My Zune Problems Are Just About Gone…was I Too Quick To judge
So what we get is the Zune Pad, to the detriment of the Zune 80. Don't get us wrong, it's not terrible (if you want terrible ask us about a Rio, any of them), it's just not great either; it's better than the first Zune and still not as good as the competition. Ultimately it's not possible to avoid fighting with the Zune Pad at least once, which is a tough sell in the iPod age. We should note that you can turn the touch-sensitive features of the Zune Pad off, which would solve some of our problems, but then we'd just be back to a d-pad.
My Zune problems are just about gone…was I too quick to judge
For users there is a regression, things that used to work are no longer working. It doesn't really matter which component is doing something wrong, the end result is that the user can't work.If the reason for this is "we opted to go with this new tool" then the cause for the regression is that new tool (even if the technical fault is that some other component doesn't quite comply with the specs, but in a way that never mattered before.This also points out the difference between the official spec of the API and the practical spec (how much of the official spec you really need to implement to make things work (the de-facto spec). As a practical matter, things that don't need to be implemented 'properly' to make things work probably won't be, so if you later change things in such a way that they break if these previously 'optional' things aren't right, you need to think really hard about what value you gain by requiring these things to be right and what fallback options you can provide (either ignoring the broken info or extracting what value you can get from the broken info) rather then breaking completely.there are quite a few people doing infrastructure work for linux that don't pay attention to this sort of thing, and this causes all sorts of problems for users. the case mentioned above where script headers didn't matter before, but break under systemd is a perfect example of this, but the systemd developers are not the only offenders.If you are really starting from scratch, with no installed base (like Android did), then you can just implement the new way of doing things without worrying about backwards compatibility, but if you are writing something that you hope to get added to an existing system (and this includes writing a new version of an existing system, android ICS, Gnome 3, KDE 4, systemd, etc), then you do have to deal with backwards compatibility and the de-facto standard.Yes, thee are times when you can decide to break the de-facto standard (not being willing to do so under any condition leads to a windows-like mess), but you should be very reluctant to do so. Poettering: systemd for Administrators, Part XII Posted Jan 27, 2012 17:06 UTC (Fri) by jeremiah (guest, #1221) [Link]
That's it. 100% specific purpose of a operating system is to:1. Make it easy to write applications. 2. Make it easy to run applications. You do not have to be aware of this fact to use or judge operating systems either. When your applications do not run on your operating system it makes your operating system worthless. When your applications do not run on your hardware it makes your hardware worthless. Linux desktop had it's chance with Linux netbooks. For quite a few months Linux netbook systems were the top most popular items sold on places like Amazon.com and quite a few. How many of the people that bought those systems actually went back and bought a second Linux system? I am sure that it is in the single digits. Linux fails because:1. It makes it harder to write applications because you have to deal with distribution's BS before you can reach your audience and that there is no standardization.2. It does not run any of the applications people want to run after decades of using Windows XP.You can get all hand-wavy and start talking out the side of your mouth about Microsoft being like McDonalds, people are too stupid for Linux, people just have not tried Linux yet, there is some grand conspiracy from Microsoft holding Linux down, etc etc... all of it is true to some amount, but picking one thing and saying that is _the_ problem really amounts to ignoring reality.And it's not even things like Photoshop or video games. It's having the ability to check your email, run one of the hundreds of thousands of special purpose applications written for this or that corporation, integration into active directory, and a hundred other really PITA and mundane things that Windows does that Linux can't or Linux makes it difficult to do. It's all the hundreds of thousands of schools teaching Microsoft stuff. It's the businesses training people in it, that depend on it, that are happy with it. It is the 'geek squads' and millions of people that make a living working with individuals and businesses to work around Window's limitations.It's not just the low hanging fruit that is killing Linux it's the long, long, long tail.Incidentally this is why you will never see ARM systems displace x86 desktops (even with Windows 8). Even if ARM manufacturers discovers out a significant market segment that x86 folks are not exploiting correctly it will take years to get to the point were the applications are up to speed and by that Intel and friends would have competitive products to eliminate any desire to migrate in the first place. Poettering: systemd for Administrators, Part XII Posted Jan 23, 2012 12:28 UTC (Mon) by mjthayer (guest, #39183) [Link]
