When you press both buttons at the same time or simply depress the top half of the mouse, it left clicks.
Oh, and here's another:
I tested right-clicking with varying degrees of skin contact on the left side of the mouse and you can't be touching it "too much" on the left side before the right click gives up and you're now left clicking.
I don't remember what happens with a regular 2-button mouse if you click both buttons no the desktop, and I don't have an OS X machine to test on right now. I thought it behaved differnetly from just a left click, though I could be mistaken. In any event, I see no indication that chording might work someday, and several indications that it does not work now.
Hey, no fair replying seriously to my utterly non-serious post.:)
I'll have to tell you that I've set up KDE on a machine that really is shared between everyone from a professional sysadmin to someone who still has a hard time differentiating between when to right or left click, and then I'll have to express appreciation for teh work Apple has done...
And there's no guarantee that you'll ever leave the city, so does the unsubstantiated claim that a Porsche might do 150 mean anything if you'll never be able to verify it?
It may also be that the click interpretation is due to a designed-in limitation/feature of the hardware. Given that both buttons were supported under Windows using the default Windows HID driver, and past experience shows that driver to support chording, I'd say that there's a pretty good possibility that the click interpretation is happening in the mouse and not the driver. If the mouse was capable of sending both click events, then the windows driver should have caught them. Of course, this is mere speculation, albeit grounded on actual verifiable observations, and it could be that our faithful reviewer never tested chording on Win32 / in a Win32 app that supports it.
It does seems plausible that Apple would have added that logic to the already complex mouse hardware, though, rather than making the driver more complex. They do have a history of making some poor implementation decisions with respect to long-term functionality.
To further support your argument, I've included with this post the beginnings of source code to a driver that doesn't do anything "right now", but one day will make this mouse capable of shooting fire from the LED. I mean, just because the driver that ships with the mouse and another third-party driver don't support it now, why should I refrain from claiming that one day the apple MightyMouse will certainly be able to shoot fire from the LED?
Oh, I almost forgot to make my argument valid: "Fuck, son of a bitch, idiot."
I'd like to see someone name packages the same (ie, don't put version numbers in the package names) and use rsync to keep a local repository in sync with a remote repository. Something along the lines of apt's cache.
BTW, several Gentoo packages have updates that consist of a patch to the original source, and you just pull down the patch if you already have the source.
If you don't have the fourth disk, then the read is presumed valid until the pairity disk is replaced and regenerated. Under normal circumstances, the implementation typically is that the pairity is calcualted and compared, despite that other ways are conceivable.
Anyway, you would want to skip the pairity block if atttempting to get maximum speed, so you could avoid the calculation all together.
Yeah, I should take OS X and NeXTStep off of my resume, because some poster on Slashdot - who thinks that a display system equals an operating system and thinks that the internet super server is the same thing as init - wants to argue about semantics. OS X is just another *nix with a pretty GUI on top - it's not the revolution fanatics want to believe.
BTW, NetInfo is going away, and inetd isn't process 1 on anything common. Normally, init is process ID 1, on BSD and OS X (and Linux). Here's the OS X machine sitting on the desk behind me:
cloudmaster@yyy ~ $ ssh zzz cloudmaster@zzz's password: Welcome to Darwin! zzz:~ cloudmaster$ ps 1
PID TT STAT TIME COMMAND
1 ?? Ss 0:02.04/sbin/init
It's the same on all of the 10.2, 10.3, and 10.4 machines, and on the FreeBSD box in my basement.
Maybe one day soon I'll tar up Linux with QTopia as the UI and use LDAP to store user information, and remove RPM from my distribution. Maybe I'll add a couple of printks to the Linux kernel, and a hook to activate them (I'll call it the "spice" kernel). Then I'll inlcude the libraries to support QTopia and LDAP, and call it cloudnix. Then I can just sit back and wait for some people on/. to get into a discussion about how cloudnix isn't Linux.
For the pairity calculation on read to be useful, it has to be compared to the pairity bit that was written. So, you have to read from all four disks in a four-disk array. You read from three to get the data, read from the fourth to get the pairity bit, then calculate the pairity to verify that your read was valid.
