This is crazy. Why should the app have to know or care about the DPI of the screen. GUI toolkits should work based on pixel-independent units, and use layout managers to sanely lay out the widgets. I guess I've never understood the HiDPI hacks that everyone is doing, even in Gnome and KDE. If I have a dialog that uses 12 point font, it should be 12 point regardless of DPI and everything in the dialog box should be the same whether the screen is 72 dpi or 300 dpi.

I've always thought the Win32 pixel-based, fixed-position dialog boxes were a bad idea. OS X also fails badly in this area. I tried to get OS X to drive a TV, which sucked horribly because you just can't scale up the display at all. Windows may do this hap-hazardly, but at least you can make things bigger for the TV screen.

Sounds like you didn't actually read the article, or understand what it said. Of course the OS is not for you. You're either not a programmer, or else you're not a programmer who's been exposed to some of the really cool ideas of the past like Smalltalk and its IDE, or LISP machines. Perhaps you would not think of anything to learn from them either.

From a programmer's point of view a shell/file explorer that's integrated the way TempleOS's is is way cool, even if it's not useful for normal users. Sure C is awkward as a shell expression language. But it's still a super cool idea (even if impractical for most computer use today).

No, his hypertext format is definitely not a reinvention of docx. Reread the article. It's more like a more flexible version of the old MS .doc format actually, which was just a serialization of in-memory OLE structures, which was part of why doc files have always been so hard to read since the OLE objects themselves were always changing as MS worked on Office.

Actually, as it presently stands, his OS is very secure indeed. It's literally impervious to remote exploits, and you'll never run any insecure software on it because odds are if you ran it its only his software or your own. Sure if it ever became network-oriented it would be a huge problem. That's not the point, though.

You're absolutely right that TempleOS will ever find its way into any sort of mainstream use, obviously. But that's not the point of the article. The point is that this is a monumental work for one person, and that there are some really cool ideas that maybe we could learn from. Such as the blurring of a text-mode shell (a la bash) and something more graphical. The idea of embedding diagrams and documentation hyperlinks in source code is genius (and automatic ones at that), as are his annotations for struct members. Having a living coding environment is also very good. I would guess that if an environment like Visual Studio implemented this (kind of like Quick Basic 4.5 had back in the day; it compiled on the fly), programming in C or C++ would be almost as fast as Python since you could test and tweak individual functions as you wrote them. They would become live and callable as soon as you syntactically completed them.

So I think you missed out on a few things that we really can learn from TempleOS. If TempleOS had a few more features and dropped the VGA-only interface, it could be a very fun learning platform for programming and system concepts. In fact I think TempleOS could itself be made into a standalone, self-contained application for teaching programming. Would be similar to Squeak or Logo. Limited in scope but a good teaching tool.

Seems like congress likes to act all indignant (certain congresscritters) and demand that executive branch agencies answer their questions and defend civil liberties, but in the end nothing ever is done. Even Nancy Pelosi, after being temporarily upset that she was the target of some surveillance (if I recall correctly) fully supports the NSA and their illegal information gathering. I am left to conclude that congress just puts on a good show for the masses while the media is focused, and then when things move on they go back to doing what they were doing before. Both parties.

First I have to say I'm pretty disappointed in Slashdot commenters today. I was hoping to have a nice discussion of some very cool ideas, reminiscences of work with smalltalk or LISP, but sadly most people are just talking about the man and his illness. Too bad. Few of us, inspired by God or not, could have coded something as complete as this is. Temple OS's concepts really make me think. When I was a teenager I dreamed of an OS that would blur the line between code, programs, and data. In some ways I envisioned a system just like he designed. Where code is live as you write it. Instead of programs, you'd have just data, and code to operate on the date. Just like his idea of embedding pictures in anything I thought why not treat all forms of data that way. Instead of a word processor, you just have frameworks that operate on text objects already native to the system.

Some of these ideas are similar to, I believe it's called, squeak, which is a smalltalk environment that is completely live and modifiable.

In some ways TempleOS (if it could be adapted and modernized) could be the learning tool to really get young hackers excited about building things on their computers. Having a live compiler jit'ing code on the fly as I type it sounds very cool, especially since it accesses the whole system, and becomes a part of the system as you code. And his programming language looks very interesting.

As cool as it is, I think it would never fly in the real world because almost all people don't want computers to work that way. They don't want to create abstractly with them, but just use them as appliances to do some task. Which is sad but understandable.

While I feel the outrage over this move is probably overblown, it does vindicate the fairly extreme positions in regards to free software held by Richard Stallman. Basically the watered down idea of free software, called "open source", has taken off and really win the world over. Even Microsoft is embracing open source. Everyone sees the benefits. The problem is that they see that it can benefit their existing proprietary models quite well. So for example Microsoft, while being more open to open source and interoperability is as proprietary and closed as ever before.

