You haven't said anything about the problem. If you want ease of scaling go with a pure functional language. Functional languages will force you to isolate state issues. Isolate state and you can operate in total parallel. But... generally you have shared objects which are mutable across the users. So you'd end up with very little meaningfully isolated. Those shared mutable objects are what is creating the scaling complexity. The language or technology doesn't solve that, though it can make the solution easier to implement or harder. You're going to need to architect around passing them around and verifying.
But at least it will get you thinking about the problem the right way.
Oh I agree is is better than editing most PostScript by hand. There is a long distance between worst possible case and comfortable. Though I have worked with gorgeous PostScript. Old fashioned hand coded PostScript was pretty awesome sometimes.
There are only two engines, and you would lose the ability to use LyX if you lost the TeX engine. 3 decades ago TeX had something that stripped those issues off dvi2tty. It is easy to drop the page orientation. Mathml may not be needed, TeX can render centered equations using jpgs easily.
Climate change is not just CO2 emissions. All those computer chips, all those plastic bottles need more than energy to make.
That's resource depletion not climate change. Our resource discoveries are outpacing our usage. While there are going to be problems that's not high up on the list.
Do we really need a new smartphone every 2 years, do we really need to buy bottled water?
I hate the word need. But having higher bacterial content in the drinking water and drinking bottled water is less environmentally taxing than a clean water system would be. As for a smartphone every 2 years. We have a lot of breakage and they are wearing out. I'd assume this cycle will slow soonish.
I think pollution is a problem. But I don't think it is nearly as complex a problem or as costly a problem as climate change. I don't think resource depletion is a serious problem.
This is mainly about energy. We could be switching energy production towards renewables. That's going to take years and short term will boost consumption. But everyone knows it is the right long term approach. We don't need to cut standards of living just incentivize the switch.
No. There are 2 TeX engines: TeX proper which goes to dvi and PDFTeX which skips.dvi. LaTeX extensions work against either engine but need to be built against both. The font system is particularly nasty since metafont isn't PDF only.
I wish that TeX would just flush.dvi and tie themselves to the PDF standard. But no, they haven't.
You get the picture. I wouldn't quite go as far as binary. But yes you do throw away the advantages of a human readable editable file. LyX is a good crutch for the learning curve for TeX. But ultimately someone is going to want more WYSIWYG or they are going to want learn TeX.
Slashdot is also remarkably conservative. You see this regularly in terms of computer technology (anti-Wayland, anti-Gnome, anti-Windows 8....) but it is also true in terms of American politics. Climate change is going to require coordinated large scale governmental actions through incentives and regulation. Libertarians don't like it so they pretend there is no underlying problem
This goes beyond typical summer of code but what about adding Framemaker type capabilities to LyX. Right now LyX is a word processor that outputs to TeX. What about an entire document authoring and organizational GUI which used LyX / TeX for the lower level stuff. Pull in DocBook.
There are far more feature rich notebooks systems that include more than just TeX: http://www.sagemath.org/tour.html But if you want primarily TeX plus computer algebra (Maple) it exists but it is commercial. http://www.mackichan.com/
I would then be happy to use LyX as an entry point for WysiWyg tweaking, and finallly jump into emacs to really finalize my document...
Understand when you are talking LyX you are talking computer generated code not human generated. It can be very complex to modify. You can do minor edits to do things LyX doesn't support (though LyX allows you to do TeX directory) but you aren't going to want to work on a LyX document in a normal text editor as a human for much.
That would seem almost trivial from a coding standpoint. I think the real thing is infrastructure. If you had servers on which the system was authenticated LyX could save files remotely or move then via. ssh. LyX can use an arbitrary TeX including an ssh connect. TeX can accept input streams....
CLSI sounds cool. I didn't know about it. But IMHO the real issue is who is willing to provide the service.
As I tried to point out above, Raster's pager in the Enlightenment window manager version 0.16 (or earlier) in 1995-96
Oh I see you aren't talking about an actual GUI pager but something for Englightenment. Well I don't know much about Enlightenment but what I do know is they have tied themselves to Tizen continuing the Framebuffer approach of MeeGo. It has gotten Enlightenment specific but Enlightenment in general isn't running against raw X11 but rather against Evas which handles many of the primitives that X11 normally would. X11 is a supported subsystem.
So while I have no idea what you are talking about. I do know the Enlightenment guys don't agree with your rosy assessment of how well Enlightenment worked with X11 which is why they are building entirely new layers in EFL to avoid X11.
So who exactly is lying to the kiddies here?
I've got it on my Nokia N900, it's on Kindles - there may be better for such situations but it has certainly delivered so that's yet another example of you not knowing what you are talking about.
Maemo was not a shinning success. And why wouldn't Kindle use X11? That's a text base operating system. If you mean in graphical mode then no, it is using Android.
