One more note between the X25-E and X25-M...
The X25-E is branded as an enterprise-level disk, and is correspondingly much more expensive. It is a SLC design (single layer controller), where the X25-M is a MLC design (multi-layer controller). SLC chips are expected to last considerably longer than MLC, because fewer bits are stored in a single cell in an SLC chip, each cell is expected to be written fewer times when writing or modifying data on the disk.
But still, as mentioned in numerous comments here, the X25-M should* last long enough for most consumer uses.
*- time may prove us wrong, but I hope not.
Sequential reads are certainly sequential reads, but that might not mean exactly what you think it means.
In a spinning disk HDD, a sequential read means move the head to the correct location, and then stream the data as the disk spins under the head. This is roughly as fast as it gets with spinning media, ignoring cache etc.
In a SSD, you can actually benefit from parallelisation. SSDs can (if supported by the controller) read simultaneously from different chips, which means either reading the file in a number of parallel sequential reads simultaneously, or being able to satisfy multiple read requests simultaneously, as long as they are stored in different memory chips.
So defragmentation in the classical sense may not actually be a desirable goal in the new world order.
Why does everyone insist on piping grep to awk when awk can also perform the regex?
Actually, there are more reasons than you might expect...
Familiarity. More people are likely to recognise grep regular expressions as search terms than awk regular expressions, even though they are often identical. Grep just means search, at least to a geek.
Substitutability. You can replace grep with any other tool that can process stdin, without worrying about which argument is actually the file providing the data. This can be very useful if you're logically building up the command.
Efficiency. Cat does nothing but open a file and read it. It probably does that extremely well. Grep is primarily a search algorithm, and its file handling routines might not be optimal; it might hold file descriptors open longer than necessary, or whatever. In grep's case, it probably doesn't, but without looking at (and understanding) the source of what you're actually using you don't actually know. Of course, if you use the right tool for each part of the job, then you don't even have to care.
Permissioning. Linux/unix/posix doesn't actually do this, but you could run grep with a very restricted permission set; it is allowed to read from stdin, write to stdout, and allocate a certain amount of memory, and that's it. Even if your grep was somehow compromised, it couldn't do anything but print bad stuff to stdout. The cat process has a higher level of trust, and is allowed to read from the file system, and write to stdout. If you happened to be doing a multi-file grep, then you could invoke grep with a higher level of permissions so it can read the filesystem itself. If you're running an object permissioned OS (eros, keykos, coyotos, or a number of other OSes that aren't actually ready for use yet) then cat | grep is actually the recommended way to do things.
But mostly, it's just what they're used to. There's nothing wrong with it, it's roughly the same efficiency either way, so do it the way that makes sense to you, and for the problem you're solving.
And don't forget words that cannot be transcribed without context, such as their/there/they're and which/witch... "I'm going over there, which is useful because I need to talk to whoever anyway" vs "I wouldn't give their witch any credibility".
After reviewing with some colleagues including 1 who was an attorney, the general consensus was that it was indeed 6 words. It's also a contraction, which is pronounced as 1 word, but if you diagram that sentence, you have to separate the subject and the verb.
Does the same reasoning hold for the sentence It isn't Windows? That is, does the word/contraction isn't count as a full-fledged verb?
Reading this article, I wondered... how much time are these supercomputers around the world actually putting into useful calculations, and how much time into benchmarks to prove how fast they are?
I want to see another benchmark... how many TFlops per year are being spent on benchmarks vs useful applications!
That doesn't make sense... If the board is in a certain state, why does it matter if it reached this state by moving a pawn or by moving a knight, if it is black's turn to move?
Being able to check for a draw isn't necessary in the state tree... it simply means we need a graph with non-directional cycles to represent play. Note that if we enforce that the graph may not contain any directional cycles, then the construction of the tree will never recurse, so each particular board state only needs to be evaluated once.
That doesn't get past the fact that no computer (that I've heard of) has sufficient resources to maintain such a tree, though. That's not to say it's not possible to store a sufficient tree to allow perfect play... I don't know enough to guess either way.
Unlike up2date and its counterparts in the other distros, however, Windows Update just updates the base OS, so you have to take additional steps to update your word processor, C++ compiler and such.
This raises an interesting question. If Microsoft were to encourage developers to register their applications and patches in an accessible repository, so Windows Update could offer to optionally patch any and all software installed on your machine, this would be considered an action intended to increase their monopoly. (Whether or not cost was involved, to the developer or the customer)
If they were to compile their own database of software patches from third party products, this would probably be considered even worse, since they would (probably) only include common applications, so niche software wouldn't get a look-in.