The firmware goes a long way to balancing the preformance across all of the drives efficiently, and to intelligently buffering the data before it goes out to the bus. Not to mention the driver dealing with the requests. Sure, it's not the hardest thing to do, but it's also not hard to mess up an efficiency tweak here and there. Like, for example, grouping requests that are close to each other on the same drive, saving seek time. Etc.
The old slow PCI bus may be the bottleneck, but not until you get at beyond 4 drives or so on the controller (and have other system devices contending for PCI resources). Sub-par performance is more likely a driver issue - either software or hardware (or a possibly a tuning issue, which shouldn't be on consumer-level stuff, IMHO).
Here's a random 7200RPM Seagate SATA drive. Note that the average sustained transfer rate is greater than 58MB/sec. Most any midrange [7200RPM] drive (UATA/100 or SATA) should be able to sustain 50MB/s at this time, and a 5400RPM drive oughtta be capable of a solid 35-40.
A RAID-1 does have performance benefits on the read side, since the read can be spread out across multiple drives. That said, for overall use, a RAID-5 gives the best bang for the buck. In specialized situations where redundancy and read speed are most important, but capacity is not, a RAID-1 is the cost leader over the capacity afforded by a RAID-10. However, a properly designed RAID-10 with the same capacity will beat a RAID-5 for writes (no pairity calculation), and be faster on reads (more heads), *and* live through at least one drive failure without the same performance degredation a RAID-5 would feel.
Of course, this all varies with the quality of the implementation. Cheap cards are generally cheap because they're not worth much money. Smart programmers aren't cheap. Put those two facts together, and we may have found the performance problem with some of the inexpensive controllers.:)
BTW, yes, I'm primarily working on storage system benchmarking and performance tuning right now. Thanks for asking.;)
I guess you didn't read the differences, and you've never used OS X (or BSD).
Ok, so changing the display system changes the OS. Are you saying that my Linux system running X.org is different from one running Xfree? Or that mine running Windowmaker is not Linux, while one running twm is? No, those are all Linux, or all BSD, and the article wasn't about the user interface.
The utilities are all BSD, except for the *additional* utilities for OS X-specific stuff (like the things for netinfo, the disk imager, etc). The libraries are all the same, except for a few changes (can count them on one hand), except for the things they added for the GUI. The API is the same, except for the extra GUI stuff.
OS X is not a new OS. It's a new GUI on top of BSD, with a couple of minor changes from the specific BSD they chose as their OS. It may be a different user experience, but it's certainly not a new OS.
If you want ultimate *write* performance (like for audio/video/etc recording), you don't want RAID5 and the pairity calculation that comes with each write. You want RAID0. Faster reads come from RAID1. Data security comes from the other RAID levels (preferably 5 or 6, I'm partial to 10). Also, you got screwed if you're only getting 50MB/s out of $900 worth of hardware RAID - you should get that out of a single modern drive.
So, I sent them a URL, and when they go to look at it, something's wrong with either the server, their internet conection, etc - and then *I* have to troubleshoot their internet conenction to figure out why they can't see my document. If I had sent a PDF then they would have the file when they received the email, with no more steps to go through.
A collection of files would be convenient on the surface, but when the prof expects stuff in a word doc, asking them to deal with extracting a.zip and maintain the structure is generally extra work that they just won't do. "The clsss requirements say that assignments should be in a word doc" and "no grade for you" are the typical response. Similarly for suggesting that they use FireFox. If you think that most college-level instructors are gonna see the error of their flawed method of working based on the suggestion of some know-it-all Linux punk, when "everyone else" manages to do things the way he asks, you've been out of academia for a while. Getting them to take a PDF is generally tough enough - I had one where I had to rename.rtf files with a.doc extension so he's think I was sending Word docs.:(
As a sysadmin, I can tell my users what they oughta be doing, and they'll generally listen to me. In that role, all of my marked-up documentation is HTML-based. In my pursuit of education, though, I was just another damned student getting in the way of research time, and PDF is the easiest way to essentially work around that kind of idiot end user. It's not a good use, but it's unfortunately a common one.