I think it is a pretty easy question. Whose pager. Who made this supposed pager?
And the arguments over Wayland have always been over future directions. As for X11 sucking. If X11 was good there wouldn't be a Wayland project in the first place. There is nothing childish about it. X11 failed at delivering a system for Linux comparable to what exists on Windows and Mac. X11 failed at delivering a system for phones. At the thing everyone claims X11 is designed to do really well, remote applications it also fails in practice.
A lot of people thought the flying car would be easier than it was. In general, the lack of improvement in cars since pre WWII has been impressive. Things got better, like they did for bicycles. As for leisure society I'd say sci-fi was rather split on that. As for money, yeah I'll agree sci-fi rarely predicted that Americans would vote for the economic system of the 20s.
What Google and Verizon have said is that their costs are high due to regulation and hurdles. It is to complex to navigate the CA agencies. There is no one they can just "do business with" but rather dozens of agencies all of which have to be passed through. What's wrong with California is you don't have political machines in CA.
As for a named account... I think that Slashdot's moderation system is a terrible failure. Among other problems, it's destructive to good discussions, absurdly easy for bad-faith users to game, and lends itself to groupthink.
As an aside I disagree. I think there is a lot of group think on it and a lot of silly moderation. But a system doesn't need to be perfect individually. In general good contributors tend to end up at 2 while bad ones (i.e. trolling, no information...) at -1. It accomplishes when it needs to in terms of moving discussions up.
I question that assumption. What's magic about 2.5x?
Nothing is magic. But
1) Disk space is much cheaper than RAM. You want RAM to be efficiently used. For most OSes (OSX after 10.7 with state saving may be an exception but this applies to Linux) this means loading software into RAM, letting it go through configuration and then swapping it out. I.E. the swap space should be swap > RAM. So something like swap = 2*RAM is a good min.
2) RAM is much faster than disk. The thing that needs to be avoided is where a large fraction of the swap can't be stored in RAM because this can result in fragging which can slow the computer down by 2 orders of magnitude. About 1/2 the loaded programs are running at any time at max, most of them have some data they aren't using but much above swap = 3*RAM starts to not get much of a speed bump and danger goes up.
So swap = 2RAM - 3RAM i.e. 2.5 is a good middle point.
The Mach pager process dynamically creates more swapfiles as the system demands more swap. Boot an OS X system up and you'll find it has a single 64MB swapfile (look in/private/var/vm for them). Push it hard enough to use all that swap up and it'll create another, push it further and it'll start ratcheting the size of the new files up in powers of 2. When need dies down, it can compact swap contents and dynamically delete some of the files. (Actually succeeding at deleting unneeded swapfiles is a new thing in OS X, always used to hear NeXT graybeards talk about the bad old days of being stuck with large swapfiles until a reboot cleaned them out.)
That's what Oracle does by default as well. We have details of this sort of process on from Oracle which uses both the Linux and the OSX strategy. For most use cases the OSX strategy is better (on Oracle). OK I'll grant you the argument if OSX is using gradually increasing file allocations that's enough of an advantage to beat static swap.
. You might manage to run yourself out of disk space, then find there's something you can't do because the system can't expand swap, but you'd have run into problems long before that with a dedicated swap partition...
You want to cap the swap in size (generally) long before you run out of disk. You want to avoid fragging.
Google seems remarkably successful in cracking the small business advertising nut. Arguably that is Google's core business. They have been successful. But the thing is most of the time when people are searching for a product on the web they are ready to buy and many times when people are buying they start by searching.
Foursquare would need to have something like local groupon type deals. Something like "I want to buy gas where is a good place" "I want a good meal" with some advantage for the customer in following their advice something like groupon.
Like I said, Stockholm Syndrome. You think it's a good thing that tools like fdisk let you get all fiddly, because you must get fiddly to make a triple-boot USB stick. But there would be literally no need to get fiddly if non-OSX operating systems could get with the times in abandoning the World's Stupidest Partition Table Format, the one literally based on 1980 Seagate ST-506 MFM controller registers.
Great post. You should get a named account. And I agree. But I don't get to run the world. Other OSes using dumb partitioning systems I can't change. Linux's FDISK handling those dumb systems well is something I can take advantage of.
Most of the time it just wastes space, and if you guess wrong about how much swap you need, you're fucked.
I assume you generally want swap at 2.5x ram so unless you are changing your RAM...
Slashdot Mirror
You haven't said anything about the problem. If you want ease of scaling go with a pure functional language. Functional languages will force you to isolate state issues. Isolate state and you can operate in total parallel. But... generally you have shared objects which are mutable across the users. So you'd end up with very little meaningfully isolated. Those shared mutable objects are what is creating the scaling complexity. The language or technology doesn't solve that, though it can make the solution easier to implement or harder. You're going to need to architect around passing them around and verifying.
