Updated 2nd April 2006: I removed the comment about battery life, since I've since found it to be fine - it must've been a case of special circumstances before. Or One Of Those Things.

I cut my Pocket PC teeth on an HP Jornada 545, and for years after judged every Pocket PC on whether it was built like a tank and had an attractive flip top and annoyingly badly-seated stylus.
I must confess that when I attended PDC 2003 I was hugely envious of the local (to NZ) Microsofties and Intergenistas who managed to pick up 4x00-series devices at obscenely low prices. I was personally unable to take advantage of any such offers since I was using most of my meagre luggage allocation on the whopping great big security book we were handed on the first day.
That's what you get for being conscientious.

So anyway, years pass and I find myself with a deceased hx4700 and a Harrier which while still happily doing its thing, is sadly lacking in Wi-Fi capability and can never be upgraded to WM5.0.

I had been thinking of buying a Jasjar, but apart from my normal condition of semi-abject poverty, I have heard less than wonderful things about performance, and there is the matter of the inherent nippiness of EVDO to be considered.
The recent announcement of UMPC clarified a few things for me: I have always leaned towards the PC end of the Pocket PC scale - I have little use for cell phones (other than for their ability to connect to the internet), and have always looked for the most computing power I could get in a mobile situation.

None of this is directly relevant to the rx1950, but with patience it will explain my choice.
It has been clear to me for some time (at least since the "Compuiting on the beach" session at PDC 2003) that the destiny of the Pocket PC was to be swallowed between ever-more-capable smartphones and shrinking tablet-like devices. In other words it will go the same way as the mini. This prospect does not fill me with alarm, since the devices that replace it will finally fulfill the promise of the technology. This is especially true in a managed world, since we will be able to make the transition from the Compact Framework (semi-famously and affectionately described by Rory Blythe as a wrapper around System.NotImplementedException) to the full framework, and before long to WinFX.

So rather than buying the most powerful Pocket PC I could afford, I decided to buy a reasonably priced and non-crappy Windows Mobile 5 device, partly in the expectation that this is the last Pocket PC I will buy (because afterwards I'm likely to be buying UMPCs and Smartphones).

So far I've been quite pleased with it: the The screen is about as good as I could hope for in a non-VGA device, and I have not discerned any trace of the yellow tinge that blighted some earlier models.
It pays to manage memory carefully on the rx1950 (in other words, shut things down rather than leaving them lying around...but then, everyone other than Microsoft has known for years that that was a good idea). It doesn't have a lot to play with, and when things get tight will become unresponsive (i.e. you'll probably find yourself doing a warm restart). On the other hand, I was relieved and gratified to see that  the .NET Compact Framework 2.0 (which takes up a lot of memory on installation) installed without any problems...which also means of course that the rx1950 can easily run the kind of software I've been writing for the last few years.
Other obvious points: I'm not a big watcher of video on mobile devices (since that frankly seems a bit pointless), but other people have reported that the rx1950 is somewhat sluggish in that regard. On the other hand, it has no problems at all with my extensive library of resolutely non-trendy music. 
So in other words it may not be a great machine to show off on, but from what I can tell it's a brilliant machine to use.
Last but not least, the 4xxx design was probably as close as anyone has come to the perfect design for a PDA without a keypad, so it's great to see an updated version (please note however that while the rx1950 is equipped with WiFi, it has no Bluetooth capability).

As a developer, my natural reaction when a new software or hardware platform is released is always "What kind of applications can I write for that?" (actually, that's my automatic reaction to anything. You name it: toast, cats, whatever. Speaking of which, if anyone knows why strange women were walking around Blenheim last week with toast on their heads, will they please tell me? Is it a special Blenheim variant of Morris Dancing? It troubles me. ).

The newly announced Ultra-Mobile PCs offer a mix of challenges and opportunities, some centring around the native 800 x 480 resolution, and some deriving from the fact that it is at heart a Tablet PC but with the addition of touch. And don't forget the small physical size (and likely to get smaller).

As Alex Yakhnin points out, this is a platform well suited to the skills of Windows Mobile developers.
Since it is running a full version of Windows, some of the software opportunities that exist on the Windows Mobile platform are absent or lessened, since a lot of Windows software will "just work". On the other hand, by comparison with desktop Windows and even the tablet marketplace I think we'll find that some software categories that were not worth attempting owing to the presence of dominant encumbents will suddenly be opened up: if an established application provides a bad user experienced on the UMPC platform, there will be opportunities for nimble-footed newcomers to get in and make some waves.

Or at least I very much hope so :-)

As for my thoughts on the UMPC itself, it's interesting (although no doubt coincidental) that Brighthand is reporting that HP will no longer be offering the hx4700: for many of those of us who lean towards the upper end of the mobile device scale, high-end PDAs lose a lot of their reason for existing with the introduction of the UMPC...we'll finally see an end to most of the necessary compromises that have made those devices in many ways almost useful, but not quite. So the day that was portended at the "Computing on the Beach: Visions of Mobility" panel discussion at PDC 2003 may finally be on the way - my bet is that the high-end PDA will go the way of all flesh (even though it doesn't have any, unless you've been very careless), just as the mini was squeezed between ever more capable PCs and miniaturising mainframes. At the very least I doubt we'll see many more VGA Pocket PCs: while the displays are gorgeous, they suck up a lot of power and introduce technical issues that have impacted software compatibility, and the Windows Mobile implementation of VGA did not give us the kind of resolution improvements many had hoped for. Having said that, there'll no doubt be a swag of VGA Pocket PC announcements next week just to make me look like an idiot...