacewalk (for Hubble maintenance, probably not safety-critical) so they could return sooner and avoid encountering the bug. To me it sounds like they did what they should have done, with safety as a priority.
Launching spacecraft is an industry in which the stakeholders usually prepare for possible or likely delays. NASA has to delay launches all the time for all sorts of reasons. I'm not sure why a possible software problem would be treated any differently. If the problem is with the managers dangerously forcing early launches, NASA should really be fixing their managers as a priority over fixing a known bug with a known workaround. Weighing it out, it's probably a lot cheaper, easier and safer to simply delay the occasional launch for a few more days, especially given that the Shuttle's remaining days are limited. Why risk the safety of future launches by making changes that will soon become obsolete?
Anyway, those are just my own thoughts. I don't work in software development where the process is quite so strict, but I'm sure they know what they're doing when they don't fix something like this.
If we could just wait a day or so and let paper ballots be counted we would not have these issues. Sure paper ballots could be miscounted but there are more eyeballs, and it would certainly be harder to pull off a massive fraud like what would be trivial with today's Diebold machines.
Definitely. I've just gone and watched the demo, and read a bit about it. Good on these people for coming up with a system where it's (apparently) impossible to prove to anyone else who you voted for, yet still allows for someone to be able to verify their vote to some extent. That said, I still think that trying to solve this is really trying to fix the wrong problem.
The only reason receipts are wanted right now is because some voters have a lot of doubt about whether their vote was counted correctly. The problem could be solved much better by fixing the cause rather than trying to treat the symptoms. Letting people have receipts won't actually improve the validity of the election, anyway, it'll only help people feel better about themselves. It certainly doesn't mean that a reliable recount can take place, because the vast majority of people will never check their vote, keep their receipt, or bother to return it on request. At best it'll indicate that there's been a problem with the election, but it's not as if we don't already know that, and it hasn't taken voter-only-verifiable receipts to figure it out.
Trying to do something this complex on the scale of a national election, or even small elections that involve a typical cross-section of the general public, is asking for trouble. An election is trusted because it's kept simple, and the people who vote can see and have a reasonable understanding of how it actually works. People can understand the concept of writing a vote on a piece of paper, dropping it in a secure box, having trusted people empty the box and count the votes, and allowing other trusted people to observe the process at all stages.
Tying the whole thing into computers, digitized logic hidden inside electronic machines, abstract metaphor (such as dragging and dropping virtual objects), and abstract automated counting methods, reduces the number of people who can understand the entire process, let alone any of the process, by orders of magnitude. It just opens up more possibilities for misunderstanding, confusion and concerned citizens who no longer trust the process.
Perhaps this system can be used for other things, but I really hope we don't resort to using it in large scale elections. There are so many other very basic things that need fixing first, and I'm skeptical whether anything like this will be beneficial once the root causes of the problems have been dealt with.
In times when it's difficult for either major side to gain a clear majority, there are plenty of reasons for both major sides prefer an ambiguous loss rather than risk letting the other side have a close but definitive victory. For example, an ambiguous outcome makes it easier to determine the actual result in the courts, and having the right lawyers, or simply throwing more money at the problem, could be preferable to relying on unpredictable voters.
In a quote that's frequently repeated on Slashdot, Robert J. Hanlon once stated: "Never attribute to malice that which is adequately explained by stupidity." Given some of the stories that have emerged about voting machines in use now and in the past, I personally have trouble believing that malice is not involved.
My main question is: Do you believe that the current situation with e-voting in the USA can be adequately explained by incompetence among politicians and partisan election officials, and corporations to understand problems with the technology? Furthermore, if corruption is involved, and if people in certain positions are influencing the decisions to use unreliable voting systems, how serious-an-effect do you think this interference will have on efforts to make voting systems more reliable?
The voter reviews it, makes sure it's correct, and then exits the booth and deposits the paper ballot in an old-fashioned ballot box. When the polls close, we have an instant count but if the result is challenged, we have the old-fashioned system to do a recount.
For this to work reliably, you'd really need a way to make sure that voters can't ever leave with their voting paper, either accidentally or deliberately, for two reasons:
It just provides a way for the two counts to become inconsistent, which opens up all sorts of ambiguity and possible legal challenges afterwards. I can quite easily see disappointed partisans claiming that it wasn't made clear enough that people weren't supposed to keep the paper slip, and so on.
