I've had a laptop with a touch screen (hp pavillon dv3) for two years and a bit, and I use touchpad AND keyboard AND touch screen to interact. When I want to select a big button or activate a window I find it far more convenient to touch the screen (with the back of my finger so it doesn't leave greasy prints), than wiggling the mouse around so I can see where the pointer is, moving it to the right place and clicking. (I sometimes even first touch approximately the point I want to hit and then move the mouse for fine tuning).
I'd even add that I miss that at work and when I use another (touchless) laptop. All the time I just want to bring that window to front or move a window away (I'm on linux so when the alt key is pressed (with my left hand) I can move a window around with my right hand as easily and naturally as moving a piece of paper around on my desk.
I get pain in my wrists and fingers due to mouse and keyboard usage, not the occasional touch.
I'd HATE having to do everything by touch however. I want my mouse AND my touch screen.
Indeed. Turing machines don't require infinite tapes, they require unbounded tapes. In particular the initial state of the tape must contain at most a finite number of non-blank cells. Working with a finite tape is therefore fine as long as you are ready to enlarge it when the head reaches the boundary (so that, to the machine, it appears infinite).
In the same sense, a physical computer could act as a Turing machine if, when it runs out of memory, an operator could come and plug in an extra hard drive (and if memory addresses were made in a way that they can be arbitrarily large).
I'm wondering something: How come this happens always at the same time of the year? That cloud of debris can't just sit still (relative to the sun), doesn't it rotate around, or fall into, the sun, as years pass by? That should make the phenomenon happen at a different time each year... From my understanding, ignoring gravity due to Earth and other planets, something sitting motionless at this distance of the sun would fall into the sun in a matter of months? Unless there's a huge trail going straight away from the sun getting slowly eaten like a spaghetti, that somehow isn't disturbed by other planets, that gets crossed by the Earth once a year. *confused*
http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html scroll down to section 14.16, Content-Range.
Most servers I've encountered support that header. For clients, wget has the --continue option (that works for both ftp and http as long as the server supports it).
I thought of exactly that! It reminded me of that very nice novel by Dan Simmons where he explores that exact theme. In his story, the brain evolved to block "brain waves" emitted by other people, but for some rare few, that doesn't work, and they could hear what other people were thinking. Maybe a bit far-fetched/not very realistic as the actual waves are probably far too faint and noisy, but a nice read all the same.
It seems fashionable to remove all functionality from browsers and then put them back into websites.
Now website designers have to add (necessarily non-standard) buttons in the web pages themselves to do things like "print", "go back", "change font size", "search in the page", "close", etc, that are all also provided by the browser, but that people no longer (know how to?) use...
That's the approach Unix has used for a long, long time now. Installed programs on a Unix system are generally root-owned and sit in directories that are also root-owned. For a normal user, both the executable and the directory in which it is located is read-only.
System-wide programs are stored in directories not writable by normal users, but that doesn't prevent a user from downloading a trojan into his own directory and running it, which is what the parent was talking about.
Unix systems do offer the option to mount/home (and other mount points like/tmp where the user has write access) with -o noexec which would close that issue, but I've never seen a linux distribution that would do that by default, because users expect to be able to run programs they've downloaded without having to jump through hoops.
You're totally right on the missing bridge to email, but the invite system is good in that it permits controlling the load on the system, by only giving as many invites as the system can handle. If they had opened it to everybody right away, it would have been *really* unusable due to overload. (And also makes people with an invite feel special and therefore want to use the system)
I agree the current situation is far from perfect (Ideally, the people at freedesktop.org would build a unified centralised password access protocol like they did with dbus etc, so applications developers wouldn't have to implement all existing protocols every time) but having each application implement its own strategy is worse.
Three reasons:
First, the user either has to type as many master password as there are implementations (Now I have to type three passwords when logging in: the session password, the kwallet password, and the firefox password because firefox doesn't integrate with kwallet) or store them in cleartext (or in an easily decrypted format). If I had to type one master password for each program that needs passwords (IM, browser, email, irc, gpg, ssh, etc), that would mostly defeat the purpose of them.
Secondly, having a single storage space enables sharing passwords securely between applications. Now I need to save my passwords separately for firefox, konqueror, and chrome. You'll say "stick to a single browser then" but it shouldn't have to be like that.
Third, writing your own implementation increases the risk of having bugs that lead to security holes, compared to a single implementation that got polished over time.
I'm not sure your statement that most users don't use those is right but know too little a sample to support my opinion (I don't know that many linux users but all of them, and not only experts, do use gnome keyring, and I use kwallet).
CanvasBlock?:)
They say it has stability problems but I'm sure those will get ironed out (possibly through an actual plugin rather than a greasemonkey script) when the need arises
Sorry but you are wrong. In the first part of your comment, the three cases (boy boy, boy girl and girl boy) are equally likely.