I would like to bring up the east of transfer gained from the use of PDF over HTML. With HML, there's no good way to include a bunch of images in a single directly readable file. So, to transfer a file to someone, I have to send either as a.zip or as a whole bunch of files. I've taken several on-line courses that require turning in reports via email, and a.zip'ed collection of files is difficult to manage for the receiving instructor. So, I used PDF for most of my reports, as it ensures a controlled presentation (as my grade is somewhat dependent on how readable the content is) in a nice little bundle that I'm sure the instructor can read - and more importantly, I'm sure that the instructor will already have a reader (other formats that require extra work, possibly including.zip, are out of the question for many lazy recpients). Since I work on Linux, but he'll likely be on a windows workstation using IE, I don't want to forget about one of the tiny IE/firefox incompatabilities when I'm formatting my HTML. PDF frees me from that concern.
I do agree, that PDF is *way* overused, and is generally irritating. I prefer to use HTML (or plain text) when possible, but in cases like this it would take more significantly work on my part to get similar functionality out of HTML as I already can get from PDF.
There's a lot of extra stuff in your version, largely due to not taking advantage of things behaving differently in array and scalar contexts. readdir in array context will return all of the elements in the dir, which you can then grep for the entries matching your criteria. Similarly, there's no reason to have $count *and* $arraysize, or really either one. @array will give you the number of elements in the array when called in scalar context - like in rand($). Finally, there's no reason to use the backticks, since you don't care about the output. You just want a return code, at best, so use system(). As an added benefit to using system(), you can specify the program and arguments in a list - so this should work reliably with files that have spaces in their names.
We'll ignore all of the checks that oughtta be in place (like verifying that entries are actually files and not dirs named "collection of.wav" and that the path seperators are correct), of course - just for those who will inevitably further correct my version.
There's always something to learn.:)
#!/usr/bin/perl $SOUNDDIR="/usr/share/sounds"; $ DESTFILE="/tmp/sound.wav"; opendir(DIR,$SOUNDDIR) || die "Can't open $SOUNDDIR: $!\n"; @files = grep(/\.wav$/, readdir(DIR));
The Focus SVT would have made a good "young people's" car, but it just wasn't advertised enough. Yeah, I had one a few years ago, but sold it. I like torque, and kids these days, well, they just don't understand the importance of torque.
What? This is totally arbitrarily shoving elements into position so they sort of look like a spiral - there are big gaps in the pattern. It's just like looking at a regular periodic table through a curved lens, with a picture of "outer space" behind it.
Wow, revolutionary! Kids will suddenly be excited about science instead of pot!
Slashdot Mirror
Oh, and here's another:
I don't remember what happens with a regular 2-button mouse if you click both buttons no the desktop, and I don't have an OS X machine to test on right now. I thought it behaved differnetly from just a left click, though I could be mistaken. In any event, I see no indication that chording might work someday, and several indications that it does not work now.
Hey, no fair replying seriously to my utterly non-serious post. :)
I'll have to tell you that I've set up KDE on a machine that really is shared between everyone from a professional sysadmin to someone who still has a hard time differentiating between when to right or left click, and then I'll have to express appreciation for teh work Apple has done...
And there's no guarantee that you'll ever leave the city, so does the unsubstantiated claim that a Porsche might do 150 mean anything if you'll never be able to verify it?
n e/ :)
To veer off topic a bit: http://www21.porsche.com/usa/models/cayenne/cayen
Relevent quote: "Top Track Speed: 133 mph". It may be that the Porsche really can't get up to 150.
It may also be that the click interpretation is due to a designed-in limitation/feature of the hardware. Given that both buttons were supported under Windows using the default Windows HID driver, and past experience shows that driver to support chording, I'd say that there's a pretty good possibility that the click interpretation is happening in the mouse and not the driver. If the mouse was capable of sending both click events, then the windows driver should have caught them. Of course, this is mere speculation, albeit grounded on actual verifiable observations, and it could be that our faithful reviewer never tested chording on Win32 / in a Win32 app that supports it.
It does seems plausible that Apple would have added that logic to the already complex mouse hardware, though, rather than making the driver more complex. They do have a history of making some poor implementation decisions with respect to long-term functionality.
To further support your argument, I've included with this post the beginnings of source code to a driver that doesn't do anything "right now", but one day will make this mouse capable of shooting fire from the LED. I mean, just because the driver that ships with the mouse and another third-party driver don't support it now, why should I refrain from claiming that one day the apple MightyMouse will certainly be able to shoot fire from the LED?