At least as critically, it also removes the anonymity of the election. If it's even possible that the electronic count is the only one that will matter, someone could quite easily threaten to kill your children if you don't bring back the slip proving who you voted for electronically.
Rather than relying on voters to deposit their paper votes themselves, a better way would be for the machine to print the paper vote and display it behind a window. The voter verifies that it's correct, and then manipulates a visibly simple, mechanical switch that either accepts their vote (and visibly drops it into the box), or rejects the vote, and visibly destroys it, or drops it somewhere clearly marked as a rejected vote. If the voter rejects the paper vote, of course, the digital vote gets withdrawn, the process re-starts, and they can select again.
I don't understand why an open voting system wouldn't work. (And yes I know the major hurdle would be beating the peoples in power to transition to one)
An open source electronic voting system could certainly be beneficial, but I'm not convinced that any system will work unless the people who administer it are both capable and motivated to administer it properly.
In response to your other suggestion:
when you vote you're given a ticket with a number, anyone can go online and see how everyone voted but only you are able to tell which vote was yours by the corresponding ticket number. That'd allow for everyone to do their own count if they wanted.
This should never be allowed in an anonymous election, because it means that votes are only anonymous for as long as the voter is not threatened to hand over their receipt, or to prove their vote to someone else, which defeats the purpose of having an anonymous election.
Besides, consider what would happen if the elected representatives did something that was unpopular. Want to cast doubt on the authenticity of the election after the event? Just find enough people to claim that their vote was recorded incorrectly. It's their word against yours.
All the idea offers is a feeling of satisfaction for voters, but it opens the door for more controversy. If voters need to feel satisfied that their votes were recorded correctly, it should be accomplished by having a stable and reliable system that all voters (within reason) can understand the complete workings of. Perhaps this would require a mandatory manual recount of the voter verified paper trail after the election to confirm the result, at least in areas where the count was reasonably close, but I think it'd be absolutely worth it to have a fair election.
Instead of making it look "gorgeous", how about focusing on making Linux look "consistent"?
I don't entirely disagree with your overall point, but the problem I have with this reasoning is that I use Linux, "consistent" usually means far less options, and I'm concerned that the options you happen to think are appropriate for some kind of plan to "market Linux to the masses" are unlikely to be what's optimal for me.
I think one of the major problems with this is that so many people are thinking of it as "Linux", when Linux really has little or nothing to do with the desktop that everyone sees.
It'd make much more sense to make KDE look consistent, and encourage people to use KDE, or make Gnome look consistent, and market that at people. Or maybe the groups in charge of KDE and Gnome might decide to combine the interfaces a bit more, and that could be marketed to people. Or maybe someone just builds some kind of super-distribution and gets lots of attention, and that distribution can be marketed at people. But it shouldn't be marketed as "Linux", because Linux is only a single component of a much larger system with many more obvious components, and the Linux part of the system is replaceable.
In particular, people don't have to tie themselves to Linux for this. They can tie themselves to any of the BSD's or to Solaris, just for starters, and still get a similar experience if they're running the same desktop. But trying to argue that "Linux should look consistent" will never work, because Linux is a kernel and it doesn't look like anything.
Firefox's original claim to fame was "Not being IE". It didn't have ActiveX, supported tabs, and was super speedy.
I agree with most of what you've said, but not quite with this. The way I remember things, Firefox's original claim to fame, at least among geeks, was not being Mozilla, as in Mozilla Seamonkey. In other words, it was Mozilla with all the extra flab cut out. Just a browser. No mail client, no composer, no newsreader, no built in IRC chat client, etc etc. This made it much faster loading, and eventually more comparable with Internet Explorer as a drop-in replacement.
It's still just a browser, but since then it seems to have been becoming a somewhat more heavy duty browser trying to meet everyone's needs, without necessarily doing everything brilliantly. The irony for me is that Mozilla Seamonkey can still be installed in modules to some extent, so it's actually possible to install just the browser. (Well, the browser seems to come with the Composer in my debian package, I'm not sure if that's standard.) I haven't done any extensive benchmarks, but anecdotally it does often feel more cut down to me than Firefox does. To be honest, it feels more stable and I almost prefer it over Firefox.
The full Mozilla Seamonkey suite seems to have been forgotten in all the Firefox hype, but the only thing that's putting me off using it more often is that it doesn't seem to have a very reliable extension handling interface. That is, I can install an extension from an XPI file, but there's not actually an extension management interface in the front end UI that I could find. Perhaps this is a good thing, because it diversity of so many third party extensions of varying quality seems to be one of the big complications of Firefox stability.