Why then can Red Hat have such a database of patches and it's a good thing? Surely it's doing the same thing?
The only thing I might wish for that it doesn't have is some kind of a 9x virtual machine to run a legacy game or two I have.
You might be interested in VDMSound
The only games I've found that it doesn't run appear to have issues with the CPU (UFO, Legacy of Kain)... I haven't gotten around to trying an emulator yet.
Ten years before my time? There's a big assumption. Just for the record, I got started on text-only adventure games long before Sierra came into the picture (no pun intended).
...when winning the game meant hunting-the-pixel...
There was no reason to hunt pixels in the original Sierra *Quest games... the entire interactive interface was still through typing, I don't think mouse was even supported! You just happened to have a visible avatar that you could maneuver through the traditional adventure 'room' by using the arrow keys. My comment in reply to the grandparent post, was suggesting that this particular addition of pictures to the adventure game was a far from academic exercise, and actually added a lot to the game. At least, IMHO.
I agree with you 99% that today's graphical adventures are mostly an exercise in attempting to find the combination of clicks required to solve a particular 'puzzle'. Then again, some of the puzzles in the old days devolved into 'hunt the exact spelling of the exact verb/noun the game designer intended to be used in this situation', which was arguably more frustrating.
Adding graphics to Nethack is just a waste of time - if it's an academic exercise fine, but please don't ask people to play it. You may as well take text adventures and make them "better" by adding pictures.
You're kidding, right? King's Quest, Space Quest, Hero's Quest (aka Quest For Glory), Police Quest? They rocked my world, back when they were still text adventures:-)
I was hoping for a home printer that could hold a ream of paper and have separate trays for labels and envelopes and plug and play networking.
One of my best purchases ever was a second hand LJ 4M Plus for AU$150 (US$80?). Fast printing, great quality, and plugs directly into the network without requiring a host computer, and drivers were already installed in Win2k... The hardest part of installation was setting up the printer's IP and Gateway addresses, since it didn't support DHCP, but now it runs beautifully!
In your opinion, what is more educational, a computer simulation or a book?
To rephrase the question (disclaimer: Probably violating the spirit of the parent poster), which is more educational, a Harry Potter novel or a Harry Potter computer/console game? The book teaches literacy through example, while the game (hopefully) teaches problem solving and hand-eye manipulation.
A lot of older games used to cover the literacy front as well - adventure games in particular - but today's push for stunning graphics and voiceovers is removing that tendency. Conversely, the immersion factor is increased, so it is more like the player is actually doing things to some end, rather than asking (telling?) a character to do those same things.
To revisit the book vs game question, I think that there's no good reason why the two should be exclusive. Rather, each should focus on its strengths, and refer to the other when weaker areas are encountered. Rather than having quizzes at the end of textbook chapters, have games included on a CD that requires knowledge (or perhaps even understanding, if the game designer is good enough?) of the chapter content to complete the next level... include a simple checkpoint system so the child is rewarded for getting through each chapter, and encouraged to move on to the next chapter. If chapters have a degree of non-linearity, then this can be reflected in non-linearity of the game plotline.
Something like that would have had me rabid to get through my highschool textbooks, if decently implemented!
There's a project in the works intending to teach (written) Japanese from scratch through a CRPG vehicle.
I hope the owner manages to finish it... After completing the game, people could then progress to those Japanese CRPGs that were never translated!
This perhaps becomes viable because of the volume of other Japanese-language games in the market... I'm not sure how well this general technique might apply to other languages, however.
Seriously though, I hope they port the registry over to a real database system along the way.
That's not a bad idea, but it's also not a great one... I would much rather see the registry system evolved, rather than rehauled.
As I see it, the registry is a repository for system, program and user settings. I would much rather see these settings being stored in special data records linked to the program and/or user in question. Copying a program with all linked settings to another machine will run the program with exactly the same configuration on the other machine... Explicitly linking a program's settings to your user then uninstalling the program could leave the settings in the system, to be reattached if the program is ever re-installed. This last could even be extended so multiple setting profiles could be maintained for a program - swap them in and out as necessary for the task at hand.
Properties on steroids. That would impress me much more than a central registry, no matter how stable or recoverable.
Once Microsoft integrates a relational database into the operating system, worm penetration will reach 100%.
With any luck (forethought?), the filesystem schema will be completely unreachable from any database-exposing APIs that are released - access should be made available exclusively through filesystem APIs... with ACL functions built-in to these APIs, file systems should be no less secure than they are now.
Then again, easier for users to find files does imply easier for virii to find files... Maybe it'll also mean easier for virus checkers to find virii?;-)
...
