I think the main appeal of desktop/workstation 10K RPM drives over 7200RPM drives was for realtime audio and video capture... the reduced seek times really help when you need to maintain constant write throughput Or Else (tm). But the folks doing that have moved onto SSDs, because disks for that purpose are basically expendable and SSDs cost less than a good 10K drive did only 3-4 years ago. And have tons of advantages for real time use (zero seek latency! No more need to tweak things and really hope the system and disk cooperate to give you full linear bandwidth with no I/O stuttering... I'm having flashbacks to 2003).
For bulk storage, there just isn't much of an advantage to justify the per gigabyte cost increase and higher failure rates. You may as well spend your money on more RAM (having a 14G disk cache is pretty nice).
C++ is an unparseable mish-mash of incoherent and unsafe components. More complex than Common Lisp with none of the benefits... AFAICT half the language exists just to support obsolete memory management strategies. And the other half exists to compensate for having an extremely unexpressive type system grounded in glorified bit fields of varying length./p
This is what I use, since my roommate has Amazon Prime, and it is great. Also, proof that the DRM on streaming services cat is out of the bag... RTMPE was known broken for years before Amazon Prime and Hulu appeared, and yet they still use it. And tons of people circumvent is regularly, and they appear to not care. Making Netflix's licensing execuses just that. Hollywood execs seem like they're pretty easy to hoodwink.
A side question... have you ever gotten rented content to work with xbmc? I'm not a fan of "buying" video from Amazon (since... I am not paying nearly full dvd price, sometimes more, unless I get a DRM free download, in which case I'd buy Jeff Bezos another yacht) but renting is fine by me... unfortunately the one time we tried it showed up in the watchlist but refused to play (real bummer, since I like to watch movies on a whim, and rarely care to ever see one again enough to justify the dvd... it's so close to being easier than piracy and worth it, except for the not playing part).
Worse than that, an impovershed and failing alliance of UHF syndication channels, that loved Star Trek more than him despite Babylon 5 clearly being superior. And then an impovershed basic cable channel that destroyed everything:(
I think you missed the part where there's no graphics acceleration and it sucks.
Also, I really just want a DRM-free RTMP stream that I can use with xbmc. Or, hell, throw some trivial drm on there that you know is broken and just look the other way like Amazon and Hulu. Until then... I guess Netflix doesn't get my money, and I will be sad that there is no legal way for me to watch JMS return to television.
Hopefully having been burned by the collapse of PTEN and then getting bent over by TNT, JMS negotiated a contract that will avoid immediate cancellation... because I want some good scifi that wasn't made 15+ years ago, and there's only one man left who can do it...
They need to use external tools to build the metaobjects required for their metaobject protocol to work because C++ is insufficiently expressive and does not provide a meaningful way to reconstruct type information at run time in some false quest for efficiency.
I bet they wouldn't be using C++ were they writing Qt from scratch in 2013 and not 1995.
Federal DOT guidelines are to give motorists a 5-10mph leeway for ticketing above 35mph or something (too lazy to look it up, I think it's +5mph below 50, then +10mph) so it's perfectly sensible that the road was designed to be safely used with typical driving habits in mind.
Have you tried setting the dirty regions option? For whatever reason the default re-renders the entire screen 60 times per second... but you can flip it to only updating when regions are damaged. Once I flipped that on I was able to use xbmc alright on my ancient athlon.
I don't think so. Right now, sure that's the alternative, but it needn't be that way.
See Emacs. You start off with tool bars and menus... maybe interacting with the graphical customize system to tweak a few things. But then you hit your first "I wish it...", something small. Next thing you know, the interface is peeled away; you check the help page for a similar command, see that you can visit the source there, realize the source is sensible and short, copy and paste and modify one or two things, and... congrats, you're a programmer and you didn't even know it. Or maybe you discover keyboard macros and do things that way, peeking at your.emacs realiziing that all they do is generate a short function.