The city's infrastructure had historically been a multi-platform environment that included about 100 Solaris servers used to run a large number of Oracle databases and applications. As these servers neared the end of their life cycles the City of Chicago began the migration to cost-effective Red Hat solutions.
Given that they're migrating their Solaris server boxes rather than Windows desktop workstations (as seems to be happening in Munich), I'm not sure if it's quite the same thing.
I think they know they Democrats will get in next time, and they're going to hand them a mountain of debt, which they'll try and sort out, causing a serious cooling of the economy. Then four years later, back will come the Republicans, saying "remember how good you had it under us?"
Don't get me wrong, I think PatchGuard probably has more holes than a slice of Swiss cheese... but the submitter's text needs redacting, and the original article could do with an -actual- statement from Microsoft.
Perhaps this link was added to the slashdot summary after you posted your comment for all I know, but the slashdot summary that I read had two links, and I found that statement quite clearly after following the first link. About the 13th paragraph down in that article states, complete with the additional link that I've included here:
Microsoft immediately responded with a angry attack stating that that the hack harmed windows users by reducing the security of Windows.
So no points for grammar in that sentance (which I copied verbatim), but it seems to explain quite clearly what the Microsoft criticism is. That second linked article begins with the paragraph:
Microsoft officials say they are unhappy that security software maker Authentium has decided to bypass the controversial PatchGuard kernel protection feature in its next-generation Vista operating system, and said that the tactic could lead to eventual problems for users of the company's software.
...and goes on for quite a while. Is this the statement you meant?
Professors are not the ones who will decide whether Wikipedia will make the grade or not. The populus will.
I think this really depends on the definition used for "making the grade". Just because something succeeds doesn't make it great, or even good, if you happen to measure greatness by a different metric. The populace determines a lot of things, including an adequate quality of television, and an adequate quality of elected government legislators. I don't agree that either of these meet my own expectations of adequacy, on most occasions, and I can fully appreciate that any individual should be able to decide that they also don't want to accept something simply because "the populace" does.
I use Wikipedia a lot, and I contribute to it a lot. I love it because it has such a wide range of information about all sorts of things, which is very difficult to find elsewhere. There aren't many other places where you could find such an in-depth description of something like the Slashdot subculture, for instance, and I'd consider directing people to that article if they wanted a summary of the site.
But I use Wikipedia cautiously, with an understanding of how it works, and that the information I get from it probably isn't authoritative and will need verifying if I'm using it for something important. If I relay the information to other people, I'll usually indicate where I got it from, and do my best to point out that it could potentially be suspect if I think it's appropriate to do so. I'm not sure that most of the populace has the same cautious approach. I'm beginning to see more and more people treat Wikipedia as if it's just automatically authoritative information, without any realistic understanding of what it actually is and where it comes from. I find that a bit disconcerting.
All that said, I guess with people's choices of TV and political representation, and many other things, there are quite a few things that I have trouble relating to in the "general populace".
I've been trying to change my default browser at work, where we tend to use a very Microsoft-centric desktop for most things. (Outlook, IE, Office, etc.) Switching the browser that runs underneath the Start Menu browser icon is quite easy, but I still haven't figured out how to make all applications use Firefox, or any other browser for that matter.
The most obvious example is Outlook, which seems to want to open links in MSIE no matter what I do. It' be nice to be able to tell Outlook to use another rendering engine entirely (such as Gecko), but I'd be keen to simply tell it to run Firefox instead of IE when it wants to open a completely new browser instance. Does anyone know how to change this? Is it even possible with Outlook?
It might be the greatest OS ever, but if you're trying to get ADSL working, and it doesn't work because it assumes it's pppoe when it's nothing of the sort, and it doesn't recognise your modem, it's as useful as DOS.
I guess my point is that as long as the enduring experience is good, it opens up other options. For one thing, it'd make it easier for OEM's or retailers to install it for you, which is effectively what they do with Windows at the moment. It also opens up other possibilities, such as having someone else install it. (My mum and dad wouldn't care about the installer if I did it all for them, but they'd care a lot about being able to keep things working after I'd gone home.)
