Yeah, I agree. That's why at my job, everyone has a macbook pro for software development, because only a young kid would like it.
No, the reason macbooks are becomming so popular with THINKING ADULTS is the build quality on the macbooks is excellent, and you'd pay a lot of money to move beyond the fragile, hinges-breaking crappy ass plastic that you buy on your cheap-as-shit will break in 9 months disposable notebook. You can also dual boot to linux, if you need to sustain your fragile ego, but OSX runs a different variant of unix perfectly fine, and has a nice GUI on top of it.
All of these studies should posit that their results are in a baysian framework. Doing increases your risk of by times. IE: eating a high sugar diet increases your risk of getting diabetes by 20%.
Just because you don't do something that makes you more likely to have some condition doesn't rule out the possibility of having that condition. There are many factors to getting some diseases, both genetic and enviornmental, and we don't yet have mappings of most contributors...often, we'll wave our hands and say "whatever I can't map as genetics must be the enviornment" or "whatever I can't map to the enviornment must be genetics" (see the problem there?)
This is comming from someone who's spending his life working in the medical genetics community. People publish papers, the media digests it, but while we know something new, the vast majority of the time we JUST DON'T KNOW.
If I hadn't already commented similarly, I'd mod you up. R has been my bread and butter for serious (and ad-hoc) analysis for a few years now. It's fast to write, easy to get data into and out of, provides fantastic stats support, and creates beautiful graphs with very little effort. The interactivity is incredibly useful when prototyping.
I use python for certain tasks, and it's great and handy. For data analysis tasks, I tend to use R, which is both far more interactive than python, has much better graphing solutions than python or excel,and supports more statistical analysis methodologies than pretty much anything else. I can prototype and figure out a methodology in R (and provide provable results) long before I get to start running my python script.
If the product works, then he tells the truth and you think the product works great. If the product doesn't work, then he lies and says the product works great.
Yup. Bioinformatics is a big field now. I started in biology 10 years ago, but I'm now a very good engineer (who is also often the subject matter expert for our group.)
You forget about those of us using compute farms to do work. My place has 700 blades that are dual core right now. If we had 8 or 16 cores, that's a lot more available compute time - I often see 10-20K jobs pending to hit the farm, so having more compute to dispatch all those processes to would be great, and it wouldn't take up more server room (which is a big issue when space becomes limited.)
You can add to that a number of bioinformatics problems, where you're trolling millions of items on an entire genome, and running the same calculations on each and every one. These ought to be done in parallel - right now we often break them up by some natural partition like chromosomes, and run them across multiple machines.
I think the backlash was due to all the love-romance whiny crap that was substituted for a plot. Oh no, who's in the love triangle now?
If I wanted that sort of "plot", I'd watch Days of Our Lives. I think it's also an excuse to keep the cost of the episodes down - less CGI, special effects, more people giving each other pathetic looks and drinking fake booze. Speaking of which, where did they continue to get all that hard, but high quality booze from, since the still episode early on showed how they were scrapping the bottom very early on in the show.
I really like R - but I do a fair amount of scientific work, visualization, and statistics. I really like working with a matrix of data, and being able to apply functions across that matrix very quickly and easily.
I find R is also great for lots of ad-hoc data exploration that I need to do - but I'm sometimes as much of an analyst as I am a programmer.
Don't you mean "reducing our carbon emissions has to be priority #1"? I mean...if you're burning kittens, you're still releasing tons of CO2 / kitten. The goal doesn't preclude a number of solutions, not does it mean picking a worse emitter of CO2 that isn't coal.
Doesn't Rhino pay money to press the disks, create box cover art, store physical media in warehouses, etc? That all costs money, and if you don't sell any product, or enough to cover the costs, you...get this...lose money. That's risk.
If they had a secure digital download system, then they'd have less risk from those fees, but they'd have to create (or license) a service, which would also cost $$$.
USB1 has poor throughput too. Actually, isn't 54g better than USB1 (54 Mbit/sec vs 11Mbit/sec)? I know wireless isn't going to give you more than 25-50% of that, but that's still better.
