Mass numbers of computations are used in numerical weather prediction for one thing - and NWP is what allows forecasters to be come more accurate about severe storms, hurricane track and intensification, etcetera. The models already operate on massively parallel machines - but faster machines or less latency will allow them to forecast for smaller grids, improving accuracy, and if they run faster, they can be run more times per day which also helps forecasts adjust to changing conditions.
Yes, I believe you _can_ get a lot of the raw data via NOAA's site, it is I believe free for US citizens since we already paid for it through funding NOAA. Maybe you can use Google's new algorithm for detecting 100,000 features in an image on a single computer to make the model run on you desktop in usable time - remember, as one of the NOAA sites quotes a scientist: if your model doesn't run in faster than real time, you might as well just swivel your chair to look out the window (assuming I'm not being an insensitive clod who doesn't pay attention to the fact that you don't have a window).
Once upon a time we were all going to have a 'user agent' that was local to our computer that performed all sorts of mediation for us, sort of an Arthur Treacher (look it up you young whippersnappers) role; this is one instance of that idea.
Once saw a FORTRAN compiler change cause _slightly_ different numerical results in a bridge engineering program. They did some work to verify that bridges built using the program compiled under the previous version of the compiler were structurally sound but that was the most fearful thing I could think of, that numerical inaccuracies in what the compiler generates could cause an engineering failure that could cause loss of life.
COBOL, which is compiled, doesn't need another language to run. It does, in many implementations, require a run-time library to supply common functions which are called by the compiled code - for example, the DISPLAY verb may be implemented as a common subroutine. On IBM mainframes, the run-time library support for many languages is provided by the Language Environment product.
Rexx is still one of the best procedural-style scripting languages... I use it in Unix (well, Unix Systems Services, the Posix-compliant interface component of z/OS)
Maybe it's just my mainframe background, and maybe I haven't searched hard enough, but I don't see in the laptop/desktop/server enviroment a lot of detailed presentations of performance issues, how they were diagnosed, and how they were solved. I see them very occasionally. Is there a lack of good performance recording (and I mean recording for analysis after the fact, not real-time monitoring) tools, or is it just that purchasing faster hardware components has probably become cheaper than spending time diagnosing the issues?
Ala carte channels would be governed by the law of supply and demand; so if you and five of your friends are the only ones who want to watch "The Channel That Shows Video from Alt.Plan9.OS.Demos" the price is going to be pretty high. None of the ala carte proposals I've ever seen say that the channel provider has to make those channels cheap.
"X-Wait" doesn't mean it's never coming back... if you notice it, it means that response time is slower than you expect but it still might come back. Many mainframe block-mode terminals and application "servers" like CICS (CICS is most analogous to Tomcat though the comparison isn't exact) "lock" the terminal until the transaction response is produced, to keep people from entering hundreds of transactions in a row and losing track of which response goes with which transaction. So, if a transaction is delayed for some reason, your terminal is still "locked". Impatient people don't like the situation, reset the terminal emulator, and enter the transaction again - but the only effect of that is to throw away the result of the first transaction when CICS goes to send it (because the terminal session was lost) and put your transaction at the back of the queue.
That's not to say that things _never_ get hung up in the locked state, of course there are failures where the response never comes back, but it's good to pick some amount of time (30 seconds, 1 minute) to wait before going through the "throw it away and start over" process.
I don't need or want my watch to check my e-mail, flash when I have voicemail or any of that stuff. I'd be happy with a watch just smart enough, with a very very simple API, to allow me to design and download new "skins" to the watch when I feel like it. It would have a screen no bigger than my current Timex (including the bezel), and have these basic UI elements:
Background
12 or 24 hour time
Digital or analog display supported
Analog display Hand designs - including transparency for those people who just want a little dot rotating around or whatever.
The font for the digits; perhaps a separate one for the day/date display
A button to cause it to brighten, dim, or be blank when I want it to not glow in the dark
Download one or two alarm sounds
Since I don't want to change the watch theme extremely frequently, I'm OK with micro USB to do that, I don't need wireless.
Of all of the above, the most important point is to _not_ make the thing look like I've strapped some cellphone to my wrist!