Oh, I almost forgot to make my argument valid: "Fuck, son of a bitch, idiot."
--begin source--
--end source--
I'd like to see someone name packages the same (ie, don't put version numbers in the package names) and use rsync to keep a local repository in sync with a remote repository. Something along the lines of apt's cache.
BTW, several Gentoo packages have updates that consist of a patch to the original source, and you just pull down the patch if you already have the source.
Wait, you share a computer with your grandparents?
I usually use Dia to create network diagrams, and then use Inkscape to clean them up (Dia's interface could use some work).
If you don't have the fourth disk, then the read is presumed valid until the pairity disk is replaced and regenerated. Under normal circumstances, the implementation typically is that the pairity is calcualted and compared, despite that other ways are conceivable.
Anyway, you would want to skip the pairity block if atttempting to get maximum speed, so you could avoid the calculation all together.
BTW, NetInfo is going away, and inetd isn't process 1 on anything common. Normally, init is process ID 1, on BSD and OS X (and Linux). Here's the OS X machine sitting on the desk behind me:It's the same on all of the 10.2, 10.3, and 10.4 machines, and on the FreeBSD box in my basement.
Maybe one day soon I'll tar up Linux with QTopia as the UI and use LDAP to store user information, and remove RPM from my distribution. Maybe I'll add a couple of printks to the Linux kernel, and a hook to activate them (I'll call it the "spice" kernel). Then I'll inlcude the libraries to support QTopia and LDAP, and call it cloudnix. Then I can just sit back and wait for some people on
For the pairity calculation on read to be useful, it has to be compared to the pairity bit that was written. So, you have to read from all four disks in a four-disk array. You read from three to get the data, read from the fourth to get the pairity bit, then calculate the pairity to verify that your read was valid.
The firmware goes a long way to balancing the preformance across all of the drives efficiently, and to intelligently buffering the data before it goes out to the bus. Not to mention the driver dealing with the requests. Sure, it's not the hardest thing to do, but it's also not hard to mess up an efficiency tweak here and there. Like, for example, grouping requests that are close to each other on the same drive, saving seek time. Etc.
The old slow PCI bus may be the bottleneck, but not until you get at beyond 4 drives or so on the controller (and have other system devices contending for PCI resources). Sub-par performance is more likely a driver issue - either software or hardware (or a possibly a tuning issue, which shouldn't be on consumer-level stuff, IMHO).
BTW, on RAID5 pairity *should* be checked on all reads, which is faster than generating, but slower than not doing it at all. :)
RAID1 reads are as fast as RAID0 reads given the same number of drives. Writes are as fast (slow?) as one drive.
:)
Otherwise, I'll generally agree.
Good site, BTW.
Here's a random 7200RPM Seagate SATA drive. Note that the average sustained transfer rate is greater than 58MB/sec. Most any midrange [7200RPM] drive (UATA/100 or SATA) should be able to sustain 50MB/s at this time, and a 5400RPM drive oughtta be capable of a solid 35-40.
:)
;)
A RAID-1 does have performance benefits on the read side, since the read can be spread out across multiple drives. That said, for overall use, a RAID-5 gives the best bang for the buck. In specialized situations where redundancy and read speed are most important, but capacity is not, a RAID-1 is the cost leader over the capacity afforded by a RAID-10. However, a properly designed RAID-10 with the same capacity will beat a RAID-5 for writes (no pairity calculation), and be faster on reads (more heads), *and* live through at least one drive failure without the same performance degredation a RAID-5 would feel.
Of course, this all varies with the quality of the implementation. Cheap cards are generally cheap because they're not worth much money. Smart programmers aren't cheap. Put those two facts together, and we may have found the performance problem with some of the inexpensive controllers.
BTW, yes, I'm primarily working on storage system benchmarking and performance tuning right now. Thanks for asking.
I guess you didn't read the differences, and you've never used OS X (or BSD).
Ok, so changing the display system changes the OS. Are you saying that my Linux system running X.org is different from one running Xfree? Or that mine running Windowmaker is not Linux, while one running twm is? No, those are all Linux, or all BSD, and the article wasn't about the user interface.
