I'm surprised that Amazon isn't using any of the existing methods of creating virtual machine images, especially since it seems likely that they are using some sort of VM technology to do all of this. VMs are the easiest way to provide identical hardware to all of the instances; VMs also provide the tools to efficently load and unload images on demand. I wonder if we'll see different pricing depending on how CPU intensive an instance is. I can see using an instance as a server for less-popular online games; you'd only pay for those times when someone is playing. Depending on how fast multiple instances could be ramped up, you could also provide slashdot-effect insurance; just bring more instances on-line as more people arrived and shut them down as the peak passes.
megaditto said: Also, there is a difference between 100 people on a plane going down and 3000 in two towers going down; so as bad as it sounds, shooting down hijacked planes might be a good idea after all.
You do realize that the terrorist were planning to blow up 10 to 12 planes, right? And trans-Atlantic flights (depending on the plane) hold 300 to 500 passengers? That would be anywhere from 3,000 to 6,000 deaths. A massive, co-ordinated attack that causes planes to be destroyed could easily exceed the 9/11 death toll, which is what the British terrorists were after.
VMWare is a nice product but A it is too expensive, and be it is too expensive because it turns any VM machine into a basket case performance wise.
Except, (a) VMware's Server and Workstation products are free, and (b) VMware runs its guests as well as virtual infrastructure. Xen has its way of dealiing with performance issues, VMware has other ways, but nobody's going to use anything that won't run the guest OS resonably fast. And once Intel's VT release arrives, do you think that VMware isn't going to use the new instructions? Right now, VMware is the only virtualizer that runs both Windows and Linux guests, and it does a damn fine job of it.
the immunity to the negative effects of LDH cholesterol developed in a single man in Italy (creating descendants among whom heart disease and strokes are vanishingly shockingly rare).
Can someone provide a reference for this? Googling for "LDH cholesterol Italy" doesn't turn up anything useful.
First, the screen shots should be on the front page. I poked around and couldn't find any, then stumbled across them by accident a few minutes later.
You claim to handle "issues and relations". Define those terms. Give me a concrete example of what the tool is used for. Arrange the screenshots to tell that story.
Who are your target users? Tier one support? Tier two? Project managers? All of the above? Have a story for each.
I wish you guys still in school or freshly graduated knew how often the rest of us roll our eyes every time we hear "my thesis" or "my prof". Your college education will be of marginal vaule to you as a programmer. Most of what you use will be from experience gained after you graduated. You won't necessarily be a good programmer just because you made good grades either.
Yeah, tell that to Larry Page and Sergey Brin (Google), Scott McNealy, Andy Bechtolsheim, Bill Joy and Vinod Khosla (Sun Microsystems) and Fred Smith (FedEx). Those companies all started out as college term papers.
Don't laugh! Most (if not all) filesystems are optimized to handle the opposite of what you want. TAR files are designed for tape, so you won't be seeking all over the disk to get meta information, instead you'll get your data at the maximum speed supported by your hardware. TAR files are designed so that you can append files to them later, so you can use a *big* partition and just keep dumping stuff into it.
The only drawbacks are that you have to read the entire partitioin sequentially to find things, and you can't delete files. Both of these can be fixed with a bit of Perl. Write a program that maintains an index of offsets to the files, then you can use "dd" to skip to the correct offset and read from there. More dangerously, write a program that deletes files from the middle of an archive and shuffles everything backwards to fill in the gaps. You'll want to make sure that no one is trying to read the TAR partition while this is running.
Actually, the latest Whedon interview has him saying that he was misquoted in his immediately previous interview. All he said was that the movie provided closure to the dangling plot threads from the truncated first season. He also said that there are other stories he'd like to tell.
Once peak oil/global warming/the singularity hits, no one will be launching anything for quite a while. This will give the crap in low orbit a chance to drop lower and eventually burn up in the atmosphere.