But at least it will get you thinking about the problem the right way.
Oh I agree is is better than editing most PostScript by hand. There is a long distance between worst possible case and comfortable. Though I have worked with gorgeous PostScript. Old fashioned hand coded PostScript was pretty awesome sometimes.
There are only two engines, and you would lose the ability to use LyX if you lost the TeX engine. 3 decades ago TeX had something that stripped those issues off dvi2tty. It is easy to drop the page orientation. Mathml may not be needed, TeX can render centered equations using jpgs easily.
Climate change is not just CO2 emissions. All those computer chips, all those plastic bottles need more than energy to make.
That's resource depletion not climate change. Our resource discoveries are outpacing our usage. While there are going to be problems that's not high up on the list.
Do we really need a new smartphone every 2 years, do we really need to buy bottled water?
I hate the word need. But having higher bacterial content in the drinking water and drinking bottled water is less environmentally taxing than a clean water system would be. As for a smartphone every 2 years. We have a lot of breakage and they are wearing out. I'd assume this cycle will slow soonish.
I think pollution is a problem. But I don't think it is nearly as complex a problem or as costly a problem as climate change.
I don't think resource depletion is a serious problem.
This is mainly about energy. We could be switching energy production towards renewables. That's going to take years and short term will boost consumption. But everyone knows it is the right long term approach. We don't need to cut standards of living just incentivize the switch.
No. There are 2 TeX engines: TeX proper which goes to dvi and PDFTeX which skips .dvi. LaTeX extensions work against either engine but need to be built against both. The font system is particularly nasty since metafont isn't PDF only.
I wish that TeX would just flush .dvi and tie themselves to the PDF standard. But no, they haven't.
Anti-firearm tends to correlate with urban more than conservative.
You get the picture. I wouldn't quite go as far as binary. But yes you do throw away the advantages of a human readable editable file. LyX is a good crutch for the learning curve for TeX. But ultimately someone is going to want more WYSIWYG or they are going to want learn TeX.
Slashdot is also remarkably conservative. You see this regularly in terms of computer technology (anti-Wayland, anti-Gnome, anti-Windows 8....) but it is also true in terms of American politics. Climate change is going to require coordinated large scale governmental actions through incentives and regulation. Libertarians don't like it so they pretend there is no underlying problem
That's a good line. I like it.
This goes beyond typical summer of code but what about adding Framemaker type capabilities to LyX. Right now LyX is a word processor that outputs to TeX. What about an entire document authoring and organizational GUI which used LyX / TeX for the lower level stuff. Pull in DocBook.
There are far more feature rich notebooks systems that include more than just TeX: http://www.sagemath.org/tour.html
But if you want primarily TeX plus computer algebra (Maple) it exists but it is commercial. http://www.mackichan.com/
What you want is .dvi to EPUB. That's a TeX/LaTeX feature not a LyX feature.
I would then be happy to use LyX as an entry point for WysiWyg tweaking, and finallly jump into emacs to really finalize my document...
Understand when you are talking LyX you are talking computer generated code not human generated. It can be very complex to modify. You can do minor edits to do things LyX doesn't support (though LyX allows you to do TeX directory) but you aren't going to want to work on a LyX document in a normal text editor as a human for much.
That would seem almost trivial from a coding standpoint. I think the real thing is infrastructure. If you had servers on which the system was authenticated LyX could save files remotely or move then via. ssh. LyX can use an arbitrary TeX including an ssh connect. TeX can accept input streams....
CLSI sounds cool. I didn't know about it. But IMHO the real issue is who is willing to provide the service.
As I tried to point out above, Raster's pager in the Enlightenment window manager version 0.16 (or earlier) in 1995-96
Oh I see you aren't talking about an actual GUI pager but something for Englightenment. Well I don't know much about Enlightenment but what I do know is they have tied themselves to Tizen continuing the Framebuffer approach of MeeGo. It has gotten Enlightenment specific but Enlightenment in general isn't running against raw X11 but rather against Evas which handles many of the primitives that X11 normally would. X11 is a supported subsystem.
So while I have no idea what you are talking about. I do know the Enlightenment guys don't agree with your rosy assessment of how well Enlightenment worked with X11 which is why they are building entirely new layers in EFL to avoid X11.
So who exactly is lying to the kiddies here?
I've got it on my Nokia N900, it's on Kindles - there may be better for such situations but it has certainly delivered so that's yet another example of you not knowing what you are talking about.
Maemo was not a shinning success. And why wouldn't Kindle use X11? That's a text base operating system. If you mean in graphical mode then no, it is using Android.
Exactly you are running X11 against a framebuffer. Which means X11 is writing to memory and someone else is handling putting that on the screen.