The problem is that "configuration" exists... there is this utterly artificial distinction between configuration and code, the user and developer, because of languages like C where the skill jump from using to writing is obscene (how many of us had to suffer CS201/202, seeing the students who hadn't taught themselves to code in high school struggling for an entire year just with the syntax and semantics of C! And even then, not being able to reason about them despite having written thousands of lines of it and being forcibly exposed three times a week in labs...). Languages like Python bring us closer to closing that gap, but interfaces are still static and difficult (maybe not Firefox, but xulrunner is basically the Emacs of the web browser world when you think about it). Why can't I just ask to inspect whatever I click at, and alter my interface at run time? We let people add and remove tool bar buttons (insert gnome 3 joke here)... why not everything?
You can already see a bit of this kind of extension in Firefox and Gnome-Shell and KDE's Plasma, but really, the desktop folks should think long and hard about why Emacs is so discoverable (if your starting point is that you are a three armed alien of course, but aren't half of us?), and integrate those principles. We're getting there, perhaps in another few years if disinterest in the desktop projects doesn't cause them to implode:(
My Grandmother's hard drive died a few years ago and... I got a cheap 1TB drive to replace it (nothing cheaper available really at the time) and of course the XP restore disk decided she was a filthy pirate and that was that.
After installing Debian and KDE for her, my support requests have become nil, except when any of my younger cousins visit and why does this flash game not work... because Adobe hates your freedom children, that's why. But my Grandmother just readers her email and plays KPatience and it Just Works (tm).
The innards are identical (ok, except for the mcu since I guess they had to swap that out to get usb support). It's just less heavy because they reduced the size of the case and removed some excess material.
The only dodgy unicomp thing is the one with the built in pointing stick. It's a first or second gen trackpoint, or a clone, and really not very useful. And the mouse buttons are simple contact foil things, so they wear out after 2-3 years and you can't replace them without taking the keyboard completely apart, which is approximately impossible. I'd kill for one of them with a real trackpoint IV and proper switches for the mouse buttons...
I think in a laptop, mechanical failure is far more likely than flash failing. Sure the drives are designed to handle being started/stopped all the time, but life for a laptop hard drive is still not a very gentle one. I can't imagine a use case for these hybrid drives outside of laptops.
It's also likely that the SLC cache at least would outlast the mechanical drive components under even ideal conditions anyway...
I think we're seeing a natural cycle in the software world. During the 80s there were dozens of architectures, operating systems, languages, etc. and the best (for some definition of best) became dominate and during the 90s consolidated. Now we're in the midst of another explosion in new technology (languages, display servers, processor architectures, perhaps even operating systems) that will eventually lead to reconciliation and consolidation in another five to ten years.
Things like Wayland have to appear, and even fail: their existence allows new ideas to be tested giving us a better idea of where to go from here.
I have an 8-core FX-8320 rig, a nice 16G of RAM, a radeon 6870 (some XFX thing with two fans that run slowly... it's quiet instead of lawnmower loud like I hear they normally are), and a serious amount of storage (a small expendable SSD for the rootfs, 2TB RAID1 for bulk storage, expanding with a 3TB RAID1 soon so that I can finally back my laptop up to it, and it has room for four more drives so I want to have hotspares and whatnot...). And it is quieter than the furnace fan.
Although, it does get to be about as loud as the fridge if someone's playing mupen64plus while I'm doing make -j16 on something (in practice, the fans are at faster than completely idle for about 20 minutes a week). It's still quieter than my old dual athlon MP rig, and that thing was pretty quiet for a machine built in 2002.
You can pick up all the Scheme syntax you need to know in a few minutes: "" delimits strings, () delimits lists, you separate words with spaces, and everything else is a generic atom. Then you only need to learn a few operators (if, quote, define, display, the basic arithmetic operators) and you are off writing the sorts of programs a beginning CS student writes. None of this undecidable grammar making level one programming a hair-pulling nightmare of useless syntax errors nonsense.