This is in stark contrast to the server offerings on Linux, where aside from directory services and management the Linux/FOSS offerings are leading edge in terms of quality, features, and performance.An SDET is a QA engineer; Software Development Engineer in Testing (compared to a regular dev at MS, called an SDE). The folks who write tests, test harnesses, etc. A lot of folks in the Windows OS SDET groups are actually Linux nerds, surprisingly enough (surprised me, at least), and Cygwin and various Linux/UNIX tools play a pretty big role in their work. You're not likely to hear that advertised much, naturally. :)XBL is the Internet services division for XBox (XBox Live) and Games for Windows Live. It's a tiny division compared to the Windows Desktop division. And, btw, not recognizing what XBL is puts you at odds with about 60% of normal consumers, which is just another tiny bit of evidence that Linux folks are completely out of touch with what normal non-nerd computer users care about or do with their PCs and electronics. Meeting computer geeks who aren't also gamers is so "weird" feeling. It's like meeting cinema buffs who don't know what Netflix is. :)Also, hating Microsoft had nothing to do with hating QA. I've met very, very few Microsoft employees who hate the company in any way; the few I have met were mostly just unhappy with how various dev divisions at Microsoft are still using the waterfall dev model (mostly the business software groups; one might note that SCRUM was invented at Microsoft in the Visual Studio division). Microsoft treats its people very well, theres a lot of really cool people who work there, lots of great parties and social events... honestly most people outside the FOSS camp love Microsofts products and services and think they're a great company. Microsoft hate is an Apple/Linux user phenomenon, not a general trend in the rest of the industry.Yeah, Microsoft's marketing and business groups can be pretty slimy, but the engineering groups are quite disconnected from those. FOSS proponents tend to forget that in a company with over 100,000 employees, the public faces and the internal cogs can be and often are quite at odds with each other. Kind of like how some groups at Sony lobby against DRM and other groups lobby for it: big companies can be quite schizophrenic. :)Most people hate test engineering because it's boring and immensely less interesting than developing new technology. Hence the problems in Linux- land where most the volunteer workers are busy rewriting half the desktop stack every few years instead of polishing and fixing bugs and such. I don't blame the developers for doing that - I would never ever accept a QA engineer job because I know damn well I'd hate it with a passion and hence would do a very lazy half-add job at it, and I much prefer doing fun new projects than maintaining boring old projects. I can blame Red Hat, Canonical, and so on for not paying for lots of QA engineers and QA testers, though, since they're trying to get money out of us in exchange for horrible desktop products. (Well, not Red Hat anymore, but I did used to buy their offerings back when they were pretending to be a desktop OS company still.)The really sad bit is how all the Linux desktop folks are echoing the "PC is dead" nonsense ( PC sales are up, not dying off - mobile is conplimentary to the desktop, not a replacement pf it) and trying to chase markets that Linux cannot succeed in. Linux-the-OS-family is NOT a significant force in the mobile market, although Linux-the-kernel is seeing good success thanks to Android. Meego, Tizen, whatever are all flops. Likewise on tablets, the release of Windows 8 is going to end things for Linux there quite quickly, especially if Intel can realize its dream of tablets on x86. Why use Linux (Android or otherwise) when Windows tends to get better battery life, has less bugs, more features, more software, and isn't fractured between 20 different incompatible offerings? One might also note that iOS is still dominating the developer market in the mobile space, due to a combination of better APIs, better dev tools, and a user base that actually buys software.The LInux API story is pretty bad, btw. You have to rely on a bazillion different libraries written by different developers with different naming and interface conventions, often needing to rewrite glue code for each of them to tie into the applications IO, memory, or debugging systems. That's a huge part of why unified platforms like Qt or .NET are so popular; there's no need to learn a ton of different API paradigms just to make a simple MP3 player. Tying that back into this thread, that's one of the great things about systemd - the unified monolithic platform approach is just so much easier to build off of than the scrambled disconnected grab bag UNIX Philosophy approach the traditional init/service/logging/session systems were all cobbled together with. OS X and iOS have a very solid unified API and Windows is making decent progress at modernizing an unifying their APIs (somebody still needs to take tr core C Win32 API out back and shoot it, though; ugh is that thing horrific). Poettering: systemd for Administrators, Part XII Posted Jan 23, 2012 0:11 UTC (Mon) by jimparis (guest, #38647) [Link]