I agree. But around NY we have rather good residential fiber. I suspect the problem with NY is cost. I don't know Chicago well enough.
America is a mess when it comes to infrastructure. California, because of Silicon Valley, gets discussed.
I think it is a pretty easy question. Whose pager. Who made this supposed pager?
And the arguments over Wayland have always been over future directions. As for X11 sucking. If X11 was good there wouldn't be a Wayland project in the first place. There is nothing childish about it. X11 failed at delivering a system for Linux comparable to what exists on Windows and Mac. X11 failed at delivering a system for phones. At the thing everyone claims X11 is designed to do really well, remote applications it also fails in practice.
A lot of people thought the flying car would be easier than it was. In general, the lack of improvement in cars since pre WWII has been impressive. Things got better, like they did for bicycles. As for leisure society I'd say sci-fi was rather split on that. As for money, yeah I'll agree sci-fi rarely predicted that Americans would vote for the economic system of the 20s.
What Google and Verizon have said is that their costs are high due to regulation and hurdles. It is to complex to navigate the CA agencies. There is no one they can just "do business with" but rather dozens of agencies all of which have to be passed through. What's wrong with California is you don't have political machines in CA.
This is not the 21st century I was told to expect.
Most scifi is rather depressing. I would have figured better than the 21st century you were told to expect by most.
As for a named account... I think that Slashdot's moderation system is a terrible failure. Among other problems, it's destructive to good discussions, absurdly easy for bad-faith users to game, and lends itself to groupthink.
As an aside I disagree. I think there is a lot of group think on it and a lot of silly moderation. But a system doesn't need to be perfect individually. In general good contributors tend to end up at 2 while bad ones (i.e. trolling, no information...) at -1. It accomplishes when it needs to in terms of moving discussions up.
I question that assumption. What's magic about 2.5x?
Nothing is magic. But
1) Disk space is much cheaper than RAM. You want RAM to be efficiently used. For most OSes (OSX after 10.7 with state saving may be an exception but this applies to Linux) this means loading software into RAM, letting it go through configuration and then swapping it out. I.E. the swap space should be
swap > RAM. So something like swap = 2*RAM is a good min.
2) RAM is much faster than disk. The thing that needs to be avoided is where a large fraction of the swap can't be stored in RAM because this can result in fragging which can slow the computer down by 2 orders of magnitude. About 1/2 the loaded programs are running at any time at max, most of them have some data they aren't using but much above swap = 3*RAM starts to not get much of a speed bump and danger goes up.
So swap = 2RAM - 3RAM i.e. 2.5 is a good middle point.
The Mach pager process dynamically creates more swapfiles as the system demands more swap. Boot an OS X system up and you'll find it has a single 64MB swapfile (look in /private/var/vm for them). Push it hard enough to use all that swap up and it'll create another, push it further and it'll start ratcheting the size of the new files up in powers of 2. When need dies down, it can compact swap contents and dynamically delete some of the files. (Actually succeeding at deleting unneeded swapfiles is a new thing in OS X, always used to hear NeXT graybeards talk about the bad old days of being stuck with large swapfiles until a reboot cleaned them out.)
That's what Oracle does by default as well. We have details of this sort of process on from Oracle which uses both the Linux and the OSX strategy. For most use cases the OSX strategy is better (on Oracle). OK I'll grant you the argument if OSX is using gradually increasing file allocations that's enough of an advantage to beat static swap.
. You might manage to run yourself out of disk space, then find there's something you can't do because the system can't expand swap, but you'd have run into problems long before that with a dedicated swap partition...
You want to cap the swap in size (generally) long before you run out of disk. You want to avoid fragging.
Google seems remarkably successful in cracking the small business advertising nut. Arguably that is Google's core business. They have been successful. But the thing is most of the time when people are searching for a product on the web they are ready to buy and many times when people are buying they start by searching.
Foursquare would need to have something like local groupon type deals. Something like "I want to buy gas where is a good place" "I want a good meal" with some advantage for the customer in following their advice something like groupon.
Like I said, Stockholm Syndrome. You think it's a good thing that tools like fdisk let you get all fiddly, because you must get fiddly to make a triple-boot USB stick. But there would be literally no need to get fiddly if non-OSX operating systems could get with the times in abandoning the World's Stupidest Partition Table Format, the one literally based on 1980 Seagate ST-506 MFM controller registers.
Great post. You should get a named account. And I agree. But I don't get to run the world. Other OSes using dumb partitioning systems I can't change. Linux's FDISK handling those dumb systems well is something I can take advantage of.
Most of the time it just wastes space, and if you guess wrong about how much swap you need, you're fucked.
I assume you generally want swap at 2.5x ram so unless you are changing your RAM...