When I was first getting started in Linux (ages ago now), the installation was relatively easy. It took a while, and I had to figure out things I hadn't done much of before, such as boot managers and repartitioning. What really caused me to go back to Windows for a few more years, though, was that the whole system slowly degraded over time in ways that I couldn't figure out how to fix. The reason for this was that although the installer set things up initially, it didn't help to keep them running. I was basically dumped all at once into a system with a heap of configuration files for programs that I knew nothing about, most of which had no obvious configuration interface. The only way I knew how to reliably change some of the settings was to reinstall from scratch, and obviously I didn't want to do that.
I think a neat thing about a lot of distro's and desktop environments lately is that they're providing a slightly more intuitive interface for helping people to fix things when they break. I'm still not sure if they're quite there yet, but it's a lot better than it's been in the past.
It's the same thing with attacking RMS. He is working for us, and you better damn appreciate it.
I'm undecided whether I agree with RMS or not on this, but I don't really agree with what you've said... in this sentance at least. To me this implies that we should be blindly accepting the personality simply because it gets attention for things that I sometimes agree with.
That said, I'm all for promoting constructive criticism, and looking at what people are actually saying rather than focusing on who the person is.
After the pain of the XP install was over it worked just fine. He loves it... it's faster than any PC he has ever owned.
This is one of the reasons that I get a little skeptical about how so many people like to judge a linux distribution by its installer. A lot of reviews I've read over time of linux distributions have focussed almost entirely on the installer. It's true that most people who use Linux (unlike Windows) will need to install it at some point, and installation is quite important because of that. But surely the entire enduring experience after the installation, including things like system maintenance to keep it running nicely, is at least as important.
That means that a significant amount of money that would have gone towards PCs this Christmas will probably go towards something else. OEMs are pissed, and rightly so. Missing the Christmas season is the unpardonable sin in the retail business.
In all honesty, though, what would retailers do about it? Ditch Microsoft and start shipping PC's with Linspire?
We do this in a lot of places too, and I think there are perfectly good reasons for it, including security. (eg. If my account ever gets hacked, someone probably still needs to know a much more secure password if they want to give the account more access.) Another is just plain robustness. It's harder for me to accidentally break things when I don't have access to them.
At least as importantly, though, I think it helps the users actually trust us more easily. Most of our users realise that we don't automatically have access to their documents, for instance. They also know from experience that we'll tend to ask them if we ever need to give ourselves access. Another example is with watching (and controlling) their desktops, usually for tech support. We could quite easily configure things to be able to connect at any time, but instead we make sure that whenever we do it, they first get a prompt which asks them if they'll give permission. Obviously it doesn't mean we can't do things without them knowing, but having policies about that sort of thing which people understand really makes it easier to work with the other people in the organisation.
For instance, "Disable the close button on background (unfocused) tabs" sounds pretty self-explanatory.
I can see this in about:config, but where is it in the options dialog? (Keeping in mind that the vast majority of people likely to get frustrated by accidental tab closures will be people who don't know about about:config.) I'm using Firefox 2.0 right now, and the four options I have in the Tabs section are to "Warn when closing multiple tabs", "Warn when opening multiple tabs that might slow down Firefox", "Always show the tab bar", or to "Switch to a new tab immediately when opening".
More importantly, the decision to go this way seems to be making it much easier to accidentally close a tab when simply trying to switch between them. Suddenly the tab-switching bar is interspersed with close buttons. I'll be adjusting it in about:config straight away, but I'm a little concerned that as well as making it a default behaviour, there's not even an option visible in the preferences dialog for people to change it.
In defence of the devs, I guess there was an entire open development process in which to raise this earlier.
Slashdot Mirror
Personally I don't find this too surprising in such a commercialised society, where people are trained to watch commercials, follow trends, and basically make less decisions for themselves.
Society in many places, as an entity, doesn't exactly want people to think critically, because then they might not respond as well to marketing, and they might spend less money.
The Netcraft link you posted does claim that it was running Linux in September, but it also claims that the web server was still IIS. Perhaps this is me mis-interpreting the way that Netcraft presents its data, but how does one run IIS on a Linux platform? Unless it was something like Apache temporarily configured to report that it was Microsoft-IIS, but I can't see the point in doing that.
The IP address is also radically different, so my guess is that the domain was temporarily redirected to another hosting service in September that had a server configured quite differently, and possibly wasn't reporting its state as accurately.
I'm pretty sure this is just Time Magazine, or at least my own impressions of it. It tells people what they want to hear and reflects what a majority of people of the time tend to be thinking, but it's not the greatest source of active journalistic investigation. Opinions may vary.