Unrelated to the above, if relational file systems take off and directory structure becomes less relevant, what kind of notation is going to be used to refer to file location in a single canonical way? How will this affect URIs, and web browsing in particular? Will we still be using slash-separated paths twenty years from now?
If I incorporate my proprietary code into GPLd software, I can still retain copyright to the code and continue to use it in my own projects. Apparently not so with SysV code.
Since the code is now "derived from a GPL work", will your future projects using this code be required to be GPLd? Or as the code author, are you permitted to use the code under different licenses in different projects?
Even if this is the case, what happens five years down the track, when someone notices that your non-GPL project uses identical code (in part) to a GPL work?
This isn't a flame, I'd honestly like to know - as a developer myself, this is the biggest single thing that stops me from even thinking about assisting GPL projects.
If the ISP is receiving 1mbps from the customer, doesn't the multicast tree mean that this will never translate to more than 1mbps down the upstream link? Even if the multicast tree had 100+ leaves up that stream, they would all correspond to a single branch in the tree, so a single stream to the upstream provider.
If any branching occurred at the ISP level, that would correspond to leaves directly below the ISP, so would be going through the ISP's internal (presumably 100/1000 Mb) network.
I guess things become more confusing if the ISP has multiple upstreams, for redundancy/speed.
Slashdot Mirror
... oh wait...
That's ok, it's your own patent
One more note between the X25-E and X25-M... The X25-E is branded as an enterprise-level disk, and is correspondingly much more expensive. It is a SLC design (single layer controller), where the X25-M is a MLC design (multi-layer controller). SLC chips are expected to last considerably longer than MLC, because fewer bits are stored in a single cell in an SLC chip, each cell is expected to be written fewer times when writing or modifying data on the disk. But still, as mentioned in numerous comments here, the X25-M should* last long enough for most consumer uses. *- time may prove us wrong, but I hope not.
Sequential reads are certainly sequential reads, but that might not mean exactly what you think it means. In a spinning disk HDD, a sequential read means move the head to the correct location, and then stream the data as the disk spins under the head. This is roughly as fast as it gets with spinning media, ignoring cache etc. In a SSD, you can actually benefit from parallelisation. SSDs can (if supported by the controller) read simultaneously from different chips, which means either reading the file in a number of parallel sequential reads simultaneously, or being able to satisfy multiple read requests simultaneously, as long as they are stored in different memory chips. So defragmentation in the classical sense may not actually be a desirable goal in the new world order.
But mostly, it's just what they're used to. There's nothing wrong with it, it's roughly the same efficiency either way, so do it the way that makes sense to you, and for the problem you're solving.
And don't forget words that cannot be transcribed without context, such as their/there/they're and which/witch... "I'm going over there, which is useful because I need to talk to whoever anyway" vs "I wouldn't give their witch any credibility".
Does the same reasoning hold for the sentence It isn't Windows? That is, does the word/contraction isn't count as a full-fledged verb?
I'm just curious.
It's taken at least two years for anyone to "find" your computer... being the only assumption you can realistically make from your statement.
Once a script kiddie finds the computer, we find the answer to the question of how long to "tunnel" into it.
Reading this article, I wondered... how much time are these supercomputers around the world actually putting into useful calculations, and how much time into benchmarks to prove how fast they are?
I want to see another benchmark... how many TFlops per year are being spent on benchmarks vs useful applications!
That doesn't make sense... If the board is in a certain state, why does it matter if it reached this state by moving a pawn or by moving a knight, if it is black's turn to move?
Being able to check for a draw isn't necessary in the state tree... it simply means we need a graph with non-directional cycles to represent play. Note that if we enforce that the graph may not contain any directional cycles, then the construction of the tree will never recurse, so each particular board state only needs to be evaluated once.
That doesn't get past the fact that no computer (that I've heard of) has sufficient resources to maintain such a tree, though. That's not to say it's not possible to store a sufficient tree to allow perfect play... I don't know enough to guess either way.
This raises an interesting question. If Microsoft were to encourage developers to register their applications and patches in an accessible repository, so Windows Update could offer to optionally patch any and all software installed on your machine, this would be considered an action intended to increase their monopoly. (Whether or not cost was involved, to the developer or the customer)
If they were to compile their own database of software patches from third party products, this would probably be considered even worse, since they would (probably) only include common applications, so niche software wouldn't get a look-in.
Why then can Red Hat have such a database of patches and it's a good thing? Surely it's doing the same thing?
Ten years before my time? There's a big assumption. Just for the record, I got started on text-only adventure games long before Sierra came into the picture (no pun intended).