I think there's a lot of interest in microkernels right now in the mobile world... single chip, running something like OKL4 or Genode with GNU/Linux as a server and the radio stack as another server, completely isolated. It's easier to trust a small microkernel won't allow resources to leak between servers than it is to trust something huge like Xen.
First of all Netflix, Hulu, Amazon, iTunes, etc all suck. I definitely won't be switching to any of them. They don't work across platforms and are infected with digital restriction crap (for instance I don't have access to ANY of them) with the Linux distributions I run. They aren't remotely standards complaint. There depending on proprietary crap like flash and silverlight. Both aren't widely supported anymore (if they ever were) and what support exists is disappearing fast. iPads, many Android tablets, and other devices don't support either format not to mention other devices on the market. Firefox for instance isn't getting updates beyond security. I don't use chrome either. Not to watch movies/tv shows online anyway.
Actually, at least Amazon Prime's "free" content (my roomate has a Prime account, I have XBMC, and we share a living room...) and Hulu are just using RTMPE... utterly broken, and it's pretty great. There are easily available XBMC plugins (bluecop repository) that integrate reasonably, and the experience is at least better than cable. Which sort of makes me wonder (given that DVDs have effectively been DRM free since ever and bluray is easily broken by people who really care) why the video industry even bothers with DRM. I'm kind of bummed that I can't use stuff like Netflix, or actually buy tv series and whatnot on Amazon (buying the permission to stream DRM encumbered crap from a third party isn't exactly buying if you ask me... just let me download the files, I don't upload my music to the pirate bay trust me I won't upload the movie either guys).
I hate being treated like a bad person just for wanting content that doesn't look horrible on a 50" screen without atrocious DRM (bluray's whole thing where new discs can prevent you from reading old discs or anything at all at the hardware level is just plain evil, and they wonder why the optical media industry is dying).
We'd all like to think that the data model is 100% clear at the beginning of a project and comes from the heavens in perfectly formed 3NF.
Reality bites. Needs change... the domain expands, human judgement is limited (there are multiple ways to model the same data, at least once you have tuples in your relational language... and we're not perfect), etc.
I mean,... you really only need triples so why bother with supporting operations on tuples?
Guile has fluids, which give you similar behavior as dynamic scoping without forcing the costs of dynamic scoping on all code. Emacs Lisp actually has optional lexical scope now, and a lot of stuff is converting to use it because it's proven to reduce code errors and runtime overhead. That, and you can't use dynamically scoped bindings to pass information between threads (a serious limitation in Emacs's implementation).
Obviously, fluids have their place in something like Emacs where the cost of breaking abstraction is lower than the incredible pain in the ass it'd be to not have them for customizing things.
The main impediment to adoption I think was barely missing the Debian Squeeze and the last set of distro releases... but it made it into Wheezy and I think is in every other major distro's latest/upcoming release.
Common Lisp only had a few years of stagnation... It more suffered from being a decade ahead of its time so folks took "you can still run code from 1995 in the aughts" as a sign of decay instead of stability. Of course, it never really did regain the GUI capabilities (except for Clozure on OS X) the proprietary stuff had since the whole compiler/tools industry just collapsed suddenly.
I've used CL professionally, and I'd put it on par with C++ for that... powerful, well supported, without the annoyances of Java, but fundamentally flawed... at least the development environment (Emacs + SLIME) is better than anything other languages provide (sadly, that's what keeps me coming back to it for more abuse).
It's close to being reality. Guile has an Emacs-Lisp compiler to its VM that can run actual elisp programs, but lacks... the emacs part. And last summer's GSoC (perhaps this summer, finishing it?) saw emacs's lisp interpreter ported to guile... as in, the C representations of Emacs's internal data types and control structures are done using libguile. The code is currently being rebased on the latest emacs trunk; hopefully it'll see public release sooner than later.
Now the two pieces just have to meet in the middle.
So that's the first 95%. Now just for the other 95%!