Personally I think the second part of this paragraph is the most important. One of the huge problems with any use of information technology as a fundamental part of an election process is trust. Above anything else, an election system should be trusted by as much of the population as is possible, and to be fully trustworthy, the election process has to be fully visible and understood by as many people as possible.
It's quite easy for most people to understand a manual election. It's as simple as voters making a mark on a paper ballot, putting it in a secure box, and then having the votes counted afterwards. Any concerned groups from nearly any cross-section of society can examine the process, provide observers, and make sure it's being done properly.
Wrapping up the selection, verification and counting process inside computers reduces the amount of people who can understand what's happening by orders of magnitude. It doesn't really matter if the voting system is open source, well designed and administered, or whatever. It's always going to exclude the majority of the population from being able to fully understand how it works, and to trust that it's working properly.
It's quite possible that IT systems can help with elections, and they already are in some places, but I don't personally think they should be used at the expense of a manual process, and I don't think they should be depended on for anything other than an early indication of the result. Voting machines, when used, should always provide voter-verified paper trails that are always deposited in a secure box in a voter-verified way using a fully visible and voter-controlled process. Manual recounts should be mandatory if there's any reasonable doubt of the outcome by anyone.
Well I can't vouch for its authenticity, but someone who edited Wikipedia claims that "each shuttle was designed for a projected lifespan of 100 launches or 10 years' operational life". I couldn't say whether this is because some components would deteriate too much after 10 years, or after 100 launches, but I guess part of the problem is that the shuttles haven't been launched anywhere near as frequently as the original plans projected, for a variety of reasons. (Budget cuts, accidents, etc). According to that projection, Columbia was more than twice the age of its orignally projected lifetime when it disintegrated on re-entry. (This isn't necessarily a definitive metric, though, given that there have only been two fatal accidents and Challenger was only a few years old when it exploded.)
It's a fair enough comment and I hope I didn't come across as condescending in any way, because I agree that it's a bit silly having a bug like this in a system being used today. I think more than anything, the problems are likely to be related to stretching the programme so far beyond what was intended. What remains is based on some very legacy technology, but probably very critical, given how much I understand the Shuttle's flight ability relies on its computer systems.
...which is more than many software development processes would reveal. Chances are that this known restriction is on a check-list which every shuttle mission has to be checked against, and the list would exist precisely because the software development and verification process is so solid and conservative.
Opinions vary, but I don't think I'd ever recommend working to the same standards, unless the customer actually had good reason to require it. (NASA does.) Even aside from your own code, doing it properly would require an extensive understanding of any and every third party library and system the code interacts with, which could add orders of magnitude to the development time and cost, even if it's open source and open hardware. I don't like hacks and yucky untested code any more than most people, but at some point it can just make sense to avoid extensive and pedantic formal development processes in favour of just getting it to work.
A lot of development processes (perhaps most) wouldn't have stopped the shuttle launching, even if this were reported as a bug. Chances are that it'd be forgotten about (if not fixed straight away), and someone would stumble on it again accidentally. Many bugs aren't even reported until someone's stumbled on them at least once. This is fine in most situations. Once it becomes a problem again, you can go and look it up, quickly find out everything that's known about it from before, apply any known workarounds, and spend time to fix it if necessary. The point, though, is that many systems wouldn't be sure to keep you informed about the restriction in a way that actively prevents someone stumbling on it later.
I still agree that it seems a little strange that this problem wasn't fixed ages ago. Realistically, though, the Shuttle was never expected to fly this long. It sounds a lot like a compromise that was made in the earlier days when computers were more limited, probably even moreso for the restricted range of systems that are certified to work under such conditions. Any update is likely to be very expensive and time consuming, simply because the software development and verification process is so solid and reliable.
From the article you quoted, it sounds more like they dropped a spacewalk (for Hubble maintenance, probably not safety-critical) so they could return sooner and avoid encountering the bug. To me it sounds like they did what they should have done, with safety as a priority.
Launching spacecraft is an industry in which the stakeholders usually prepare for possible or likely delays. NASA has to delay launches all the time for all sorts of reasons. I'm not sure why a possible software problem would be treated any differently. If the problem is with the managers dangerously forcing early launches, NASA should really be fixing their managers as a priority over fixing a known bug with a known workaround. Weighing it out, it's probably a lot cheaper, easier and safer to simply delay the occasional launch for a few more days, especially given that the Shuttle's remaining days are limited. Why risk the safety of future launches by making changes that will soon become obsolete?
