http://www.businessweek.com/technology/content/may2008/tc2008056_305255.htm?campaign_id=rss_tech

Here’s one of the first articles in a more mainstream publication, that reflects on some of the conflicts of interest that may exist between iPhone and Android; and also between Apple and Google, especially with Schmidt, CEO of Google sitting on Apple’s board, when the companies at competing against it each other in the mobile phone space.

http://www.businessweek.com/technology/content/may2008/tc2008056_305255.htm?campaign_id=rss_tech

I wasn’t “trying to state that Windows at the lowest level isn’t portable to other machine processors,” that’s just the strawman you found easier to attack that the point I actually made: that Microsoft tried to deliver NT as a portable OS and failed miserably, years after the much smaller NeXT seamlessly accomplished the same task.

All Microsoft had to do was copy. Instead, it did the typical sloppy job where its products were designed by marketing pinheads instead of engineers. Cutler did make great pains to develop a hardware abstraction layer, and that attempted to accomplish what MS planned for NT, but by the time NT was functional (v5 Windows 2000), cross platform efforts were dead.

You can blame chickens and eggs and the market for not having enough Alpha and MIPS machines to deliver a functional NT port, but the obvious question is: Why could the much smaller NeXT deliver functional ports to not only Alpha and MIPS, but also SPARC and HP PA-RISC?

WinCE was a product MS undertook because NT was not optimized to run on anything smaller than a workstation class desktop. It was an attempt to replace Unix with a new wheel, albeit a wheel that didn’t work as well, had a poor process model, had severe security problems, and was slathered with other MS absurdities such as the Registry. NT was not a general purpose, high quality OS.

Meanwhile, once NeXT got some funding, it was able to not only allow Apple to migrate its Mac platform to new hardware, but also scale to mobiles using yet another processor.

MS did fail in trying to put the (horrible) Win95 interface on mobile devices, but it also failed in developing NT as a purportedly portable OS that couldn’t port or scale effectively.

If you blame MS’ problems on the era of technology in the mid 90s, I’ll point out that NeXT developed its OS in the late 80s.

If you blame MS’ problems on difficulties in the hardware market, I’ll point out that MS maintained a strict monopoly over the PC desktop, and every other vendor (again, NeXT) struggled far harder.

What other excuses can one imagine for explaining away why MS’ technical incompetence was not the company’s fault? That it didn’t have enough money? That it lacked access to talent? That it was restrained by an outside monopolist? That it didn’t have political clout? That it faced a tough economy?

As some point, you have to stop making excuses for the most ridiculous company to ever exist in technology, and just admit that MS is a horrible disease that needs to be eradicated, just like its namesake.

Dan was trying to state that Windows at the lowest level isn’t portable to other machine processors, which is blatantly false, as it was designed from the start to be portable.

That there was a massive failure in the market due to chicken/egg problems of there not being enough machines of the other processor types to make it profitable for Microsoft for the amount of returns, and also not profitable for the number of potential titles copies sold for Windows developers (who generally rely on the huge market base in numbers of x86 users) is part of that problem.

Microsoft did themselves a disservice, I believe, by attempting to do what they did with all their mutations they’ve called “Windows Mobile, CE, xxx” in that it is far enough away from the main Windows code as to being notably different to develop for, though admittedly, some of the things they stripped out (Multiple Document Interface) had no real reason to be kept at all. But, the stripped down Windows mutations are not the question he answered wrong that I am harping on here: it’s the basic premise that he stated Windows itself was not portable, and he claims this all without having any personal experience at a level that would allow him to be remotely qualified to judge this.

And as I pointed out towards the end of the most recent post, if Apple had tried to do an exact duplication of their desktop interface on a 320*240 screen (and perhaps even on the 480*320 screen without multitouch) while it might have looked nicer, and while Apple may keep it locked down so applications don’t make it as crashy as I hear a lot of Windows Mobile systems are, it’d still be a tough GUI to shoehorn into such a small amount of real estate. But, the GUI thing (interface-wise) is not the original question I answered that was brought up about portability, though it does go directly to usability and thus also to marketability: Microsoft failed to insist that portable devices have the capacity to do a WIMP GUI justice, and I expect that’s at least part of the equation for their failure (but certainly not all by a long stretch) when going for the mobile world.

Right, so what exactly was your point again?

UMPC = epic market fail.
Windows anywhere beyond the x86 desktop = ditto.

Some pretty hokey portability on show there, whether it sounds good in the dev libraries or otherwise.

Microsoft proved the OS was very portable: long before x86 went to 64 bit, Windows was 64 bit. It also was deployed on hardware platforms of opposite endianness, which is a major measure of portability. When it comes to large projects, most developers do NOT use an IDE like Visual Studio, and instead use the command line tools: thus, your arguments don’t hold water. I maintain that you claiming that you know Windows (NT-based) was not/is not portable is without merit whatsoever, and I insist that you post your sources from online whereby you get that information, because having taken a careful look at your online resume, I carefully note you’ve not ever programmed in a single systems language: they’ve all been internal languages to content creation products, or some web scripting languages. I also found it entertaining that you claim the University of California claims you’ve got the equivalent of a Masters degree in Computer Science: what you’ve done may accurately depict something similar in Information Technology or Informatics, but is far too weak to qualify as a Masters of Computer Science, or even a Bachelors.

