I just installed the Windows 7 Phone SDK and all is component parts, which was a long but easy chore.  Well, the first time it hung forever and I had to remove a bunch of components before it would install correctly.

The process of getting a simple phone app built and running on the emulator was about like I expected.  I already learned basic XAML programming, and I found nothing extraordinary about the Phone classes (so far).

However-- (big pause) The level of support for real network programming on the WP7 is abysmal.   First, you can only use the WCF "service" model, which currently only runs over HTTP.  In other words, you must have a server hosting a WCF service or no go.  The claim (see this forum among many others) is that resource-poor Microsoft didn't have time to test sockets correctly before releasing this still-wet-behind-the-ears new product.  Maybe so.

On the other hand, maybe their urgent desire to establish Azure as the cloud platform of choice had something to do with this.  If you can't do your own networking you're forced to rely on hosted faciliies, either web servers or cloud communications.  Either way, Microsoft gets more money from you.  

Since HTTP runs over TCP and HTTP is supported in its web browser, I sense something altogether fishy in Microsoft's limitations for this first version of the WP7. Why couldn't they simply allow synchronous (blocking) communications over single TCP sockets, which is probably what the built-in web browser is doing?  Reason-- because they want to kill peer-to-peer.

What really (really) concerns me as a civil libertarian is that this coincides with ongoing, onerous attacks against peer-to-peer operations from both government and big business.  I sense that there are deep, powerful forces at work that want to work with ISPs to "white-list" the internet and kill peer-to-peer operations.  The crippled version of the WPF7 may just another example of a product brought to market too soon.  Or it could be part of a larger, decentralized effort to kill arbitrary point-to-point internet connections.  Government wants to monitor everything, and PTP is messy to monitor.  Businesses assume everyone doing PTP is an IP pirate and want to block it all.   Now, arguably the biggest product launch of the year for a smartphone fails to include the most basic communications capability for the internet.  Accident?  Hmmmm.

Suffice it to say that without sockets or multitasking, this product is more a toy than a platform.  On the iPhone, most people run their applications an average of twice.  Microsoft seems to be depending on games and social networking sites to drive WPF7 sales.  They will probably fail.

Well, I think I finally got the Windows 7 partition on my MacBook backed up.  The support staff at Paragon told me that their product, Drive Backup Pro 10, couldn't back up under Windows (using either Volume Shadow Service or their "Hot" technology).  What I had to do, they said, was use the CD recovery drive and run it from there.
So I did, and it appeared to work.   However, there's no way to be sure unless and until I attempt a restore, and I do not do restores lightly.
But my confidence grew today when I successfully used the Drive Backup Pro recovery disk to "upsize" the main drive on my webserver.  Yesterday I used the program to back up the boot drive on it, but it reported an unrecoverable read error.  Luckily, it only reported one.  So I hit "Ignore" and carried on.  Then I went to Hard Drives Northwest and bought one of a vanishing stock of new EIDE drives.  After installing it as the primary, I partitioned the new drive and restored the full partition backup.
At almost every step there were errors that could have been better explained.  I wondered, for example, why it reported read errors when an extensive "chkdsk" (check disk) found nothing.   In addition, the boot disk program reported read errors on the BACKUP drive until I partitioned and formatted the new C: drive.   After that, it ran flawlessly.
So now I have a new drive in my server, and about a 70% confidence level that I'm fully backed up on my dual-boot Mac.  As an intelligent consumer, what I really want is programs that give intelligible errors.  For example, why couldn't Paragon's program simply have an error message saying, "Live (hot) backup not supported under Bootcamp.  Please use recovery boot disk image."  That would have saved me literally hours of nonsense.
Oh, well.

Well, I just completed rebuilding Bonnie's web site, http://groupwyse.com, and adding a blog and a forum.  The blogware is the same as this blog, BlogEngine.Net; the forum is the same as our forum, AspNetForums.  The real learning was about using LunarPages 'plesk' and the 'mylittleadmin' tool for SQL Server 2005.

Plesk does let you do a lot.  Indeed, I keep finding new things I really can control as administrator.  However, the overall documentation is very poor (as usual), and some things remain elusive, such as controlling redirection.

MyLittleAdmin, on the other hand, is so much like the SQL Server 2005 Enterprise Manager that it usually only takes me a little poking around to find exactly what I want.  I'm so used to being able to control EVERYTHING that it's refreshing to have something so similar available to me. 

I've been briefly reviewing solutions for mobile device browser rendering.  It seems that not all devices support the 'handheld' media type for a CSS 'link'.  This is unfortunate, but appears to be improving steadily.   From what I've seen of mobile stylesheets, they can be extraordinarily large and complex.

It seems to me that mobile devices may require significant reengineering of a web site's look and feel.  Merely 'skinning' with small screen boundaries and copious use of the 'visibility' attribute won't really, I don't think, turn a sow into a piglet.  Active recognition of target devices would require correct, stable reporting of device characteristics in the GET/POST request.  An alternative would be to emblazon "click here for mobile devices" on the front of my web site(s), but that seems ungainly.

In addition, the inconsistent treatment of even standard CSS by many mobile browsers leads to unfortunate testing requirements.  How many phones is one supposed to have?  Why don't browser makers or some clever web hacks produce some 'proofing' tools where you can set the phone/PDA type and see it rendered?  Mobilify talks about this, but many folks still recommend downloading the Andriod SDK and using its emulator.  I've done that, and the set of steps and requirements is not trivial.

Here's hoping for a brighter, if smaller, future.