Good Lord, I'd almost repressed my memories of Digg's comments. I quickly learned to never look at them unless there were more than a hundred, otherwise you'd rot your brain trying to find anything worthwhile. I think that part of the problem is that recaps are limited to just a sentence or two. It encourages a telegraphic writing style that everyone copies in the comment sections. And you'll notice that no one *ever* moderates any comments. For the first few months, everyone there seemed to be cutting-edge types, now it's being taken over by the other end of the bell curve. I've thought that they need an affinity system of some sort; you know "Peoply who dugg this article also dugg these as well." That way there'd be some chance of escaping the seething mass of banality that's increasingly being posted. Or maybe I just need to go back over everything I ever dugg and see who posted it, then make them my friends or something. No, too much work. Forget I said anything about it.
"The problem with Slashdot is that the signal to noise ratio of comments has been getting worse and worse." Not as bad as over at digg. For a while, I found myself spending more and more time at digg, and correspondingly less time here. Over the past couple of weeks, though, the front page stories at digg have become less and less interesting to me, so I'm spending more time at slashdot again. I fear that any sort of automatic promotion system will eventually suffer a similar fate. I guess editorial control really is useful, after all.
"Also, we are unable to log into the PCs when they are disconnected from the network. Local accounts aren't allowed. You have to log in to the domain server."
I think that you mean that you have to log in *using* the domain server. It still sounds bogus, though. Windows caches account info locally. I have a laptop and I use domain accounts while off-net all the time; it pops up a window saying that it can't connect, but lets me log on anyway. This could be a configuration option; you should check into it.
"However, we had to keep the PC connected to the network because the test scripts called software that had to contact our license server."
Insert your own firewall server between your test boxes and the rest of the LAN that allows traffic to the license server to pass through but blocks the rest of the world.
An array of mirrors reflects light into a dome atop a tower. The dome contains a circulating supply of water that is heated up into steam and used to drive a turbine. The mirrors are automatically angled to reflect the sun (at pretty much any angle) into the dome.
Ever hear of dust? Those mirrors only reflect well for a few weeks or maybe months, then they get dirty from airborne particles settling on them, and then you need an army of homeless people with squeegies and spray bottles to make everything work again.
The bad news is that successful consultants need good interpersonal skills, because the interact face-to-face with their clients. The good news is that since we interact face-to-face with our clients, we don't see our jobs out-sourced to India.
Can vi(m) run a shell in an editable buffer, or are you just talking about the:shell command?
Do you mean the '!' commands? '!!' lets you type a command, runs it with the current line as stdin, and replaces the current line with stdout. '!}' does the same with the current paragraphs, '!G' to the end of the file. I'll often use an 'o', combination to create an empty line to sacrifice to commands like 'ls'.
Here's another killer feature that I'll bet vi(m) doesn't have: you can remotely connect to a running copy of XEmacs.
No, vi(m) adheres to the Unix philosophy of having many small tools that do one thing well, instead of one Swiss-Army-knife that tries to be all things to all people. VNC (for X11) and screen (for tty) both let you do everything you describe, and for more that just your editor.
I've used Emacs for fifteen years, and XEmacs almost as long.
I've used vi for twenty.
First, I can use Emacs without taking my hands from the keyboard, ever. I can compile, debug, run a shell - you name it, I can do it without having to reach for the mouse.
Ditto for vi.
Second, it is customizable in the extreme. Everything from key bindings to highlighting is driven by Elisp and regular expressions. Don't like the way something works? You can quickly and easily change it by rebinding a lisp function; most importantly, you can make these mods on the fly, without having to run a separate compile step, without having to restart the editor.
Try vim.
What I'd like to see is an editor that combines the best of Emacs and Eclipse. You'd never have to take your hands from the keyboard. You'd get the attractive UI of Eclipse without the Visual envy. You'd get an editor that makes you more productive and happy than any other.
"The first and only time it was tried and, it worked... better. "
Read your whitepaper:
"The version of Apache that was being used was not multi-threaded so each request was handled by another Apache process that was spawned off by the parent process. Spawning a new process is costly and Perl is an interpreted language so the performance of these machines was not optimal. One of the first tasks undertaken by the dev team when Microsoft purchased Hotmail was to convert all the CGIs from Perl to C++."