Anyway, those are just my own thoughts. I don't work in software development where the process is quite so strict, but I'm sure they know what they're doing when they don't fix something like this.
Definitely. I've just gone and watched the demo, and read a bit about it. Good on these people for coming up with a system where it's (apparently) impossible to prove to anyone else who you voted for, yet still allows for someone to be able to verify their vote to some extent. That said, I still think that trying to solve this is really trying to fix the wrong problem.
The only reason receipts are wanted right now is because some voters have a lot of doubt about whether their vote was counted correctly. The problem could be solved much better by fixing the cause rather than trying to treat the symptoms. Letting people have receipts won't actually improve the validity of the election, anyway, it'll only help people feel better about themselves. It certainly doesn't mean that a reliable recount can take place, because the vast majority of people will never check their vote, keep their receipt, or bother to return it on request. At best it'll indicate that there's been a problem with the election, but it's not as if we don't already know that, and it hasn't taken voter-only-verifiable receipts to figure it out.
Trying to do something this complex on the scale of a national election, or even small elections that involve a typical cross-section of the general public, is asking for trouble. An election is trusted because it's kept simple, and the people who vote can see and have a reasonable understanding of how it actually works. People can understand the concept of writing a vote on a piece of paper, dropping it in a secure box, having trusted people empty the box and count the votes, and allowing other trusted people to observe the process at all stages.
Tying the whole thing into computers, digitized logic hidden inside electronic machines, abstract metaphor (such as dragging and dropping virtual objects), and abstract automated counting methods, reduces the number of people who can understand the entire process, let alone any of the process, by orders of magnitude. It just opens up more possibilities for misunderstanding, confusion and concerned citizens who no longer trust the process.
Perhaps this system can be used for other things, but I really hope we don't resort to using it in large scale elections. There are so many other very basic things that need fixing first, and I'm skeptical whether anything like this will be beneficial once the root causes of the problems have been dealt with.
In times when it's difficult for either major side to gain a clear majority, there are plenty of reasons for both major sides prefer an ambiguous loss rather than risk letting the other side have a close but definitive victory. For example, an ambiguous outcome makes it easier to determine the actual result in the courts, and having the right lawyers, or simply throwing more money at the problem, could be preferable to relying on unpredictable voters.
In a quote that's frequently repeated on Slashdot, Robert J. Hanlon once stated: "Never attribute to malice that which is adequately explained by stupidity." Given some of the stories that have emerged about voting machines in use now and in the past, I personally have trouble believing that malice is not involved.
My main question is: Do you believe that the current situation with e-voting in the USA can be adequately explained by incompetence among politicians and partisan election officials, and corporations to understand problems with the technology? Furthermore, if corruption is involved, and if people in certain positions are influencing the decisions to use unreliable voting systems, how serious-an-effect do you think this interference will have on efforts to make voting systems more reliable?
For this to work reliably, you'd really need a way to make sure that voters can't ever leave with their voting paper, either accidentally or deliberately, for two reasons:
Rather than relying on voters to deposit their paper votes themselves, a better way would be for the machine to print the paper vote and display it behind a window. The voter verifies that it's correct, and then manipulates a visibly simple, mechanical switch that either accepts their vote (and visibly drops it into the box), or rejects the vote, and visibly destroys it, or drops it somewhere clearly marked as a rejected vote. If the voter rejects the paper vote, of course, the digital vote gets withdrawn, the process re-starts, and they can select again.
An open source electronic voting system could certainly be beneficial, but I'm not convinced that any system will work unless the people who administer it are both capable and motivated to administer it properly.
In response to your other suggestion:
This should never be allowed in an anonymous election, because it means that votes are only anonymous for as long as the voter is not threatened to hand over their receipt, or to prove their vote to someone else, which defeats the purpose of having an anonymous election.
Besides, consider what would happen if the elected representatives did something that was unpopular. Want to cast doubt on the authenticity of the election after the event? Just find enough people to claim that their vote was recorded incorrectly. It's their word against yours.
All the idea offers is a feeling of satisfaction for voters, but it opens the door for more controversy. If voters need to feel satisfied that their votes were recorded correctly, it should be accomplished by having a stable and reliable system that all voters (within reason) can understand the complete workings of. Perhaps this would require a mandatory manual recount of the voter verified paper trail after the election to confirm the result, at least in areas where the count was reasonably close, but I think it'd be absolutely worth it to have a fair election.