Slashdot Mirror
I think the main appeal of desktop/workstation 10K RPM drives over 7200RPM drives was for realtime audio and video capture... the reduced seek times really help when you need to maintain constant write throughput Or Else (tm). But the folks doing that have moved onto SSDs, because disks for that purpose are basically expendable and SSDs cost less than a good 10K drive did only 3-4 years ago. And have tons of advantages for real time use (zero seek latency! No more need to tweak things and really hope the system and disk cooperate to give you full linear bandwidth with no I/O stuttering... I'm having flashbacks to 2003).
For bulk storage, there just isn't much of an advantage to justify the per gigabyte cost increase and higher failure rates. You may as well spend your money on more RAM (having a 14G disk cache is pretty nice).
It has the one thing that matters: a real type system.
C++ is an unparseable mish-mash of incoherent and unsafe components. More complex than Common Lisp with none of the benefits... AFAICT half the language exists just to support obsolete memory management strategies. And the other half exists to compensate for having an extremely unexpressive type system grounded in glorified bit fields of varying length./p
If they want my money they are.
This is what I use, since my roommate has Amazon Prime, and it is great. Also, proof that the DRM on streaming services cat is out of the bag... RTMPE was known broken for years before Amazon Prime and Hulu appeared, and yet they still use it. And tons of people circumvent is regularly, and they appear to not care. Making Netflix's licensing execuses just that. Hollywood execs seem like they're pretty easy to hoodwink.
A side question... have you ever gotten rented content to work with xbmc? I'm not a fan of "buying" video from Amazon (since ... I am not paying nearly full dvd price, sometimes more, unless I get a DRM free download, in which case I'd buy Jeff Bezos another yacht) but renting is fine by me... unfortunately the one time we tried it showed up in the watchlist but refused to play (real bummer, since I like to watch movies on a whim, and rarely care to ever see one again enough to justify the dvd... it's so close to being easier than piracy and worth it, except for the not playing part).
Worse than that, an impovershed and failing alliance of UHF syndication channels, that loved Star Trek more than him despite Babylon 5 clearly being superior. And then an impovershed basic cable channel that destroyed everything :(
We can all have dreams.
I think you missed the part where there's no graphics acceleration and it sucks.
Also, I really just want a DRM-free RTMP stream that I can use with xbmc. Or, hell, throw some trivial drm on there that you know is broken and just look the other way like Amazon and Hulu. Until then... I guess Netflix doesn't get my money, and I will be sad that there is no legal way for me to watch JMS return to television.
Hopefully having been burned by the collapse of PTEN and then getting bent over by TNT, JMS negotiated a contract that will avoid immediate cancellation... because I want some good scifi that wasn't made 15+ years ago, and there's only one man left who can do it...
They need to use external tools to build the metaobjects required for their metaobject protocol to work because C++ is insufficiently expressive and does not provide a meaningful way to reconstruct type information at run time in some false quest for efficiency.
I bet they wouldn't be using C++ were they writing Qt from scratch in 2013 and not 1995.
Federal DOT guidelines are to give motorists a 5-10mph leeway for ticketing above 35mph or something (too lazy to look it up, I think it's +5mph below 50, then +10mph) so it's perfectly sensible that the road was designed to be safely used with typical driving habits in mind.
Have you tried setting the dirty regions option? For whatever reason the default re-renders the entire screen 60 times per second... but you can flip it to only updating when regions are damaged. Once I flipped that on I was able to use xbmc alright on my ancient athlon.
I don't think so. Right now, sure that's the alternative, but it needn't be that way.
See Emacs. You start off with tool bars and menus... maybe interacting with the graphical customize system to tweak a few things. But then you hit your first "I wish it...", something small. Next thing you know, the interface is peeled away; you check the help page for a similar command, see that you can visit the source there, realize the source is sensible and short, copy and paste and modify one or two things, and ... congrats, you're a programmer and you didn't even know it. Or maybe you discover keyboard macros and do things that way, peeking at your .emacs realiziing that all they do is generate a short function.