The utilities are all BSD, except for the *additional* utilities for OS X-specific stuff (like the things for netinfo, the disk imager, etc). The libraries are all the same, except for a few changes (can count them on one hand), except for the things they added for the GUI. The API is the same, except for the extra GUI stuff.
OS X is not a new OS. It's a new GUI on top of BSD, with a couple of minor changes from the specific BSD they chose as their OS. It may be a different user experience, but it's certainly not a new OS.
Monkeys are not at all cute. Reference "you look like a monkey" and its common use as an insult for suport that others agree.
Umm, OS X is (Free)BSD with a few relatively minor changes. Those changes don't change that it's still BSD.
If you want ultimate *write* performance (like for audio/video/etc recording), you don't want RAID5 and the pairity calculation that comes with each write. You want RAID0. Faster reads come from RAID1. Data security comes from the other RAID levels (preferably 5 or 6, I'm partial to 10). Also, you got screwed if you're only getting 50MB/s out of $900 worth of hardware RAID - you should get that out of a single modern drive.
In that case, we totally agree. :)
So, I sent them a URL, and when they go to look at it, something's wrong with either the server, their internet conection, etc - and then *I* have to troubleshoot their internet conenction to figure out why they can't see my document. If I had sent a PDF then they would have the file when they received the email, with no more steps to go through.
.zip and maintain the structure is generally extra work that they just won't do. "The clsss requirements say that assignments should be in a word doc" and "no grade for you" are the typical response. Similarly for suggesting that they use FireFox. If you think that most college-level instructors are gonna see the error of their flawed method of working based on the suggestion of some know-it-all Linux punk, when "everyone else" manages to do things the way he asks, you've been out of academia for a while. Getting them to take a PDF is generally tough enough - I had one where I had to rename .rtf files with a .doc extension so he's think I was sending Word docs. :(
A collection of files would be convenient on the surface, but when the prof expects stuff in a word doc, asking them to deal with extracting a
As a sysadmin, I can tell my users what they oughta be doing, and they'll generally listen to me. In that role, all of my marked-up documentation is HTML-based. In my pursuit of education, though, I was just another damned student getting in the way of research time, and PDF is the easiest way to essentially work around that kind of idiot end user. It's not a good use, but it's unfortunately a common one.
I would like to bring up the east of transfer gained from the use of PDF over HTML. With HML, there's no good way to include a bunch of images in a single directly readable file. So, to transfer a file to someone, I have to send either as a .zip or as a whole bunch of files. I've taken several on-line courses that require turning in reports via email, and a .zip'ed collection of files is difficult to manage for the receiving instructor. So, I used PDF for most of my reports, as it ensures a controlled presentation (as my grade is somewhat dependent on how readable the content is) in a nice little bundle that I'm sure the instructor can read - and more importantly, I'm sure that the instructor will already have a reader (other formats that require extra work, possibly including .zip, are out of the question for many lazy recpients). Since I work on Linux, but he'll likely be on a windows workstation using IE, I don't want to forget about one of the tiny IE/firefox incompatabilities when I'm formatting my HTML. PDF frees me from that concern.
I do agree, that PDF is *way* overused, and is generally irritating. I prefer to use HTML (or plain text) when possible, but in cases like this it would take more significantly work on my part to get similar functionality out of HTML as I already can get from PDF.
We'll ignore all of the checks that oughtta be in place (like verifying that entries are actually files and not dirs named "collection of
There's always something to learn.
The Focus SVT would have made a good "young people's" car, but it just wasn't advertised enough. Yeah, I had one a few years ago, but sold it. I like torque, and kids these days, well, they just don't understand the importance of torque.
Macs have lots of useful keyboard shortcuts. Mice are only useful if you're not doing things that require typing.
What? This is totally arbitrarily shoving elements into position so they sort of look like a spiral - there are big gaps in the pattern. It's just like looking at a regular periodic table through a curved lens, with a picture of "outer space" behind it.
Wow, revolutionary! Kids will suddenly be excited about science instead of pot!
What? BONE, available to the public? I guess my order for Duke Nukem Forever will probably be filled soon, then...