I don't entirely disagree with your overall point, but the problem I have with this reasoning is that I use Linux, "consistent" usually means far less options, and I'm concerned that the options you happen to think are appropriate for some kind of plan to "market Linux to the masses" are unlikely to be what's optimal for me.
I think one of the major problems with this is that so many people are thinking of it as "Linux", when Linux really has little or nothing to do with the desktop that everyone sees.
It'd make much more sense to make KDE look consistent, and encourage people to use KDE, or make Gnome look consistent, and market that at people. Or maybe the groups in charge of KDE and Gnome might decide to combine the interfaces a bit more, and that could be marketed to people. Or maybe someone just builds some kind of super-distribution and gets lots of attention, and that distribution can be marketed at people. But it shouldn't be marketed as "Linux", because Linux is only a single component of a much larger system with many more obvious components, and the Linux part of the system is replaceable.
In particular, people don't have to tie themselves to Linux for this. They can tie themselves to any of the BSD's or to Solaris, just for starters, and still get a similar experience if they're running the same desktop. But trying to argue that "Linux should look consistent" will never work, because Linux is a kernel and it doesn't look like anything.
I agree with most of what you've said, but not quite with this. The way I remember things, Firefox's original claim to fame, at least among geeks, was not being Mozilla, as in Mozilla Seamonkey. In other words, it was Mozilla with all the extra flab cut out. Just a browser. No mail client, no composer, no newsreader, no built in IRC chat client, etc etc. This made it much faster loading, and eventually more comparable with Internet Explorer as a drop-in replacement.
It's still just a browser, but since then it seems to have been becoming a somewhat more heavy duty browser trying to meet everyone's needs, without necessarily doing everything brilliantly. The irony for me is that Mozilla Seamonkey can still be installed in modules to some extent, so it's actually possible to install just the browser. (Well, the browser seems to come with the Composer in my debian package, I'm not sure if that's standard.) I haven't done any extensive benchmarks, but anecdotally it does often feel more cut down to me than Firefox does. To be honest, it feels more stable and I almost prefer it over Firefox.
The full Mozilla Seamonkey suite seems to have been forgotten in all the Firefox hype, but the only thing that's putting me off using it more often is that it doesn't seem to have a very reliable extension handling interface. That is, I can install an extension from an XPI file, but there's not actually an extension management interface in the front end UI that I could find. Perhaps this is a good thing, because it diversity of so many third party extensions of varying quality seems to be one of the big complications of Firefox stability.
From here, regarding Chicago:
Given that they're migrating their Solaris server boxes rather than Windows desktop workstations (as seems to be happening in Munich), I'm not sure if it's quite the same thing.
Here's an interesting chart.
Perhaps this link was added to the slashdot summary after you posted your comment for all I know, but the slashdot summary that I read had two links, and I found that statement quite clearly after following the first link. About the 13th paragraph down in that article states, complete with the additional link that I've included here:
So no points for grammar in that sentance (which I copied verbatim), but it seems to explain quite clearly what the Microsoft criticism is. That second linked article begins with the paragraph:
...and goes on for quite a while. Is this the statement you meant?
I think this really depends on the definition used for "making the grade". Just because something succeeds doesn't make it great, or even good, if you happen to measure greatness by a different metric. The populace determines a lot of things, including an adequate quality of television, and an adequate quality of elected government legislators. I don't agree that either of these meet my own expectations of adequacy, on most occasions, and I can fully appreciate that any individual should be able to decide that they also don't want to accept something simply because "the populace" does.
I use Wikipedia a lot, and I contribute to it a lot. I love it because it has such a wide range of information about all sorts of things, which is very difficult to find elsewhere. There aren't many other places where you could find such an in-depth description of something like the Slashdot subculture, for instance, and I'd consider directing people to that article if they wanted a summary of the site.
But I use Wikipedia cautiously, with an understanding of how it works, and that the information I get from it probably isn't authoritative and will need verifying if I'm using it for something important. If I relay the information to other people, I'll usually indicate where I got it from, and do my best to point out that it could potentially be suspect if I think it's appropriate to do so. I'm not sure that most of the populace has the same cautious approach. I'm beginning to see more and more people treat Wikipedia as if it's just automatically authoritative information, without any realistic understanding of what it actually is and where it comes from. I find that a bit disconcerting.