So, when MS says "better", do they mean better than the original site, or better than the one running five minutes earlier, because it seems to me that most of the improvements were related to moving from Perl to C++ and from CGI to long-running processes ("The next big performance increase came from eliminating the need to load and parse the Hotmail-specific configuration files for every request.").
And finally, "All the memory management calls in the code are overridden. Each time new, delete, malloc, or another similar call is made, the memory is allocated or freed from the threads private heap. That heap is thrown away between each request so the memory leaks went away." I think that says it all right there. The C++ code that MS wrote in-house using Visual C++ was buggy WRT memory leaks, so they fixed it by rewriting the entire memory management subsystem.
From the Microsoft whitepaper: "Acquired by Microsoft in 1997..." and "During June and July of 2000, the Hotmail site was converted from FreeBSD running Apache Web services to Windows 2000 Server running Microsoft Internet Information Services 5.0."
So, you'd have us believe that during those three years, no effort was made to run the site on anything other than Win2K? Re-read your exact words used: "The first and only time it was tried and, it worked... better." Exactly what does "it" mean in that sentence. Does it mean the first and only time the migration to Win2K was tried, it worked? Because I refuse to believe that no attempts were made to move it to 1997's primary server OS, WinNT. Instead, I suspect that the stories are true, a migration was attempted to WinNT, it didn't work, and three years were spent optimizing Win2K to handle the task.
And as for the "better" part, don't forget that there were hardware upgrades as well. A migration to the latest FreeBSD and Apache versions would have worked better as well.
So here's what you want. You want highly-skilled developers to volunteer months of their time to write this free of charge. Then you'll turn around and charge MA $100/hr to implement OOo. Sounds wonderful. I'll get right on that. I love how all the Open Source junkies want developers to donate so they can charge big bucks to deploy the Open Source software.
Just yesterday, I was just thinking about the similarities between open source and community theatre. Both are built on the efforts of unpaid volunteers who nonetheless work long, hard hours to perfect something that is unlikely to be a financial success. Both have corporate sponsors who pick up the tab for needed capital expenditures. And both have a few paid professionals to work full-time to keep everyone moving in the same direction and hopefully make the deadlines.
You gotta love Ozzie's use of buzzword generators.;-) Heck, the entire paragraph is a great example of obfuscation by buzzword:
Some problems are inherently complex; there is surely no silver bullet to reducing complexity in extant systems. But when tackling new problems, I've found it useful to dip into a toolbox of simplification approaches and methods. One such tool is the use of extensive end-to-end scenario-based design and implementation. Another is that of utilizing loosely-coupled design of systems by introducing constraints at key junctures - using standards as a tool to force quick agreement on interfaces. Many such tools are not rocket science: for example, by forcing a change in practices to increase the frequency of release cycles, scope and complexity of any given release by necessity is greatly reduced. Another simple tool I've used involves attracting developers to use common physical workspaces to naturally catalyze ad hoc face-time between those who need to coordinate, rather than relying solely upon meetings and streams of email and document reviews for such interaction. Embracing change at a local level through such tools can make a real difference - one project at a time.
"Extant" -- that's a great word; I said it to someone on the elevator just last week. Of course, she then slapped me.
Slashdot Mirror
I'm surprised that Amazon isn't using any of the existing methods of creating virtual machine images, especially since it seems likely that they are using some sort of VM technology to do all of this. VMs are the easiest way to provide identical hardware to all of the instances; VMs also provide the tools to efficently load and unload images on demand. I wonder if we'll see different pricing depending on how CPU intensive an instance is. I can see using an instance as a server for less-popular online games; you'd only pay for those times when someone is playing. Depending on how fast multiple instances could be ramped up, you could also provide slashdot-effect insurance; just bring more instances on-line as more people arrived and shut them down as the peak passes.
