@spivonious: it depends on what you mean by ".NET". To me (and I guess to quite a few developers out there), .NET is a set of languages, libraries and technologies; the CLR is just an implementation detail. They already made major changes to the CLR in the past, I don't see why this should be any different.
@W3bbo: hmm... if that's what came across, my English must have become a trainwreck. My apologies.
I'll try to rephrase:
I installed the (official) developer Mango build on the (official) Samsung firmware. All went well, but the compass doesn't work. It wasn't supported in NoDo and it isn't mandatory for Mango, so things may stay as they are, for all I know.
The presence of a new leaked firmware, with compass support, indicates that Samsung may release a new firmware someday. If that's not included with the Mango RTM update, though, I would rather not rush the release of my little augmented reality app.
And for the record: I'm not complaining
I recently found out that my Samsung Omnia 7 reports that the digital compass is not supported, despite the fact that the hardware includes one. Apparently that's due to a missing device driver, which is included in a leaked build that has been floating around for a while.
Anyway, the fact that I could install Mango just fine on the current firmware poses an interesting question: will the new firmware become a prerequisite for Mango RTM or will that be handled as an OOB update by the manufacturer? The reason why I ask is that without a compass the whole Motion class becomes unavailable and considering that the Omnia is a popular WP7 model some places, this might affect the available market for a number of apps.
Does anybody have any pointers on the matter?
@wastingtimewithforums: I don't see why you should expect a BSOD. New hardware won't have any driver installed, so it should just not work; and since each driver is reinitialized on resume, any driver that refers to missing hardware should just fail gracefully (assuming that drivers are written correctly). At worst, the new hardware you just popped in won't seem to work.
At this point, I would expect that any driver failing to reinitialize on resume would trigger a full enumeration, and that the same happens as part of the driver installer. This would cover all the scenarios I can think of, with "power users" messing with the command line, and everybody else just popping in the CD that came with the device (something they may have to do anyway).
Considering that the vast majority of users will never ever mess with their hardware, I think that's an excellent idea.
@Doctor Who: In Mango, there's a keypad button to the right of "end call" that brings up the full keypad. It's not a feature I use very often, so I can't say if that's new or there was something similar in NoDo... the good thing is that apparently the problem was already taken care of.
Thinness is the least of Microsofts worries here. IMO:
- Windows was written first for CISC instruction set. RISC versions have come and gone (alpha, powerpc). One of the many reasons Windows did not take off on RISC chips early on is because its performance was not overwhelming better on the far superior chip. RISC to Windows is second class and as such will perform poorly as they'll be using shims and shortcuts to get Windows running on RISC.
I'm confused. I regularly port code from RISC to CISC and viceversa and never had to write a shim: the C compiler will take care of optimizing my code for the specific architecture. Also, unless I'm much mistaken, the only area where the Alpha really shined was floating point math; not something an OS would have much use for.
@Richard.Hein: hard to tell from a video, but the quality of the scanner looks good. My only issue with the device is that considering the number of documents I scan vs the miles I push the pointer around the screen, I'd rather have an excellent mouse that also happens to be an ok scanner than the other way around.
@DeathByVisualStudio: Once all the competitors have an even remotely comparable feature set, the commercial success of a device bear little or no correlation to its technical merits.
As technical enthusiasts, we tend to get carried away about some feature that, in our opinion either crushes the competition flat or makes the whole platform a total failure. The general public will happily ignore it (probably because they can't even spell it) and base their decisions on more relevant parameters like "what did my friends buy?", "who offers the best deal?" or "does it come in pink?".
P.S. Yep, the Google conspiracy theory is ludicrous.
Welcome back, Courier... I won't hold my breath this time, though.
As for patenting gestures, I agree that it's ludicrous. I would venture to say that it shouldn't be possible to patent something for which no future alternative is imaginable... if you patent a device, a sensor or a technology, even if unique, chances are that a viable alternative will turn up someday. But there's only one way I can make a specific gesture and no amounts of research will ever change that.