If you don't make the "one of the children is a boy" hypothesis, then these three cases all have 25% probability, and girl-girl has 25% probability as well.
You're right that same sex (girl girl + boy boy = 25% + 25%) and different sex (girl boy + boy girl = 25% + 25%) both have probability 50%.
But the problem says *knowing that one of the children is a boy*, what's the probability the other also is, i.e. what's the probability of the boy boy case assuming we aren't in the girl girl case. That gives 25% divided by 75%. IOW, knowing that one of the children is a boy, we're in case boy boy with probability 33%, boy girl with probability 33% and girl boy with probability 33%.
In the second part of your comment you're counting twice the families that have two boys, both born on a Tuesday, which is why you get probability 14/28=50% instead of 13/27.
What you said about the Monty Hall problem is correct.
Sorry, replying to myself but I forgot this: All window managers I've tested (kde's, gnome's and ion) lets you set a keyboard shortcut to put a window full screen, i.e. hiding window decorations and task bars for when I want a window completely full screen, whether the application supports it or not.
I constantly need to see many windows at a time... When coding I often need a documentation window, when designing a webpage I need to see the webpage while I type the source, when making a document I need to see the rendered version, when following instructions from a webpage for doing something in a terminal I need to see those two simultaneously, etc. Also I like keeping that IRC window at the back and just showing the last line below other windows so I can see what people are talking about without interrupting my work.
I only ever maximise windows when browsing the web or doing something in the terminal that doesn't require looking at documentation...
And the fact that it won't look at what's on the clipboard, and use those dimensions when I go to file->new
Shift-control-V creates a new buffer with the content (and dimensions) of the clipboard. If you want a *blank* new buffer of the same size as the copied object, do shift-control-V and then control-dot to erase the image (I'm not sure what's the need for this but as you can see it's easy to do)
Slashdot Mirror
My datapoint contradicts that claim a bit:
I've had a laptop with a touch screen (hp pavillon dv3) for two years and a bit, and I use touchpad AND keyboard AND touch screen to interact. When I want to select a big button or activate a window I find it far more convenient to touch the screen (with the back of my finger so it doesn't leave greasy prints), than wiggling the mouse around so I can see where the pointer is, moving it to the right place and clicking. (I sometimes even first touch approximately the point I want to hit and then move the mouse for fine tuning).
I'd even add that I miss that at work and when I use another (touchless) laptop. All the time I just want to bring that window to front or move a window away (I'm on linux so when the alt key is pressed (with my left hand) I can move a window around with my right hand as easily and naturally as moving a piece of paper around on my desk.
I get pain in my wrists and fingers due to mouse and keyboard usage, not the occasional touch.
I'd HATE having to do everything by touch however. I want my mouse AND my touch screen.
Indeed. Turing machines don't require infinite tapes, they require unbounded tapes. In particular the initial state of the tape must contain at most a finite number of non-blank cells. Working with a finite tape is therefore fine as long as you are ready to enlarge it when the head reaches the boundary (so that, to the machine, it appears infinite). In the same sense, a physical computer could act as a Turing machine if, when it runs out of memory, an operator could come and plug in an extra hard drive (and if memory addresses were made in a way that they can be arbitrarily large).
Personally I use bumblebee every time I want to clear my /usr, it works great! :)
I'm wondering something: How come this happens always at the same time of the year? That cloud of debris can't just sit still (relative to the sun), doesn't it rotate around, or fall into, the sun, as years pass by? That should make the phenomenon happen at a different time each year... From my understanding, ignoring gravity due to Earth and other planets, something sitting motionless at this distance of the sun would fall into the sun in a matter of months? Unless there's a huge trail going straight away from the sun getting slowly eaten like a spaghetti, that somehow isn't disturbed by other planets, that gets crossed by the Earth once a year. *confused*
http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html scroll down to section 14.16, Content-Range. Most servers I've encountered support that header. For clients, wget has the --continue option (that works for both ftp and http as long as the server supports it).
I thought of exactly that! It reminded me of that very nice novel by Dan Simmons where he explores that exact theme. In his story, the brain evolved to block "brain waves" emitted by other people, but for some rare few, that doesn't work, and they could hear what other people were thinking. Maybe a bit far-fetched/not very realistic as the actual waves are probably far too faint and noisy, but a nice read all the same.
It seems fashionable to remove all functionality from browsers and then put them back into websites. Now website designers have to add (necessarily non-standard) buttons in the web pages themselves to do things like "print", "go back", "change font size", "search in the page", "close", etc, that are all also provided by the browser, but that people no longer (know how to?) use...
I still haven't found a player for my Mac (or Linux laptop) that can run songs/movies at double speed without making everyone sound like chipmunks.
Not developed by Apple, but vlc does that fine (at least on my Linux box, with VLC media player 1.0.6 Goldeneye).
I'm afraid you got wooosh-ed, my friend.