You do realize that the terrorist were planning to blow up 10 to 12 planes, right? And trans-Atlantic flights (depending on the plane) hold 300 to 500 passengers? That would be anywhere from 3,000 to 6,000 deaths. A massive, co-ordinated attack that causes planes to be destroyed could easily exceed the 9/11 death toll, which is what the British terrorists were after.
Except, (a) VMware's Server and Workstation products are free, and (b) VMware runs its guests as well as virtual infrastructure. Xen has its way of dealiing with performance issues, VMware has other ways, but nobody's going to use anything that won't run the guest OS resonably fast. And once Intel's VT release arrives, do you think that VMware isn't going to use the new instructions? Right now, VMware is the only virtualizer that runs both Windows and Linux guests, and it does a damn fine job of it.
Depending on where and how you drop the aluminum, nickle, iron and silicon from the asteroid, you can get rid of all of the warlords at the same time.
Get a copy of VMware Player (it's free at http://www.vmware.com/download/player/) and load any of several Jabber virtual appliances. (You may have to change the virtual ethernet adaptor to use the bridged network.) Here's one applicance that looks like a close fit: http://www.vmware.com/vmtn/appliances/directory/24 8
Can someone provide a reference for this? Googling for "LDH cholesterol Italy" doesn't turn up anything useful.
On the other hand, http://www.nytimes.com/2006/05/09/health/09canc.ht ml?ex=1148097600&en=1832c1470567bcd5&ei=5070 tells of a mutation in mice that renders them immune to cancer.
You claim to handle "issues and relations". Define those terms. Give me a concrete example of what the tool is used for. Arrange the screenshots to tell that story.
Who are your target users? Tier one support? Tier two? Project managers? All of the above? Have a story for each.
> Does that mean that Richard Nixon was German? His tapes of everything, IMHO, exceeded even the Nazis.
Yeah, tell that to Larry Page and Sergey Brin (Google), Scott McNealy, Andy Bechtolsheim, Bill Joy and Vinod Khosla (Sun Microsystems) and Fred Smith (FedEx). Those companies all started out as college term papers.
The only drawbacks are that you have to read the entire partitioin sequentially to find things, and you can't delete files. Both of these can be fixed with a bit of Perl. Write a program that maintains an index of offsets to the files, then you can use "dd" to skip to the correct offset and read from there. More dangerously, write a program that deletes files from the middle of an archive and shuffles everything backwards to fill in the gaps. You'll want to make sure that no one is trying to read the TAR partition while this is running.
Actually, the latest Whedon interview has him saying that he was misquoted in his immediately previous interview. All he said was that the movie provided closure to the dangling plot threads from the truncated first season. He also said that there are other stories he'd like to tell.
Once peak oil/global warming/the singularity hits, no one will be launching anything for quite a while. This will give the crap in low orbit a chance to drop lower and eventually burn up in the atmosphere.
Good Lord, I'd almost repressed my memories of Digg's comments. I quickly learned to never look at them unless there were more than a hundred, otherwise you'd rot your brain trying to find anything worthwhile. I think that part of the problem is that recaps are limited to just a sentence or two. It encourages a telegraphic writing style that everyone copies in the comment sections. And you'll notice that no one *ever* moderates any comments. For the first few months, everyone there seemed to be cutting-edge types, now it's being taken over by the other end of the bell curve. I've thought that they need an affinity system of some sort; you know "Peoply who dugg this article also dugg these as well." That way there'd be some chance of escaping the seething mass of banality that's increasingly being posted. Or maybe I just need to go back over everything I ever dugg and see who posted it, then make them my friends or something. No, too much work. Forget I said anything about it.
"The problem with Slashdot is that the signal to noise ratio of comments has been getting worse and worse." Not as bad as over at digg. For a while, I found myself spending more and more time at digg, and correspondingly less time here. Over the past couple of weeks, though, the front page stories at digg have become less and less interesting to me, so I'm spending more time at slashdot again. I fear that any sort of automatic promotion system will eventually suffer a similar fate. I guess editorial control really is useful, after all.
"Also, we are unable to log into the PCs when they are disconnected from the network. Local accounts aren't allowed. You have to log in to the domain server."
