There are 10 kinds of people in the world: those who understand ternary, those who don't, and those insufferably pretentious arsewipes who thought that this was a binary joke...

IIRC you could create a small partition on the primary HDD and install onto it. Then you'd install drivers, apply patches to ahci.sys and then turn on EnableBigBLA (or whatever it was called) registry key. After that use diskpart to extend the partition worked ok.

But the point is that it gets to be a royal pain to do this for every install due to the stock install not being able to handle modern hardware.

XP can't go over 2TB for the boot volume, doesn't recognise 4k sectored storage devices (it still works, but with suboptimal performance until a sector-alignment tool is used), and (correct me if I'm wrong) doesn't moderate its wear of solid state drives.

So, as I said above, it's quite nice to move away from XP.

That's the plan with SCCM. We'll see how it goes.

Sysadmin here:

We've migrated about 50 users (salesforce, most are aged 40 to 60) to Win7 from XP about 2 months ago. Here's what's happening for us:

* No major problems adjusting to Win7 (I've had a couple of quick questions, that's it).
* Running users as standard users is almost viable (we're having a lot of pain and suffering from all the crapware we have to install (Adobe Flash, Reader, Shockwave; Quicktime, iTunes; Java; etc, etc) -- almost everything on this list wants admin rights to update itself). Users can't install much or tweak much, so much less user-induced OS failure/slowness/malware. We're trialling SCCM for this, so we'll see...
* Win7 seems less prone to malware infection. I doubt it's anywhere near secure, but it's already doing a lot better than XP. (I'm forced to use Symantec for AV, which is about as much protection as a pincushion condom.)
* Device drivers for modern PCs on XP is a royal pain; Win7 is ok for now (a couple of bad device drivers for Win7 x64, but much better than XP x64 and good enough for use), and updating device drivers from Windows Update works about half the time.
* Imaging tools are much nicer.
* Sleep and hibernate seem to be more reliable. XP would fail to resume 1 in every 200 resumes or so.

So for us, Win7 is a major step up -- it isn't that it's good so much as it sucks much less than XP (which sucked much less than 98, etc.). Furthermore, ReactOS (last I checked) is far, far, far away from being a viable replacement. MS could sit still for 5 to 10 years and ReactOS would still be far away. Give those guys several more good programmers and the story might be different...

Not sure about the US, but in AU companies are set up to execute their charters. Most charters say "make profit", but they don't have to. CEOs are expected to fulfill the charter, which is not necessarily the same as make a profit. (I think many charters have some lip service toward social responsibility and so on as well, but this is usually ignored when money is at stake.)

I don't know the hardware, but the thought that immediately comes to mind is that 90% of the hardware that the target audience will use does not support wifi -- machines with 128MB of RAM and the like.

Would you care to elaborate on this? It'd be interesting to know what they got right (especially the separation of presentation layer from the code).

If it's any interest for you, the word is gynoid (Greek for woman-like, and with the same roots as misoGYNy and polyGYNy). ANDRoid is man-like, same roots as misandry and polyandry.

Hi, new poster here but have been lurking for about a decade -- but as fucked up as IPSec is, there are some important benefits:

* IPSec tunnels your traffic over an unreliable datagram protocol (either IP protocol ESP or over some UDP port -- I forget the number). This avoids the performance problems of running a reliable protocol (TCP) over another reliable protocol (TCP). Some time since I looked at this, but IIRC, retransmits in the upper protocol kill you. Probably not too bit a problem if you aren't running significant traffic.

* IPSec is processed in kernel mode which improves processing performance. This isn't as important on the client which is only handling one tunnel as it is on the gateway which is handling many connections and where the CPU load could be important. Disadvantage is that a bug in IPSec is a bug in kernelspace.

* Of course anyone doing something like this should terminate the IPSec connection on a network outside their LAN and should also consider blocking comms between indials.

Just wish whoever designed IPSec had done a proper job.