In fact, they are expanding it -- they are putting in a brand new data center on the site that was the former Counterpane Security, on LaAvenida across from their SV HQ, and they also have leased a huge building a couple of blocks away on Pear Street. There's also rumors that they're behind the demolishing of an entire block of tilt-ups between LaAvenida and Pear to be replaced by six-story office buildings. In any event Microsoft isn't leaving the Silicon Valley, just Microsoft Research is leaving -- all fifty employees. Every single one of them who can have a job tomorrow by walking down the street to the Googleplex. Not a single one of whom have ever created a product for Microsoft, because Microsoft doesn't create products anymore, they just re-invent other people's products (or their own previously-good products), badly.
You're actually talking about MAID (Massive Array of Idle Disks), a technology that I first encountered in 2002. Now-bankrupt Copan Systems was the company I first encountered that was doing MAID, and New SGI (i.e. former Rackable Systems) bought their assets out of bankruptcy in 2010. Most storage companies now offer MAID add-ons for their storage arrays, though not all of them allow completely powering down the drive like Copan's solution did.
The upsides of MAID: Disks are cheap. Turning on and spinning up a hard drive to pull up some bits is faster than a robot fetching a Blu-Ray disk, placing it into a drive in the jukebox, and waiting for the disk to spin up and come online. You could store many more bytes in a cabinet with MAID than you could in an optical disk cabinet.
Downsides: The disk drives in a MAID array simply don't last that long, comparatively speaking. Spinning them up and down all the time is hard on a drive. So you end up having to replicate data and from time to time migrate data to new drives as old drives reach their service life. The service life of rarely used Blu-Ray media that has always been handled robotically (i.e., nothing touching its surfaces ever) is such that Blu-Ray media from ten years ago is probably still usable, the technology itself will become obsolete like DVD-RAM long before the media wears out. Not so much with hard drives, though disk arrays basically have unlimited life given typical failure patterns (i.e., if you're using RAID6, a drive develops errors, you remove the failing drive from the array, rebuild the array on a new drive, and chances of having two more drives fail during rebuild and thus losing the array are slim for a 12-drive array). So MAID has not really taken off the way we expected ten years ago.
At the time I first encountered MAID I was working for a company called DISC Storage, which had a NAS head which would automatically migrate little-used data to an optical jukebox in a way similar to what Facebook appears to be attempting. I designed and implemented the clustering function that would replicate the data between two NAS heads / optical jukeboxes, since the DVD-RAM platters were not themselves RAID'ed, as well as implemented a lot of the back end functionality for jukebox control and so forth. In any event, it looked like a NAS head but most of the files had been migrated to the DVD-RAM platters, and if you accessed one of those files, you would (at some point maybe 15 seconds later) get your data back as the file got read back onto the hard drive. It worked. But it was somewhat slow and cumbersome, because you're relying on a robot to go out and fetch the disk and put it in a drive, and disk robots then, and now, simply aren't that fast compared to media that's already in a drive ready to be spun up and read.
So anyhow, it was fairly obvious to me by mid 2003 that optical jukeboxes simply weren't going to be the future. In the ten years since DISC went under (there is a German company by that name now but it isn't the same company, it bought the name and some of the IP), I have not had any inclination to work for a company doing optical storage, because it's clear that for most problems it isn't the solution. It's too slow, too bulky, and magnetic disk drives and magnetic tape drives just continue getting bigger and cheaper every day. And now, with SSD coming on strong, optical jukeboxes look even less compelling.
So color me amazed. Optical jukebox and optical media technology essentially has barely moved on in the past ten years and what wasn't particularly compelling then, is even less compelling now. If you have need to keep data for a *long* time, this is how you do it... but frankly, I will be surprised if Facebook even exists ten years from now given the pace of innovation in the industry (though I'm just as surprised that Slashdot still exists!), so I question why they would do this rather than invest in LTO tape libraries, which have the advantage of being significantly denser.
Orson Scott Card is proof that ideology doesn't stop you from winning Hugos. All that is necessary is to write something interesting and mind-blowing. Writing formulatic space operas that are basically 1950's Don Pendleton manly war-fighting men set in space rather than in the wilds of darkest Africa or Southeast Asia is not the sort of thing that gathers awards, those have been done so many times that they're mind candy, something entertaining if in a certain mood but hardly mind blowing. Neither are fifteen page dissertations on why Libertarianism is the only proper ideology for running a space station. Nobody wants an ideological tract when they read science fiction, they want to have their mind blown. Orson Scott Card's later books haven't won Hugos not because of his right wing ideology, but, rather, because they've been boring, having nothing new to offer over his earlier fiction that did win awards. Which is a shame, because his 80's output was ridiculously good, then he got full of himself and his output got boring.