I hate to be pedantic, but 5081 Hollerith cards had 12 holes per column, which means you could actually encoded 12 bits per column if you wanted to. 12 bits x 80 columns = 960 bits per card, or 120 bytes. 120 x 3,000 = 360,000 bytes; divide by 1024 = 351.5 KB
"I'm just a simple caveman,..." with a mainframe background, so I have a question of curiosity here
At what point does the bandwidth/throughput of the DMA start limiting the performance of your backup?
In my world, DMA for I/O is called a "channel". We have many, and while there are a lot of nuances we could discuss, basically we try to segregate the I/O for the input to backup (disk) and the output of backup (usually tape) , and have the backup task process in parallel as much as possible - my nightly backup, for example, runs 9 parallel tasks, 9 being the limit that this particular backup program has. I could run multiple instances of the program, but then I have to have mechanisms to make sure I don't back up the same disk twice between two concurrent executions; with one instance and 9 tasks I can just say 'back up everything that's online at the moment'. So, the throughput is limited by the performance of the slowest devices, multiplied by the parallelism we are able to achieve. In the PC / server environment, does the DMA limit the I/O capability?
Chaos Monkey start up get working
Chaos Monkey is a hoot
Chaos Monkey's preferred new target
is instance of a group...
(the rest is left as an exercise for Coulton fans)
As many have said, this is Windows on a blade, in a frame that is part of the mainframe box
It will most benefit Windows-based applications which access mainframe things on the back end (such as GUI.Net apps with DB2), because the servers are attached through a high-speed internal network.
The system management tool for the hardware will provide unified management of the z/box and the blades, which will help some folk.
There actually was a "Windows" implementation on the mainframe at one time, Bristol had ported Wind/U (a Windows API implementation) to z/OS Unix Systems Services - but after some pushback from Microsoft I believe their license to do so was revoked.
Let's not forget that Palm had a lot of good experience developing simple UIs for use on portable devices and they had some good design ideas for not wasting battery life in applications, either. Some of the PDA functionality that Palm was so good at wouldn't be bad to have on a Kindle, really.
What is your philosophy on software tools? Do you prefer to use a lot of small pieces, loosely assembled, using scripts to join things together and get things done, or do you like to find a software suite (such as Office) and work within that?
"Look at me! I said look at me!! There now, that's better. That's a good boy... now, type in the password... . Good, good... Now turn away, I don't want you to see the routing code and account numbers for these funds. You just don't need to know now, do you? "
Slashdot Mirror
Mass numbers of computations are used in numerical weather prediction for one thing - and NWP is what allows forecasters to be come more accurate about severe storms, hurricane track and intensification, etcetera. The models already operate on massively parallel machines - but faster machines or less latency will allow them to forecast for smaller grids, improving accuracy, and if they run faster, they can be run more times per day which also helps forecasts adjust to changing conditions.
Yes, I believe you _can_ get a lot of the raw data via NOAA's site, it is I believe free for US citizens since we already paid for it through funding NOAA. Maybe you can use Google's new algorithm for detecting 100,000 features in an image on a single computer to make the model run on you desktop in usable time - remember, as one of the NOAA sites quotes a scientist: if your model doesn't run in faster than real time, you might as well just swivel your chair to look out the window (assuming I'm not being an insensitive clod who doesn't pay attention to the fact that you don't have a window).
Once upon a time we were all going to have a 'user agent' that was local to our computer that performed all sorts of mediation for us, sort of an Arthur Treacher (look it up you young whippersnappers) role; this is one instance of that idea.
See this page; the Campanile movie is from SIGGRAPH 97. How is Disney's tech different?
How do you store the instructions for using this thing?
Once saw a FORTRAN compiler change cause _slightly_ different numerical results in a bridge engineering program. They did some work to verify that bridges built using the program compiled under the previous version of the compiler were structurally sound but that was the most fearful thing I could think of, that numerical inaccuracies in what the compiler generates could cause an engineering failure that could cause loss of life.
COBOL, which is compiled, doesn't need another language to run. It does, in many implementations, require a run-time library to supply common functions which are called by the compiled code - for example, the DISPLAY verb may be implemented as a common subroutine. On IBM mainframes, the run-time library support for many languages is provided by the Language Environment product.