Let me know when you can stream HD media from your time capsule to your new Mac. Wireless is great, but throughput sucks for continuous file transfer, which is why I always plug my mac in to move big files.
I had something very similar happen to me, involving a bunch of black ice. The car started spinning out at 70mph on ice, and we rotated towards a large bridge overpass, so the passenger door (me) was heading right toward it. I had very similar thoughts. I think I somehow got the word "f*ck" out about 200 times in those 4-5 seconds. It was like my brain was racing, but I couldn't do much with my mouth but sound a primal alarm.
"I say good driver. It appears as though we're about to be good and f*cked!
Dear scientist, do you know what the Hapmap project is? Do you know how many variants are cataloged there? Here's a hint: it's 3.1 million SNPs, and 270 people. Did you also know that you can have "statistical sample sizes" with N ~15. Tell me what the p-value is of a coin being unbiased if you flip 14 heads and 1 tail. What kind of scientist are you anyway - I'm guessing neither a biologist nor a statistician.
I've published a paper using the hapmap data as a backing for selection, and we had a few absurdly high p-values (like p1e-100). It's no small amount of data.
I think the argument is that even if the larger servers cost more to purchase, they use less space and need less cooling for the amount of CPU power you have. So they save you a lot more money over a 3 year period. It's one of those TCO style arguments. I think there's also the bonus of having an easy way to manage the machines if the hardware is heterogeneous, as you can still put the same VM image on all the machines.
That's changing back as people realize how much you can do with virtualization. When you consider how much space you need for racks, cooling costs, etc, it often makes sense to buy a few big boxes instead of a lot of tiny ones, and run virtual machines. So I'd think the market there is going to swing back as VMs get better and better. (Full disclosure, my wife works for VMWare, so I've heard a fair number of talks on how to run these numbers.)
Slashdot Mirror
Yeah, I agree. That's why at my job, everyone has a macbook pro for software development, because only a young kid would like it.
No, the reason macbooks are becomming so popular with THINKING ADULTS is the build quality on the macbooks is excellent, and you'd pay a lot of money to move beyond the fragile, hinges-breaking crappy ass plastic that you buy on your cheap-as-shit will break in 9 months disposable notebook. You can also dual boot to linux, if you need to sustain your fragile ego, but OSX runs a different variant of unix perfectly fine, and has a nice GUI on top of it.
All of these studies should posit that their results are in a baysian framework. Doing increases your risk of by times. IE: eating a high sugar diet increases your risk of getting diabetes by 20%.
Just because you don't do something that makes you more likely to have some condition doesn't rule out the possibility of having that condition. There are many factors to getting some diseases, both genetic and enviornmental, and we don't yet have mappings of most contributors...often, we'll wave our hands and say "whatever I can't map as genetics must be the enviornment" or "whatever I can't map to the enviornment must be genetics" (see the problem there?)
This is comming from someone who's spending his life working in the medical genetics community. People publish papers, the media digests it, but while we know something new, the vast majority of the time we JUST DON'T KNOW.
If I hadn't already commented similarly, I'd mod you up. R has been my bread and butter for serious (and ad-hoc) analysis for a few years now. It's fast to write, easy to get data into and out of, provides fantastic stats support, and creates beautiful graphs with very little effort. The interactivity is incredibly useful when prototyping.
Prototype in R, and if you need to make it repeatable, maybe push it into python (or write a wrapper that gives nice command line options in python.)
I use python for certain tasks, and it's great and handy. For data analysis tasks, I tend to use R, which is both far more interactive than python, has much better graphing solutions than python or excel,and supports more statistical analysis methodologies than pretty much anything else. I can prototype and figure out a methodology in R (and provide provable results) long before I get to start running my python script.
If the product works, then he tells the truth and you think the product works great. If the product doesn't work, then he lies and says the product works great.
Hooking the salesman up proves nothing.
Yup. Bioinformatics is a big field now. I started in biology 10 years ago, but I'm now a very good engineer (who is also often the subject matter expert for our group.)
You forget about those of us using compute farms to do work. My place has 700 blades that are dual core right now. If we had 8 or 16 cores, that's a lot more available compute time - I often see 10-20K jobs pending to hit the farm, so having more compute to dispatch all those processes to would be great, and it wouldn't take up more server room (which is a big issue when space becomes limited.)