I liked Heinlein's The Moon is a Harsh Mistress as a work of fiction, and probably would have nominated it for an award that year if I hadn't been 2 years old at the time. L. Neil Smith's work, on the other hand, read like Libertarian tracts with cardboard characters and stereotypical plots, I read a couple of them and was, like, "meh", even though at the time I considered myself somewhat Libertarian and liked a number of other Prometheus Award winners. But ideology is *boring*. In the end, what I want when I read science fiction is something that's going to make me say "wow!", with a story that isn't predictable and characters that seem like real (if perhaps quirky) people and ideas that make me go "Hmm." Heinlein managed that for me, as do writers like Vernor Vinge, Ken MacLeod, and Neal Stephenson. Too many other "Libertarian" or "right-wing" authors don't. If you're going to turn it into an ideological tract with cardboard characters spouting ideological talking points, or just another by-the-numbers space opera where manly men of the finest cardboard kill lots of mooks while spouting pretentious jargon... YAWN.
This is pretty much my calculation. I don't swap glasses, I use progressive bifocals (the newer ones which have multiple focus zones rather than being totally continuous so that you can pretty much always have your workpiece in focus). Even if my vision without glasses was 20-20, I'd still need progressive bifocals, since I need multiple focal points to deal with my daily work (focus point for viewing computer monitor, focus point for viewing iDevice in my hand, focus point for viewing the speedometer in my car, focus point for dealing with tiny screws in computer cases, etc.). Given that, what's the point? I still end up needing glasses, I just add a bit of needless risk to the equation too.
Uhm, it does in fact cost money to post job ads on Craigslist. See http://www.craigslist.org/abou... . It's one of their major sources of money, along with hooker ads (oops, "therapeutic services", my bad).
Why you would do so, however, eludes me. As with the poster above, at multiple employers that have tested the Craigslist waters we've never had anybody remotely qualified respond to one of our Craigslist ads. They're just pocket change compared to Dice.com or compared to paying a recruiter's fee, so it was a "why not?". But every one of our hires seems to have been either a personal referral from a current employee, or a recruiter lured them away from another company.
Manual? What is that? Paper service manuals have gone the way of buggy whips in the auto service industry. Nobody publishes them anymore except maybe General Motors. It's all regularly-updated computer-based manuals for everybody else. Independents use Alldata, which gets updates on a regular basis from manufacturers, and dealerships use their manufacturer's computer system that does the same thing (such as Chrysler TechAuthority). For most newer cars if you are an end user who wants factory service info you go to AlldataDIY.com because paper service manuals are no longer published. For parts lists it's even simpler, you don't have to pay Alldata, the modern mechanic goes to the manufacturer's web site like Mopar.com or to some third party vendors which similarly give access to the manufacturer parts database and look up the part number there. If you pull an old part and put the part number off of it into the parts site search box, the computer will say "Superseded by" and give you the new part number. If there's no part number on the part, you drill down the assemblies list on the parts site until you see the labeled picture with your part on it, and the current part number will be in the list beneath that illustration.
And of course if you go to the dealership parts window, they put in the old part number into their computer, it says "superseded by part XYYZ", and they give you part XYYZ instead.
This is the 21st century. We have these COMPUTER thingies now. Just sayin'. There's no longer an excuse to *not* change the part number when the part has, in fact, changed. And I know for a fact that several part numbers on my Jeep have changed multiple times since it was manufactured in 2011. Which is why any modern automotive engineer has to be suspicious when GM did not change the part number on that ignition switch when, in fact, it's an entirely redesigned ignition switch...
Which, in combination with $1, will buy you a cup of coffee. I haven't noticed that Eastern European or Chinese spammers and attackers have been deterred one whit by those bilateral trade agreements.
Spammers definitely will *not* get a whole Amazon netblock blacklisted. Amazon firewalls outgoing port 25 traffic. If you want to send email from AWS you need to bounce it via authenticated port 465 or 537 through a mail service on some other ISP.
