Why Stack Exchange Isn’t in the Cloud

Post by fleitz » Thu, Nov 17, 2011, 11:01 PM UTC

wmf wrote: post #36

quanticle wrote: post #19

The reason I'm still wary of the cloud is that the abstractions are still too leaky (to borrow Joel Spolsky's turn of phrase). When you start abstracting away core system calls (like fsync), things work great 99.99% of the time. But when that .01% bites you, it bites hard . We don't expect core system calls (like fsync) to fail. And when they fail, the fact that our code is two or three levels levels of abstraction h…

In what way does fsync work differently in the cloud?

Post by ianburrell » Thu, Nov 17, 2011, 11:13 PM UTC

wulczer wrote: post #21

quanticle wrote: post #19

The reason I'm still wary of the cloud is that the abstractions are still too leaky (to borrow Joel Spolsky's turn of phrase). When you start abstracting away core system calls (like fsync), things work great 99.99% of the time. But when that .01% bites you, it bites hard . We don't expect core system calls (like fsync) to fail. And when they fail, the fact that our code is two or three levels levels of abstraction h…

Cloud or not, if you're writing serious software (and by that I mean software that depends on fsync succeeding) you'd better expect system calls to fail and have code to handle that.

Post by Donwangugi » Thu, Nov 17, 2011, 11:36 PM UTC

mechanical_fish wrote: post #13

jtbigwoo wrote: post #5

I suspect that this is part of the problem with many big manufacturing companies these days. Ford was a leader for so many years was because their leadership was passionate about engineering and manufacturing. Henry Ford was an engineer for the Edison Company before he was a manager and is famous for obsessing about every detail of the manufacturing process. That sort of spirit infected the company up through the 196…

A Ford bought today is less expensive, considerably safer, more efficient, and more reliable than a Ford from the days of Henry Ford. Meanwhile, today's manufacturing companies routinely make objects – like the processor driving the machine that you are reading this on – that are orders of magnitude more complex than the most advanced technology of 1959. And all of this was done by companies that employ lots of accou…

Post by mey » Thu, Nov 17, 2011, 11:57 PM UTC

reinhardt wrote: post #42

gbeech wrote: post #35

Earlier quoted context omitted.

> eliminating internal sysadmin staff completely (and outsourcing all the remaining sysadmin work to Amazon) Man, this idea is one of the most insane things I hear. Amazon does not install your operating systems, Amazon does not install and configure you applications, Amazon does not troubleshoot anything related to the operation of the servers, Amazon does not do analysis on your traffic patterns and decide when you…

You conveniently left out the first part of the quoted sentence, "allow programmers to actually program infrastructure". Almost everything you wrote can be programmed these days without a "traditional" sysadmin.

Post by tedunangst » Fri, Nov 18, 2011, 12:13 AM UTC

MattyDub wrote: post #31

MichaelGG wrote: post #4

What a silly argument. "If you just want to use someone else’s computers, it means you don’t love computers — at least not every aspect to them." Uh, don't they use Dell? If they love them so much, why not build their own? Why stop there? I love software and programming languages, but it doesn't mean I'm going to use my own compiler or run a company on my own libc.

Note that Joel's company Fog Creek software actually did develop their flagship product, FogBugz, on their own compiler: http://www.joelonsoftware.com/items/2006/09/01b.html