The problem is that "configuration" exists... there is this utterly artificial distinction between configuration and code, the user and developer, because of languages like C where the skill jump from using to writing is obscene (how many of us had to suffer CS201/202, seeing the students who hadn't taught themselves to code in high school struggling for an entire year just with the syntax and semantics of C! And even then, not being able to reason about them despite having written thousands of lines of it and being forcibly exposed three times a week in labs...). Languages like Python bring us closer to closing that gap, but interfaces are still static and difficult (maybe not Firefox, but xulrunner is basically the Emacs of the web browser world when you think about it). Why can't I just ask to inspect whatever I click at, and alter my interface at run time? We let people add and remove tool bar buttons (insert gnome 3 joke here)... why not everything?
You can already see a bit of this kind of extension in Firefox and Gnome-Shell and KDE's Plasma, but really, the desktop folks should think long and hard about why Emacs is so discoverable (if your starting point is that you are a three armed alien of course, but aren't half of us?), and integrate those principles. We're getting there, perhaps in another few years if disinterest in the desktop projects doesn't cause them to implode :(
My Grandmother's hard drive died a few years ago and ... I got a cheap 1TB drive to replace it (nothing cheaper available really at the time) and of course the XP restore disk decided she was a filthy pirate and that was that.
After installing Debian and KDE for her, my support requests have become nil, except when any of my younger cousins visit and why does this flash game not work... because Adobe hates your freedom children, that's why. But my Grandmother just readers her email and plays KPatience and it Just Works (tm).
The innards are identical (ok, except for the mcu since I guess they had to swap that out to get usb support). It's just less heavy because they reduced the size of the case and removed some excess material.
The only dodgy unicomp thing is the one with the built in pointing stick. It's a first or second gen trackpoint, or a clone, and really not very useful. And the mouse buttons are simple contact foil things, so they wear out after 2-3 years and you can't replace them without taking the keyboard completely apart, which is approximately impossible. I'd kill for one of them with a real trackpoint IV and proper switches for the mouse buttons...
I think in a laptop, mechanical failure is far more likely than flash failing. Sure the drives are designed to handle being started/stopped all the time, but life for a laptop hard drive is still not a very gentle one. I can't imagine a use case for these hybrid drives outside of laptops.
It's also likely that the SLC cache at least would outlast the mechanical drive components under even ideal conditions anyway...
I think we're seeing a natural cycle in the software world. During the 80s there were dozens of architectures, operating systems, languages, etc. and the best (for some definition of best) became dominate and during the 90s consolidated. Now we're in the midst of another explosion in new technology (languages, display servers, processor architectures, perhaps even operating systems) that will eventually lead to reconciliation and consolidation in another five to ten years.
Things like Wayland have to appear, and even fail: their existence allows new ideas to be tested giving us a better idea of where to go from here.
Awesome.
I have an 8-core FX-8320 rig, a nice 16G of RAM, a radeon 6870 (some XFX thing with two fans that run slowly... it's quiet instead of lawnmower loud like I hear they normally are), and a serious amount of storage (a small expendable SSD for the rootfs, 2TB RAID1 for bulk storage, expanding with a 3TB RAID1 soon so that I can finally back my laptop up to it, and it has room for four more drives so I want to have hotspares and whatnot...). And it is quieter than the furnace fan.
Although, it does get to be about as loud as the fridge if someone's playing mupen64plus while I'm doing make -j16 on something (in practice, the fans are at faster than completely idle for about 20 minutes a week). It's still quieter than my old dual athlon MP rig, and that thing was pretty quiet for a machine built in 2002.
And it only set me back about a grand.
You can pick up all the Scheme syntax you need to know in a few minutes: "" delimits strings, () delimits lists, you separate words with spaces, and everything else is a generic atom. Then you only need to learn a few operators (if, quote, define, display, the basic arithmetic operators) and you are off writing the sorts of programs a beginning CS student writes. None of this undecidable grammar making level one programming a hair-pulling nightmare of useless syntax errors nonsense.