What you’ve done amounts to a lot of content creation and network and machine management: this is technician’s stuff and MIS/IT stuff, outside of whatever web development you’ve done, along with project management in there somewhere, which doesn’t mean squat for knowing how to program. What you need to remember is how many of your readers are qualified, experienced developers that have worked at a low level, who have (like I have at 8-bit, 16-bit, 32-bit and 64-bit with differing byte orders) developed at the OS and driver level, and work on large, complicated applications, and have done so over multiple machine hardware and OS platforms, who actually know what we’re talking about when the term “portability” is bandied about. I find your articles interesting, and I’m definitely on the Apple platform (first on it in the Apple 2 days, now working towards iPhone software development) but stepping outside of your experience is a dangerous thing to do, because you can almost be certain someone that knows what’s what will call you on it, and it detracts greatly from the rest of the article in question.

Oh, as regards to the statement you make about UMPC and usability? I guess that’s all a matter of semantics: is OS X (the kernel) the whole OS, or is that just Darwin? As-is, Darwin is just as “useable” as Linux on the command line: without Aqua, which is made possible to develop for by the Cocoa Frameworks/API, it is just a command line. So, too, you can still technically be running Win32 from a command line, though you’ll obviously be stuck without any GUI apps: you still have the same power of console apps the Linux and Darwin people have. If OS X were also stuck running on a 320*240 screen, its usability would be greatly reduced, too, because that’s just not a sane amount of screen real estate to make a nice GUI with icons, and is similar to the concept of trying to drink an ocean through a straw: you’ll eventually accomplish it (assuming you are immortal), but it’ll really suck! If Apple hadn’t come up with the multitouch touchscreen, it’d be hard to see how they’d do much better with 320*240 resolution, and even at that resolution, it’d be diminished greatly: Apple took the higher practical road with what was currently available (not the highest possible: just the highest practical for all engineering constraints) and saw the weaknesses with being stuck with that (relatively) low resolution, and having to use it with fingers without carrying around a mouse or stylus (yuck, for a phone???) and solved the problems, mostly pretty well, considering that too much larger, they’d have a PDA that’s too big to stick in a pocket, and too much smaller, they’d have something that was “cute” but not suited to the majority of the human population: very much like their simplified product lines that have general purpose machine categories that may not be a perfect fit for everyone, but there’s something that’s close enough to get the majority in that group mostly happy.

“UMPC is entering the middle of the S-curve product cycle where expansion and growth accelerate from the early adopters.”

Not every product has a growth curve. UMPC is a rebranding of the same failed Tablet PC products Microsoft has been pushing since the early 90s, and there has been absolutely no uptake AT ALL.

“Atom is also great for fanless set-top media devices. I believe Apple TV under-clocks the 1GHZ CPU to around 350MHZ and has a fan. With the Atom, it could run at full speed, 1GHZ, and could go fanless.”

That may be the case, but Apple TV is a pretty low-profit, experimental device. Look at Apple’s high volume AirPort business (I haven’t seen the latest figures from NPD, but last year Apple was the top consumer WiFi base station vendor): fanless, compact, does lots of stuff, and based on ARM processors.

Apple clearly doesn’t need Atom. Intel would love to get Apple’s business however.

NT failed to work cross platform because MS did a poor job of developing a cross platform development and deployment system. It also never really supported the vast complications involved with running a fully insecure system across multiple processor architectures. The big problem for Alpha/MIPS/PPC NT users was that MS didn’t release service packs in tandem with x86 versions.

MS now says there “wasn’t demand,” which is BS-speak for “nobody wanted it because it was poorly done.” Perhaps it will take the same historical revisioning to the Zune so that future generations can be assured that it was a great product, but the demand never materialized for it.

As for “Windows XP Embedded,” it’s not mobile at all, its just MS’ “fix it with Windows” solution (ie every thing looks like a nail when you only have a hammer).

Windows-based ATM, register, and Coinstar machines ARE A PC attached to some specialized hardware, not some sophisticated mobile systems running a scaled down version of Windows that can also be used in mobile CE devices. Look at UMPC: its a total joke in usability.

As you point out, Apple’s WiFi mobile platform is purpose-designed. If Apple had tried to scale down an Intel based Mac (with the desktop UI) into a UMPC form factor, it’d do about as well as UMPC, too.

Nobody believes Symbian is the future. It only makes for good PR to recount how many licensees of Symbian there are worldwide, ignoring that the three Symbian platforms are actually different. Again, this would be like roping Mac OS X, Linux, Solaris, AIX, etc together and calling it the Posix Platform to hide the reality that they’re all really separate, competing platforms.