I think that you mean that you have to log in *using* the domain server. It still sounds bogus, though. Windows caches account info locally. I have a laptop and I use domain accounts while off-net all the time; it pops up a window saying that it can't connect, but lets me log on anyway. This could be a configuration option; you should check into it.
"However, we had to keep the PC connected to the network because the test scripts called software that had to contact our license server."
Insert your own firewall server between your test boxes and the rest of the LAN that allows traffic to the license server to pass through but blocks the rest of the world.
Ever hear of dust? Those mirrors only reflect well for a few weeks or maybe months, then they get dirty from airborne particles settling on them, and then you need an army of homeless people with squeegies and spray bottles to make everything work again.
The bad news is that successful consultants need good interpersonal skills, because the interact face-to-face with their clients. The good news is that since we interact face-to-face with our clients, we don't see our jobs out-sourced to India.
Do you mean the '!' commands? '!!' lets you type a command, runs it with the current line as stdin, and replaces the current line with stdout. '!}' does the same with the current paragraphs, '!G' to the end of the file. I'll often use an 'o', combination to create an empty line to sacrifice to commands like 'ls'.
Here's another killer feature that I'll bet vi(m) doesn't have: you can remotely connect to a running copy of XEmacs.
No, vi(m) adheres to the Unix philosophy of having many small tools that do one thing well, instead of one Swiss-Army-knife that tries to be all things to all people. VNC (for X11) and screen (for tty) both let you do everything you describe, and for more that just your editor.
BTW, http://slashdot.org/~vrmlguy/journal/13611
I've used vi for twenty.
First, I can use Emacs without taking my hands from the keyboard, ever. I can compile, debug, run a shell - you name it, I can do it without having to reach for the mouse.
Ditto for vi.
Second, it is customizable in the extreme. Everything from key bindings to highlighting is driven by Elisp and regular expressions. Don't like the way something works? You can quickly and easily change it by rebinding a lisp function; most importantly, you can make these mods on the fly, without having to run a separate compile step, without having to restart the editor.
Try vim.
What I'd like to see is an editor that combines the best of Emacs and Eclipse. You'd never have to take your hands from the keyboard. You'd get the attractive UI of Eclipse without the Visual envy. You'd get an editor that makes you more productive and happy than any other.
Try gVIM.
Read your whitepaper: "The version of Apache that was being used was not multi-threaded so each request was handled by another Apache process that was spawned off by the parent process. Spawning a new process is costly and Perl is an interpreted language so the performance of these machines was not optimal. One of the first tasks undertaken by the dev team when Microsoft purchased Hotmail was to convert all the CGIs from Perl to C++."
So, when MS says "better", do they mean better than the original site, or better than the one running five minutes earlier, because it seems to me that most of the improvements were related to moving from Perl to C++ and from CGI to long-running processes ("The next big performance increase came from eliminating the need to load and parse the Hotmail-specific configuration files for every request.").
And finally, "All the memory management calls in the code are overridden. Each time new, delete, malloc, or another similar call is made, the memory is allocated or freed from the threads private heap. That heap is thrown away between each request so the memory leaks went away." I think that says it all right there. The C++ code that MS wrote in-house using Visual C++ was buggy WRT memory leaks, so they fixed it by rewriting the entire memory management subsystem.
So, you'd have us believe that during those three years, no effort was made to run the site on anything other than Win2K? Re-read your exact words used: "The first and only time it was tried and, it worked... better." Exactly what does "it" mean in that sentence. Does it mean the first and only time the migration to Win2K was tried, it worked? Because I refuse to believe that no attempts were made to move it to 1997's primary server OS, WinNT. Instead, I suspect that the stories are true, a migration was attempted to WinNT, it didn't work, and three years were spent optimizing Win2K to handle the task.
And as for the "better" part, don't forget that there were hardware upgrades as well. A migration to the latest FreeBSD and Apache versions would have worked better as well.
Seems to me that the Chinese government is taking the whole peak oil thing pretty seriously.
Yep, you'll need those to lift the damn thing.