I think there's a lot of interest in microkernels right now in the mobile world... single chip, running something like OKL4 or Genode with GNU/Linux as a server and the radio stack as another server, completely isolated. It's easier to trust a small microkernel won't allow resources to leak between servers than it is to trust something huge like Xen.
Seriously?
First of all Netflix, Hulu, Amazon, iTunes, etc all suck. I definitely won't be switching to any of them. They don't work across platforms and are infected with digital restriction crap (for instance I don't have access to ANY of them) with the Linux distributions I run. They aren't remotely standards complaint. There depending on proprietary crap like flash and silverlight. Both aren't widely supported anymore (if they ever were) and what support exists is disappearing fast. iPads, many Android tablets, and other devices don't support either format not to mention other devices on the market. Firefox for instance isn't getting updates beyond security. I don't use chrome either. Not to watch movies/tv shows online anyway.
Actually, at least Amazon Prime's "free" content (my roomate has a Prime account, I have XBMC, and we share a living room...) and Hulu are just using RTMPE... utterly broken, and it's pretty great. There are easily available XBMC plugins (bluecop repository) that integrate reasonably, and the experience is at least better than cable. Which sort of makes me wonder (given that DVDs have effectively been DRM free since ever and bluray is easily broken by people who really care) why the video industry even bothers with DRM. I'm kind of bummed that I can't use stuff like Netflix, or actually buy tv series and whatnot on Amazon (buying the permission to stream DRM encumbered crap from a third party isn't exactly buying if you ask me... just let me download the files, I don't upload my music to the pirate bay trust me I won't upload the movie either guys).
I hate being treated like a bad person just for wanting content that doesn't look horrible on a 50" screen without atrocious DRM (bluray's whole thing where new discs can prevent you from reading old discs or anything at all at the hardware level is just plain evil, and they wonder why the optical media industry is dying).
We'd all like to think that the data model is 100% clear at the beginning of a project and comes from the heavens in perfectly formed 3NF.
Reality bites. Needs change... the domain expands, human judgement is limited (there are multiple ways to model the same data, at least once you have tuples in your relational language... and we're not perfect), etc.
I mean, ... you really only need triples so why bother with supporting operations on tuples?
Guile has fluids, which give you similar behavior as dynamic scoping without forcing the costs of dynamic scoping on all code. Emacs Lisp actually has optional lexical scope now, and a lot of stuff is converting to use it because it's proven to reduce code errors and runtime overhead. That, and you can't use dynamically scoped bindings to pass information between threads (a serious limitation in Emacs's implementation).
Obviously, fluids have their place in something like Emacs where the cost of breaking abstraction is lower than the incredible pain in the ass it'd be to not have them for customizing things.
The main impediment to adoption I think was barely missing the Debian Squeeze and the last set of distro releases... but it made it into Wheezy and I think is in every other major distro's latest/upcoming release.
The perils of distro release cycles...
Common Lisp only had a few years of stagnation... It more suffered from being a decade ahead of its time so folks took "you can still run code from 1995 in the aughts" as a sign of decay instead of stability. Of course, it never really did regain the GUI capabilities (except for Clozure on OS X) the proprietary stuff had since the whole compiler/tools industry just collapsed suddenly.
I've used CL professionally, and I'd put it on par with C++ for that... powerful, well supported, without the annoyances of Java, but fundamentally flawed... at least the development environment (Emacs + SLIME) is better than anything other languages provide (sadly, that's what keeps me coming back to it for more abuse).
It's close to being reality. Guile has an Emacs-Lisp compiler to its VM that can run actual elisp programs, but lacks... the emacs part. And last summer's GSoC (perhaps this summer, finishing it?) saw emacs's lisp interpreter ported to guile... as in, the C representations of Emacs's internal data types and control structures are done using libguile. The code is currently being rebased on the latest emacs trunk; hopefully it'll see public release sooner than later.
Now the two pieces just have to meet in the middle.
So that's the first 95%. Now just for the other 95%!