I'm not quite sure what you're talking about. Auto-scaling groups are the only thing that can restart a terminated instance (actually, they start a *new* instance). If you somehow managed to create an auto-scaling group and don't know how to set its parameters (min/max/desired) down to zero, when it's right there on the GUI, I don't know what to tell you.
Nobody really writes new applications in it, but there's a bajillion Java applications running most major corporations now the way COBOL used to in the old days.
That said, when looking for an engineer I'm not looking for someone who knows Java (even though that's what a lot of our product is written in). I'm looking for someone who understands computational complexity, is familiar with common algorithms and data structures, and has some notion of object oriented programming and software engineering. Anybody who's written a lot of code can pick up Java fairly swiftly at least to the "getting s**t done" stage, it took me roughly a week to do so ("oh, so it's like Python with C++ syntax, except with only single inheritance and with templates!"), but if you don't understand the why of what you're doing, you're not going to do well in our shop.
So: Get a computer science degree. Or at least significant computer science coursework. And not from Joe's Plumbing and Programming School, get one from some place that teaches actual computer science, not programming. Either that, or write some Open Source applications and contribute to the Linux kernel. Nobody cares what school you went to if you can write Linux drivers, all they care about is that you know the difference between a BIO and an SK_BUF. But they want to see your name in the Linux changelogs first.
A company which sells a solution at a higher price than other companies because of a higher cost of developing the software is soon to be an ex-company.
You're talking technology. But technology does not determine whether a company stays in business. Delivering in a timely manner a solution that works well enough for a cost equal to or less than the competition is what determines whether a company stays in business. There's a large number of application areas where Windows is what allows that. Luckily that's not *every* industry, or else I would have problems. (Disclaimer: I have been writing commercial software for Linux since 1996, yep, 18 years now).
My brother works in the SCADA industry. All of their stuff is Windows, mostly Windows XP Embedded. Why? Simple. It's the tools. There are various toolkits out there that make building a SCADA application almost drag-and-drop. It'd take four times as many people twice as long to hack all that up in C or C++ under Linux. And they simply don't sell enough SCADA systems to justify that kind of effort -- it's a crowded market where no single vendor manages to sell more than a few hundred instances of any particular model, so per-unit development cost difference between Windows and Linux far outweighs the OS cost difference.
As for why SCADA toolchain vendors don't port their tools to Linux, usually their tools are a large array of components from various vendors strung together with DCOM. Distributed SCADA systems in particular are heavily invested in Microsoft's DCOM OPC for communications between SCADA components such as pipeline pressure monitors, valve position sensors, billing stations, and operational monitoring stations. Linux doesn't support DCOM OPC as such, or any equivalent to it, with any standard libraries though there are emulators that may or may not work. The industry standardized on DCOM OPC for practical reasons -- it existed at the time they started doing all this (back in Windows NT days) while nothing like it existed on Linux back then, and they can write binary components that work pretty much on any Windows system, as versus with Linux where the distributions are not binary compatible and where five year old binaries will rarely run on a modern Linux system. Linux is great when you're selling a whole solution from top to bottom, but if you're trying to sell commercial software to SCADA system developers, Linux presents significant practical difficulties compared to WIndows. So there simply is no incentive to move off of Windows even though they're likely going to now be targeting later embedded Windows versions rather than embedded XP.
I'm not up on ATM's. But it would surprise me if ATM developers did not in fact use similar tools to create their product -- tools that are Windows-centric not because of Linux hatred, but because of history and the practical problems of trying to sell binary-blob commercial software on Linux (which is a task akin to nailing jelly to a tree).
H1B visas *can* be transferred, but it's a major annoyance. One of our H1B's was "officially" still on the payroll of the parent company after our division was split off as a separate company. This annoyed our former parent company greatly, since they had to pay him then bill us for his pay and benefits, but part of the divestment agreement called for them to do that for him until we could get the visa transferred. It took roughly six months to get the visa transferred
BTW, the United States is not the world. For example, the UK likes to poach the best and brightest H1B's from the US, I know three Indians formerly in the US on H1B's who've ended up in London. Canada also loves Indian immigrants who have a computer science degree and five years of work experience in computer fields, they put such immigrants on a fast track to a work permit and citizenship since they're perceived as having valuable skills and as being easy to assimilate. The notion that the H1B's only have a choice of India and the US as places to work is a false one. The US is still a major draw but if the US mistreats immigrants, they'll go elsewhere and the US will basically have spent a small fortune training them up for the UK, Canada, and other nations to get the advantage of such training.