All that said, I guess with people's choices of TV and political representation, and many other things, there are quite a few things that I have trouble relating to in the "general populace".
Thanks for the suggestions. I'll try them out on Monday when I get back to work.
I've been trying to change my default browser at work, where we tend to use a very Microsoft-centric desktop for most things. (Outlook, IE, Office, etc.) Switching the browser that runs underneath the Start Menu browser icon is quite easy, but I still haven't figured out how to make all applications use Firefox, or any other browser for that matter.
The most obvious example is Outlook, which seems to want to open links in MSIE no matter what I do. It' be nice to be able to tell Outlook to use another rendering engine entirely (such as Gecko), but I'd be keen to simply tell it to run Firefox instead of IE when it wants to open a completely new browser instance. Does anyone know how to change this? Is it even possible with Outlook?
I guess my point is that as long as the enduring experience is good, it opens up other options. For one thing, it'd make it easier for OEM's or retailers to install it for you, which is effectively what they do with Windows at the moment. It also opens up other possibilities, such as having someone else install it. (My mum and dad wouldn't care about the installer if I did it all for them, but they'd care a lot about being able to keep things working after I'd gone home.)
When I was first getting started in Linux (ages ago now), the installation was relatively easy. It took a while, and I had to figure out things I hadn't done much of before, such as boot managers and repartitioning. What really caused me to go back to Windows for a few more years, though, was that the whole system slowly degraded over time in ways that I couldn't figure out how to fix. The reason for this was that although the installer set things up initially, it didn't help to keep them running. I was basically dumped all at once into a system with a heap of configuration files for programs that I knew nothing about, most of which had no obvious configuration interface. The only way I knew how to reliably change some of the settings was to reinstall from scratch, and obviously I didn't want to do that.
I think a neat thing about a lot of distro's and desktop environments lately is that they're providing a slightly more intuitive interface for helping people to fix things when they break. I'm still not sure if they're quite there yet, but it's a lot better than it's been in the past.
I'm undecided whether I agree with RMS or not on this, but I don't really agree with what you've said... in this sentance at least. To me this implies that we should be blindly accepting the personality simply because it gets attention for things that I sometimes agree with.
That said, I'm all for promoting constructive criticism, and looking at what people are actually saying rather than focusing on who the person is.
This is one of the reasons that I get a little skeptical about how so many people like to judge a linux distribution by its installer. A lot of reviews I've read over time of linux distributions have focussed almost entirely on the installer. It's true that most people who use Linux (unlike Windows) will need to install it at some point, and installation is quite important because of that. But surely the entire enduring experience after the installation, including things like system maintenance to keep it running nicely, is at least as important.
In all honesty, though, what would retailers do about it? Ditch Microsoft and start shipping PC's with Linspire?
We do this in a lot of places too, and I think there are perfectly good reasons for it, including security. (eg. If my account ever gets hacked, someone probably still needs to know a much more secure password if they want to give the account more access.) Another is just plain robustness. It's harder for me to accidentally break things when I don't have access to them.
At least as importantly, though, I think it helps the users actually trust us more easily. Most of our users realise that we don't automatically have access to their documents, for instance. They also know from experience that we'll tend to ask them if we ever need to give ourselves access. Another example is with watching (and controlling) their desktops, usually for tech support. We could quite easily configure things to be able to connect at any time, but instead we make sure that whenever we do it, they first get a prompt which asks them if they'll give permission. Obviously it doesn't mean we can't do things without them knowing, but having policies about that sort of thing which people understand really makes it easier to work with the other people in the organisation.
I can see this in about:config, but where is it in the options dialog? (Keeping in mind that the vast majority of people likely to get frustrated by accidental tab closures will be people who don't know about about:config.) I'm using Firefox 2.0 right now, and the four options I have in the Tabs section are to "Warn when closing multiple tabs", "Warn when opening multiple tabs that might slow down Firefox", "Always show the tab bar", or to "Switch to a new tab immediately when opening".
More importantly, the decision to go this way seems to be making it much easier to accidentally close a tab when simply trying to switch between them. Suddenly the tab-switching bar is interspersed with close buttons. I'll be adjusting it in about:config straight away, but I'm a little concerned that as well as making it a default behaviour, there's not even an option visible in the preferences dialog for people to change it.
In defence of the devs, I guess there was an entire open development process in which to raise this earlier.