There was no reason to hunt pixels in the original Sierra *Quest games... the entire interactive interface was still through typing, I don't think mouse was even supported! You just happened to have a visible avatar that you could maneuver through the traditional adventure 'room' by using the arrow keys. My comment in reply to the grandparent post, was suggesting that this particular addition of pictures to the adventure game was a far from academic exercise, and actually added a lot to the game. At least, IMHO.
I agree with you 99% that today's graphical adventures are mostly an exercise in attempting to find the combination of clicks required to solve a particular 'puzzle'. Then again, some of the puzzles in the old days devolved into 'hunt the exact spelling of the exact verb/noun the game designer intended to be used in this situation', which was arguably more frustrating.
One of my best purchases ever was a second hand LJ 4M Plus for AU$150 (US$80?). Fast printing, great quality, and plugs directly into the network without requiring a host computer, and drivers were already installed in Win2k... The hardest part of installation was setting up the printer's IP and Gateway addresses, since it didn't support DHCP, but now it runs beautifully!
To rephrase the question (disclaimer: Probably violating the spirit of the parent poster), which is more educational, a Harry Potter novel or a Harry Potter computer/console game? The book teaches literacy through example, while the game (hopefully) teaches problem solving and hand-eye manipulation.
A lot of older games used to cover the literacy front as well - adventure games in particular - but today's push for stunning graphics and voiceovers is removing that tendency. Conversely, the immersion factor is increased, so it is more like the player is actually doing things to some end, rather than asking (telling?) a character to do those same things.
To revisit the book vs game question, I think that there's no good reason why the two should be exclusive. Rather, each should focus on its strengths, and refer to the other when weaker areas are encountered. Rather than having quizzes at the end of textbook chapters, have games included on a CD that requires knowledge (or perhaps even understanding, if the game designer is good enough?) of the chapter content to complete the next level... include a simple checkpoint system so the child is rewarded for getting through each chapter, and encouraged to move on to the next chapter. If chapters have a degree of non-linearity, then this can be reflected in non-linearity of the game plotline.
Something like that would have had me rabid to get through my highschool textbooks, if decently implemented!
There's a project in the works intending to teach (written) Japanese from scratch through a CRPG vehicle.
I hope the owner manages to finish it... After completing the game, people could then progress to those Japanese CRPGs that were never translated!
This perhaps becomes viable because of the volume of other Japanese-language games in the market... I'm not sure how well this general technique might apply to other languages, however.
That's not a bad idea, but it's also not a great one... I would much rather see the registry system evolved, rather than rehauled.
As I see it, the registry is a repository for system, program and user settings. I would much rather see these settings being stored in special data records linked to the program and/or user in question. Copying a program with all linked settings to another machine will run the program with exactly the same configuration on the other machine... Explicitly linking a program's settings to your user then uninstalling the program could leave the settings in the system, to be reattached if the program is ever re-installed. This last could even be extended so multiple setting profiles could be maintained for a program - swap them in and out as necessary for the task at hand.
Properties on steroids. That would impress me much more than a central registry, no matter how stable or recoverable.
With any luck (forethought?), the filesystem schema will be completely unreachable from any database-exposing APIs that are released - access should be made available exclusively through filesystem APIs... with ACL functions built-in to these APIs, file systems should be no less secure than they are now.
Then again, easier for users to find files does imply easier for virii to find files... Maybe it'll also mean easier for virus checkers to find virii? ;-)
...
Unrelated to the above, if relational file systems take off and directory structure becomes less relevant, what kind of notation is going to be used to refer to file location in a single canonical way? How will this affect URIs, and web browsing in particular? Will we still be using slash-separated paths twenty years from now?
Maybe if you're a really bad player? :-)
That or a power-gamer (Some might say munchkin ;-)... I know I've spent several days doing something in a game that should only take hours or less.
Since the code is now "derived from a GPL work", will your future projects using this code be required to be GPLd? Or as the code author, are you permitted to use the code under different licenses in different projects?
Even if this is the case, what happens five years down the track, when someone notices that your non-GPL project uses identical code (in part) to a GPL work?
This isn't a flame, I'd honestly like to know - as a developer myself, this is the biggest single thing that stops me from even thinking about assisting GPL projects.
I'm curious about that logic.
If the ISP is receiving 1mbps from the customer, doesn't the multicast tree mean that this will never translate to more than 1mbps down the upstream link? Even if the multicast tree had 100+ leaves up that stream, they would all correspond to a single branch in the tree, so a single stream to the upstream provider.
If any branching occurred at the ISP level, that would correspond to leaves directly below the ISP, so would be going through the ISP's internal (presumably 100/1000 Mb) network.
I guess things become more confusing if the ISP has multiple upstreams, for redundancy/speed.