As we know from recent history, many companies have pursued criminal charges against people who have broken into their systems, even if they had no intention to do harm. And DAs are very happy to prosecute these cases. So this would be a great virus to write if you wanted to add a prison sentence to your resume. (And you could be held responsible for the cost of restoring the servers after your failed upgrades, which could run into the millions of dollars if you damaged enough servers.)

Back in 1996, I heard two different people complain that since Microsoft convinced anti-virus companies to support Windows 95, and drop support for Windows 3.11, new viruses for Windows 3.11 were forcing them to upgrade to Windows 95. At least one of these folks was perfectly happy with Windows 3.11, had a huge stash of Windows 3.11 games on CDs and just plain didn't want to upgrade. He saw the whole thing as a charade to force people to shell out for upgrades.

This is playing out again pretty much word-for-word with the termination of support for Windows XP.

It would have been a more valid complaint for someone compelled to upgrade from Win3.1 to Windows 95.

At this stage, Windows XP users are like Japanese soldiers stranded on an island, growing old not knowing they lost the war. Except without the excuse of being on an island.

I run Win7 on my main machine but I have laptop that still runs XP (and which isn't connected to the internet, in general), because I have a large piece of audio hardware whose driver support stopped with XP but which is otherwise fine, and which I have no desire to replace ahead of time. I'm not alone in this; while XP is far from ideal as an audio production environment, later versions of Windows suffer from really terrible MIDI (control protocol for synthesizers) drivers with unreliable timing.

Of course I could get a Mac, bt I'd really rather not switch platforms or buy a bunch of extra expensive hardware for this one purpose. I could also run Linux, but until very recently the selection of music production software has been poor.

That said, I don't have any complaint with MS about their decision to abandon the platform, I think they've done a good job in supporting it this long. I do wish they'd rethink their driver priorities though, MIDI has been around for over 30 years, and being 7-bit ~32kbps it can't be that hard to get decent realtime performance.

I don't know what you're doing but I have rock solid midi timing from Windows 7 via a £5 generic MIDI/USB cable to my Korg Triton Studio. I'm using Renoise as a DAW.

XP was crappy - it hit the disk so often and hung.

It's a problem on multi-output USB interfaces for a lot of people. I suspect system configuration issues but I don't think you should have to take a barebones system approach to get decent timing. I mainly sequence in hardware so I haven't gone to great lengths to get to the bottom of it.

DYI http://www.gearslutz.com/board/electronic-music-instruments-... although you'll have to wade through some non-techie voodoo interpretations if you peruse the whole thread. Also, bear in mind that this is a particularly fussy demographic. I sometimes think if you are that obsessed then maybe MIDI isn't the protocol for you, but OTOH there isn't much in the way of OSC hardware on the market :-/ It's notable that there's a bit of a CV renaissance underway though, a lot of people prize timing/resolution over versatility.

Depends how they implemented midi and the timers. If it's win32 MIDI API stuff and Win32 timers the messaging isn't real time so you will get jitter. That's going to be worse on win vista and above due to a number of issues.

If it uses ASIO MIDI back end it'll miss all the problematic message loops.

I only use SYSEX and use the Triton Studio sequencer though as its the most awesome thing ever made by man.

Is there any reason for such a single-purposed machine to be connected to the internet?

Occasionally I grab a software update or somesuch so I'll connect via wifi rather than bothering with USB transfers. It's not in a commercial environment and there's only a few trusted sites I might be visiting.

As someone stuck supporting a Windows XP server, I know we lost the war. We raised a big white flag and rolled out a welcome mat for our new overlords. Still waiting for someone to show up, though.

(To break the metaphor, our server is hooked up to a PCI-X card which runs a radiation sensor. The drivers for the card will not run under anything after XP. The vendor has discussed possibly considering looking into designing a new card that would might run under Vista, but that's probably five years down the road.)

Sounds like you need a new vendor? As an alternative, could you run XP in a minimum VM, which has that particular PCI card's address passed directly to the VM? Then make sure that there is no other way into the XP VM (i.e., don't run any other apps on it, and keep it only on the host-only network).

XP Embedded? Still has support.

Not for long. XPe is already at EOL. Support ends in 2016 and license availability ends in 2017.

Our products using XP Embedded are expected to still be in the field well past 2020. We are already making plans to transition to Windows 8 simply because it will soon be impossible to buy XPe licenses.

> Even if it fails, a broken server that used to process credit cards is better than a vulnerable one, right?

"I saw that you dropped your wallet on the floor, so I threw it into the river so nobody could steal your credit cards."

If returning the wallet (securing the server) is out of the question, throwing my wallet in the river could easily be preferable to leaving it loose. Of course if someone told me that in meatspace I wouldn't believe them... hm.

You should only intentionally kill their server if you're willing to compensate them for the revenue they lost while they try to figure out what the hell happened.

An alternate analogy would be something like: "I saw you forgot to install a lock on door, so I superglued it shut."

You don't fix bad behavior with bad behavior.