I installed Folding@home for precisely that reason. I used to do a "yes > /dev/null" but then thought I could donate those cycles for something useful.
That's the approach Unix has used for a long, long time now. Installed programs on a Unix system are generally root-owned and sit in directories that are also root-owned. For a normal user, both the executable and the directory in which it is located is read-only.
System-wide programs are stored in directories not writable by normal users, but that doesn't prevent a user from downloading a trojan into his own directory and running it, which is what the parent was talking about.
Unix systems do offer the option to mount /home (and other mount points like /tmp where the user has write access) with -o noexec which would close that issue, but I've never seen a linux distribution that would do that by default, because users expect to be able to run programs they've downloaded without having to jump through hoops.
You're totally right on the missing bridge to email, but the invite system is good in that it permits controlling the load on the system, by only giving as many invites as the system can handle. If they had opened it to everybody right away, it would have been *really* unusable due to overload. (And also makes people with an invite feel special and therefore want to use the system)
I agree the current situation is far from perfect (Ideally, the people at freedesktop.org would build a unified centralised password access protocol like they did with dbus etc, so applications developers wouldn't have to implement all existing protocols every time) but having each application implement its own strategy is worse.
Three reasons:
First, the user either has to type as many master password as there are implementations (Now I have to type three passwords when logging in: the session password, the kwallet password, and the firefox password because firefox doesn't integrate with kwallet) or store them in cleartext (or in an easily decrypted format). If I had to type one master password for each program that needs passwords (IM, browser, email, irc, gpg, ssh, etc), that would mostly defeat the purpose of them.
Secondly, having a single storage space enables sharing passwords securely between applications. Now I need to save my passwords separately for firefox, konqueror, and chrome. You'll say "stick to a single browser then" but it shouldn't have to be like that.
Third, writing your own implementation increases the risk of having bugs that lead to security holes, compared to a single implementation that got polished over time.
I'm not sure your statement that most users don't use those is right but know too little a sample to support my opinion (I don't know that many linux users but all of them, and not only experts, do use gnome keyring, and I use kwallet).
CanvasBlock? :)
They say it has stability problems but I'm sure those will get ironed out (possibly through an actual plugin rather than a greasemonkey script) when the need arises
A proof that video quality does have long term effects: http://www.xkcd.com/598/
Sorry but you are wrong. In the first part of your comment, the three cases (boy boy, boy girl and girl boy) are equally likely.
If you don't make the "one of the children is a boy" hypothesis, then these three cases all have 25% probability, and girl-girl has 25% probability as well.
You're right that same sex (girl girl + boy boy = 25% + 25%) and different sex (girl boy + boy girl = 25% + 25%) both have probability 50%.
But the problem says *knowing that one of the children is a boy*, what's the probability the other also is, i.e. what's the probability of the boy boy case assuming we aren't in the girl girl case. That gives 25% divided by 75%. IOW, knowing that one of the children is a boy, we're in case boy boy with probability 33%, boy girl with probability 33% and girl boy with probability 33%.
In the second part of your comment you're counting twice the families that have two boys, both born on a Tuesday, which is why you get probability 14/28=50% instead of 13/27.
What you said about the Monty Hall problem is correct.
I don't think chat networks like Yahoo, MSN chat (or whatever it's called today), ICQ encrypt the passwords...
Sorry, replying to myself but I forgot this: All window managers I've tested (kde's, gnome's and ion) lets you set a keyboard shortcut to put a window full screen, i.e. hiding window decorations and task bars for when I want a window completely full screen, whether the application supports it or not.
I constantly need to see many windows at a time... When coding I often need a documentation window, when designing a webpage I need to see the webpage while I type the source, when making a document I need to see the rendered version, when following instructions from a webpage for doing something in a terminal I need to see those two simultaneously, etc. Also I like keeping that IRC window at the back and just showing the last line below other windows so I can see what people are talking about without interrupting my work.
I only ever maximise windows when browsing the web or doing something in the terminal that doesn't require looking at documentation...
Ah, this. :)
That video got deleted for not complying to site rules or something. What was it?
Someone posted this in the previous H.264 versus Ogg story, which gives a data point including screenshots pointing to ogg being better: http://people.xiph.org/~greg/video/ytcompare/comparison.html
Why call the rule FH_DATE_PAST_20XX, shouldn't it be FH_DATE_PAST_201X? At least then the hack would be documented.
Changing the rule name would break existing configuration files changing the rule score and/or description (like in that comment).
And the fact that it won't look at what's on the clipboard, and use those dimensions when I go to file->new
Shift-control-V creates a new buffer with the content (and dimensions) of the clipboard. If you want a *blank* new buffer of the same size as the copied object, do shift-control-V and then control-dot to erase the image (I'm not sure what's the need for this but as you can see it's easy to do)
you got it backward - it should be tail -f | grep -v