Often the lowest-paid H1B's, the ones working for contracting companies like Tata, are stacked in communal apartments. One 1-bedroom apartment I visited had eight people living in it, they slept on futon mattresses thrown on the floor of the living room and bedroom and eat communally in the kitchen/dining area (there are multiple Indian grocers within walking distance of that apartment complex so it is easy to get cheap eats, though they tend to not be good cooks -- I almost swore off of Indian food after tasting a sample!). They rarely have cars, they rely on mass transit or employer bus to get to work. I drive by a bus stop every morning that has two dozen H1B's waiting for the bus so they can get to Cisco where they work. (I know it's Cisco because they have various Cisco-related stickers, binders, backpacks, etc. as well as Cisco badges that serve as their bus fare due to a special agreement Cisco made with the local transit authority).
Then there are the H1B's who are here because other work visas are too expensive, but otherwise are just normal employees. Several of my co-workers at my last company were H1B's. They had been working in the India office of the company when the India office was closed down as the company downsized, but were critical employees (they'd "owned" major parts of the technology as engineers and basically were irreplaceable due to their institutional knowledge of the innards of the technology) so were brought over on H1B. There needs to be a better way of handling that kind of thing other than the H1B with all its limitations and restrictions, but right now there really isn't, not unless your name is Linus Torvalds and you're brought in on a "genius" visa. We were always nervous when one of them went back home on vacation as to whether they'd be able to get back into the country. We endured the legal nightmare that was their work visa because they really were that important to our company. I presume they were paid accordingly.
In general I have no problem with the notion that it's a good idea to have work visas that can be issued to the best and brightest from all over the world. But let's not use bogus reasons to justify it. And the H1B with its myriad of limitations and restrictions is just plain obnoxious, we need something better if we really are trying to bring in the best and the brightest.
There are literally hundreds of thousands of trained and experienced American software engineers "on the bench" right now and 50% of graduates of U.S. computer science and IS programs don't even manage to find a job in the field. Google gets over 1,000,000 resumes per year. Microsoft may not get that many due to Microsoft's reputation as a "stodgy" and "yesterday's technology" company amongst the best and brightest (yes, I know that's unfair, but that's MS's reputation outside the MS "bubble") but I would be seriously surprised if Microsoft got less than 500,000 resumes per year. Microsoft hires about 3,000 people per year -- or roughly 0.6% of the people who submit resumes.
Given that, the notion that Microsoft can't find sufficient talent without H1B's says more about Microsoft's hiring process and Microsoft's reputation than it says about the availability of Americans with a background and training in software engineering. Especially telling is your notion that Microsoft should only look at the "top 20%" from a few "elite" universities. Frankly, given the criteria you mentioned, I wouldn't have been qualified to work for Microsoft upon college graduation because I wasn't "in the top 20%" -- mostly because I'd been working multiple jobs while in college, including writing actual software products shipped to actual real paying customers either as a contractor for various local companies or as an employee of a relative's company. I think my record over the past twenty years (multiple products shipped in multiple technologies ranging from PIC firmware for a front panel processor to Linux kernel driver work to Groovy/Grails code for a web app) shows just how silly Microsoft's criteria really are. There's a lot of talent out there that never makes it past that initial pre-screen where Microsoft immediately discards 80% of the applicants as "not good enough" without a single technical person ever talking to the applicant.
Frankly, our biggest problem when we go to hire people is not a shortage of candidates. It's too *many* candidates. My team doesn't have time to interview all the possible candidates who are submitted when we have a job opening, meaning it's a heavy filtering process. We rely on recruiters that we trust to do the initial prescreening, the ones we work with have technical backgrounds. Once they do the prescreening my boss is the guy doing the next level of filtering, and luckily he has a technical background too. The handful of candidates who make it to actual interviews generally all would be capable of doing the job, it becomes a case of deciding whether a candidate would fit with the team, stands out in some way, etc.
Unfortunately at many major corporations the people doing the initial filtering don't have a technical background and end up filtering on trivial criteria that discard good candidates for no good reason. That mostly lame people get dumped on your desk doesn't surprise me. It seems to be the norm for major corporations today where HR is doing the filtering. But that says more about a broken hiring process than it says about any shortage of trained and often experienced software talent.