I wish my world wasn't stuck on either GCC or MSVC. Then we might have more portable software. (Actually, I'm stuck with either GCC 4.7 or MIPSPro for what I want to use).

Careful, or they might start putting all the content in JS....

Are you paying Google for Google Maps? You don't own the product, you don't even lease/rent it. Since it is a web application, they can discontinue it at their discretion.

I've got fairly even performance here with a Quadro FX880m in this laptop going from Windows to Linux. I haven't tried a laptop with an AMD GPU in it for a very long time so I will refrain from commenting (I think the last one I had was a Radeon HD 3200) on that category.

Actually, that is something I have to agree with. I've generally found that ATI/AMD GPUs are on a hardware level as good (sometimes better) then nVidia, but God help them if they could write drivers under any OS to get the most out of them (I've also gotten the best performance out of an ATI card right before it was dropped from support...)

Which, unless you are a devout member of RMS's Church of GNU, doesn't matter to the rest of the world.

Until reality sets in that the /, crowd just wants to bitch and moan rather then try to improve or port anything. I'll take a working driver any day over a crippled open source one.

AMD is good for keeping Intel's prices down.

You missed the part where AMD chip design changed after the Phenom II (which was a nice processor in it's day and did compete against the Intel offerings). The Bulldozer based processors were a step backwards - a Phenom II x6 ran circles around any FX 81xx processor. The Bulldozer design is as big of a failure as Itanium.

I'd like to see a major update in the Radeon line, but they still preform well for certain workloads (like OpenCL). AMD could make a killing if they with a major GPU update, as long as it's not a flop like Bulldozer.

This is Slashdot you're asking the question on - a crowd that is much more inclined to tax their hardware then most other places.

For myself, it's not uncommon for me to have many machines running under very high loads (often BOINC, sometimes games). And the boost in gaming performance I saw going from an AMD FX-8120 to an i5 3570K (both were equal priced at the time) was incredible. AMD's newer FX chips (which are very old at this point) don't even try to compete with the i7 - not even in AMD's marketing material. Plus the heat generated by them is much greater then what I see out of Intel chips.

I have not seen the need just yet to upgrade my desktop from the 3570K to something newer, I might when the next generation Intel comes out, but I still have several older i7 devices still in use (all first gen - a desktop, and a Dell M4500) that manage just fine as well (I'd argue the 81x0 competes with the i7 9xx series).

Even still, direct profit isn't the only measure of success - are people using an open source (which isn't always free of cost) product of ours? Is it leading to them using other Microsoft products or buying support for open source products? Does it shut up a few people that might otherwise be looking at non-Microsoft solutions to other problems? Many different ways it can lead to profit.

It's a pain in my rear for one. Breaks compatibility with other Unix/Unix-like operating systems, for one. For two, (not sure if systemd or Fedora related issue) my laptop boots slower since the introduction of systemd, and prints out a crap-ton of extra information on screen and in the system log making it much harder to find the actually useful information that I want to see. Final big complaint - it's far more complex then it needs to be. I want init to be just that - init. I don't need it handling networking, authentication or any number of the other million things that systemd wants to do. You have no idea how much more simple it is to manage a box running Slackware, or and old UNIX like IRIX then it is to fight with modern Linux. Linux has lost the concept of KISS.

Do we really care one way or another? I mean, honestly. Attitudes like that won't encourage them to open up more code in the future - they need to see some level of success in the small set of stuff they put out to be convinced to do more (how success is measured is a different question...).

I've only had one issue with multiple screens and nvidia - it doesn't always save the order of the screens. Otherwise, the driver works flawless for me. KMS has been nothing but a pain in my rear (its introduction, along with the dropping of DRI1, forced me to stop using Linux on quite a few laptops that I had/was supporting at the time) and the fact that 'nomodeset' is required for nvidia is a plus in my book. I really can't speak about proper XrandR support as I have never dealt with it directly.

Different archs? Yes, support for Linux on some platforms is lacking (no nvidia on PPC, I've not seen an nvidia card in a MIPS, POWER, or IA-64 system so not surprised there is no driver support there), but nvidia supports more then fgrlrx does (nvidia works on Linux on ARM, plus FreeBSD AND Solaris on Intel).