Rexx is still one of the best procedural-style scripting languages... I use it in Unix (well, Unix Systems Services, the Posix-compliant interface component of z/OS)
Maybe it's just my mainframe background, and maybe I haven't searched hard enough, but I don't see in the laptop/desktop/server enviroment a lot of detailed presentations of performance issues, how they were diagnosed, and how they were solved. I see them very occasionally. Is there a lack of good performance recording (and I mean recording for analysis after the fact, not real-time monitoring) tools, or is it just that purchasing faster hardware components has probably become cheaper than spending time diagnosing the issues?
Ala carte channels would be governed by the law of supply and demand; so if you and five of your friends are the only ones who want to watch "The Channel That Shows Video from Alt.Plan9.OS.Demos" the price is going to be pretty high. None of the ala carte proposals I've ever seen say that the channel provider has to make those channels cheap.
"Which is more impressive: Wi-Fi hotspot in 2013, or E-ink display in 2008?" sounds like a new iteration of " Quien es mas macho ?"
"X-Wait" doesn't mean it's never coming back... if you notice it, it means that response time is slower than you expect but it still might come back. Many mainframe block-mode terminals and application "servers" like CICS (CICS is most analogous to Tomcat though the comparison isn't exact) "lock" the terminal until the transaction response is produced, to keep people from entering hundreds of transactions in a row and losing track of which response goes with which transaction. So, if a transaction is delayed for some reason, your terminal is still "locked". Impatient people don't like the situation, reset the terminal emulator, and enter the transaction again - but the only effect of that is to throw away the result of the first transaction when CICS goes to send it (because the terminal session was lost) and put your transaction at the back of the queue. That's not to say that things _never_ get hung up in the locked state, of course there are failures where the response never comes back, but it's good to pick some amount of time (30 seconds, 1 minute) to wait before going through the "throw it away and start over" process.
Since I don't want to change the watch theme extremely frequently, I'm OK with micro USB to do that, I don't need wireless. Of all of the above, the most important point is to _not_ make the thing look like I've strapped some cellphone to my wrist!
Plastics are primarily made from petrochemicals - i.e. petroleum-based, so their prices fluctuate with the price of oil.
I hate to be pedantic, but 5081 Hollerith cards had 12 holes per column, which means you could actually encoded 12 bits per column if you wanted to. 12 bits x 80 columns = 960 bits per card, or 120 bytes. 120 x 3,000 = 360,000 bytes; divide by 1024 = 351.5 KB
"I'm just a simple caveman, ..." with a mainframe background, so I have a question of curiosity here
At what point does the bandwidth/throughput of the DMA start limiting the performance of your backup?
In my world, DMA for I/O is called a "channel". We have many, and while there are a lot of nuances we could discuss, basically we try to segregate the I/O for the input to backup (disk) and the output of backup (usually tape) , and have the backup task process in parallel as much as possible - my nightly backup, for example, runs 9 parallel tasks, 9 being the limit that this particular backup program has. I could run multiple instances of the program, but then I have to have mechanisms to make sure I don't back up the same disk twice between two concurrent executions; with one instance and 9 tasks I can just say 'back up everything that's online at the moment'. So, the throughput is limited by the performance of the slowest devices, multiplied by the parallelism we are able to achieve. In the PC / server environment, does the DMA limit the I/O capability?
Chaos Monkey start up get working Chaos Monkey is a hoot Chaos Monkey's preferred new target is instance of a group... (the rest is left as an exercise for Coulton fans)
Isn't this the Lytro camera concept?
This is not the answer you're looking for... Could you post a reply here with whatever you chose to do?
Perhaps the discovery of graphene will be viewed as the event which tipped us over into the Singularity !
I believe "graphene" is single-molecule-thick silicon, isn't it?
Let's not forget that Palm had a lot of good experience developing simple UIs for use on portable devices and they had some good design ideas for not wasting battery life in applications, either. Some of the PDA functionality that Palm was so good at wouldn't be bad to have on a Kindle, really.
What is your philosophy on software tools? Do you prefer to use a lot of small pieces, loosely assembled, using scripts to join things together and get things done, or do you like to find a software suite (such as Office) and work within that?
"Look at me! I said look at me!! There now, that's better. That's a good boy... now, type in the password... . Good, good... Now turn away, I don't want you to see the routing code and account numbers for these funds. You just don't need to know now, do you? "