You need a half way decent video card - they do a lot of the heavy lifting for HD video compression now.
You can add to that a number of bioinformatics problems, where you're trolling millions of items on an entire genome, and running the same calculations on each and every one. These ought to be done in parallel - right now we often break them up by some natural partition like chromosomes, and run them across multiple machines.
I think the backlash was due to all the love-romance whiny crap that was substituted for a plot. Oh no, who's in the love triangle now?
If I wanted that sort of "plot", I'd watch Days of Our Lives. I think it's also an excuse to keep the cost of the episodes down - less CGI, special effects, more people giving each other pathetic looks and drinking fake booze. Speaking of which, where did they continue to get all that hard, but high quality booze from, since the still episode early on showed how they were scrapping the bottom very early on in the show.
But one shouldn't even mention Torchwood.
I really like R - but I do a fair amount of scientific work, visualization, and statistics. I really like working with a matrix of data, and being able to apply functions across that matrix very quickly and easily.
I find R is also great for lots of ad-hoc data exploration that I need to do - but I'm sometimes as much of an analyst as I am a programmer.
Don't you mean "reducing our carbon emissions has to be priority #1"? I mean...if you're burning kittens, you're still releasing tons of CO2 / kitten. The goal doesn't preclude a number of solutions, not does it mean picking a worse emitter of CO2 that isn't coal.
Doesn't Rhino pay money to press the disks, create box cover art, store physical media in warehouses, etc? That all costs money, and if you don't sell any product, or enough to cover the costs, you...get this...lose money. That's risk.
If they had a secure digital download system, then they'd have less risk from those fees, but they'd have to create (or license) a service, which would also cost $$$.
I plug my Macbook Pro into a 720p projector and a dedicated 24bit DAC (usb in, line level out) to power my home theater.
USB1 has poor throughput too. Actually, isn't 54g better than USB1 (54 Mbit/sec vs 11Mbit/sec)? I know wireless isn't going to give you more than 25-50% of that, but that's still better.
Comcast is doing this in the states with sandvine. It sounds like your ISP may have the exact same equipment installed.
Let me know when you can stream HD media from your time capsule to your new Mac. Wireless is great, but throughput sucks for continuous file transfer, which is why I always plug my mac in to move big files.
I had something very similar happen to me, involving a bunch of black ice. The car started spinning out at 70mph on ice, and we rotated towards a large bridge overpass, so the passenger door (me) was heading right toward it. I had very similar thoughts. I think I somehow got the word "f*ck" out about 200 times in those 4-5 seconds. It was like my brain was racing, but I couldn't do much with my mouth but sound a primal alarm.
"I say good driver. It appears as though we're about to be good and f*cked!
http://www.brainyquote.com/quotes/quotes/d/douglasada109730.html
Dear scientist, do you know what the Hapmap project is? Do you know how many variants are cataloged there? Here's a hint: it's 3.1 million SNPs, and 270 people. Did you also know that you can have "statistical sample sizes" with N ~15. Tell me what the p-value is of a coin being unbiased if you flip 14 heads and 1 tail. What kind of scientist are you anyway - I'm guessing neither a biologist nor a statistician.
I've published a paper using the hapmap data as a backing for selection, and we had a few absurdly high p-values (like p1e-100). It's no small amount of data.
You mean the 1,000 repetition at the end of book 3? I almost pulled my braid I was so frustrated.
I think the argument is that even if the larger servers cost more to purchase, they use less space and need less cooling for the amount of CPU power you have. So they save you a lot more money over a 3 year period. It's one of those TCO style arguments. I think there's also the bonus of having an easy way to manage the machines if the hardware is heterogeneous, as you can still put the same VM image on all the machines.
That's changing back as people realize how much you can do with virtualization. When you consider how much space you need for racks, cooling costs, etc, it often makes sense to buy a few big boxes instead of a lot of tiny ones, and run virtual machines. So I'd think the market there is going to swing back as VMs get better and better. (Full disclosure, my wife works for VMWare, so I've heard a fair number of talks on how to run these numbers.)