Intel too has embraced open source and Linux, but the philosophy of free software is only embraced to the extent that it will help their business.

In the end then open source software has both won and lost. It hasn't changed the corporate attitudes like some of the earlier visionaries wanted. I fear we're all the losers here. This move by Intel confirms that to me. In this modern post stallman world, open source is mostly a way to placate the masses. And it works well. And is quite profitable.

Yes the idea that a mini physical keyboard on a phone device could be patented is ridiculous. But you can rest easy knowing that Blackberry is in its death throws anyway. Unfortunately that means its patents will be soon up for sale and will be bought up by other greedy companies who will continue to "leverage" them.

No it's not a clear violation of the GPL. Go read the GPL text itself. The GPL refers to the binary itself, not the installer. The binary itself, unmodified (or modified with source code), with a clear link to the source can be distributed easily, wrapped in whatever proprietary installer you want. The GPL is only transmitted through direct linkage. What SF did was certainly unethical, but it wasn't illegal.

Not sure where you're checking. ARM has been supported as a target for some time now, and as a host. Of course we aren't talking about the ARM target; we're talking about the x86 target on an ARM host. And it will definitely compile and run on an ARM system. Both full system emulation (a virtual machine) and user-mode emulation, though it's not really that fast yet. The latter mode is closer to the software described in the article. Years ago I used the QEMU x86 user mode system on my PowerPC to run a few x86 binary-only linux programs and even browser plugins (Adobe PDF reader, Adobe flash, and wine). User mode emulation often appears faster because only the program itself is running through the emulator. All calls to the kernel are thunked through to the real kernel. So you get native I/O speed, for example.

3.11 was Windows for workgroups, which actually was very good, probably better than 3.1. More stable anyway. Though 3.1 was way more stable than 3.0. No more UAEs. apps could actually crash without crashing the whole OS, if I recall correctly.

I remember my neighbor running a brand new installation of Windows 3.0 a 386. The only native app was, if I recall, was Word, and it was pretty crappy back then. Windows 3.0 would UAE at the drop of a hat and hang completely. It wasn't until 3.11 that Windows became actually usable, though the architecture (cooperative multitasking) was so bad that I'm surprised any programmers stuck with the system long enough to develop any apps. I guess the promise of a stable GUI API and a standardized hardware abstraction layer (printers, etc) was enough. And Windows 3.11 introduced truetype fonts, which were pretty amazing compared to what we had before that time in Windows and MacOS.

At college we used to say that only a fool would have win at the end of his autoexec.bat. The rest of us would run windows when we needed it, from the DOS prompt as God intended. I had a friend who ran OS/2 2.1 with a text-mode shell that multitasked MS-DOS apps, and that was far more useful at the time than Windows was, since all our apps were DOS apps back then.

Palemoon is potentially such a fork. Meets my needs. Perhaps it will meet yours too.

This is nothing like the America space shuttle. Probably closer to the secret US Airforce mini-shuttle. It's a small vehicle (1.5 tonnes) that goes up on a rocket and then re-enters and lands (eventually) like an aircraft. It won't carry people as it's much too small.

Meh. Android 4.4 broke SD cards completely. My phone runs android 4.2, and it works, so I don't want to mess with it. I think that's how a lot of people are, despite security bug risks. I like Android in general but there's a lot I don't like. One of them is that updates are dependent on the vendor. The other is the murky world of semi-legal firmware distributions that rely on crappy forums for developer interaction with no public version control, no nice spots for download. Who knows what's in Joe's firmware posted on some random forum post? Leaves a bad taste in my mouth the way most android development is done.

Hardly. Windows Me (and Win 9x), as bad as it was, was lightyears ahead of classic MacOS from an architecture pov. It was fully preemptive, multitasking. And it was to a large degree really 32-bit, though it was bootstrapped from a 16-bit environment, and some of the drivers appear to have been 16-bit. But it did run in protected mode. Wasn't nearly as good as Windows NT of course, which was also pretty darn good, based on the venerable VMS operating system's architecture.

As for Vista, under the hood it was similar to XP and Windows 7. There were horrible UI decisions (the UAC mainly), but the core was solid and stable, and fast. Windows 8 was just fine too. It was just a UI mess.

OS X, while not particularly speedy under the hood, is solid and stable also. Of late their UI has started to suck more and more though.

In Australia and western Canada, neonic-coated seads are typically placed in the ground via a gravity-fed metering system (box drill), or via an air drill that blows the seed into the ground behind shanks that open the soil. So dust particles laden with neonics get buried in the soil where bees won't be exposed directly to them. In the midwest US and eastern Canada, where the crops are predominantly things like soybeans or corn, they use vacuum planters which suck the seeds from storage one at time and drop them into the ground. Unfortunately the vacuum planters blow a lot of dust from the seeds into the air. So neonic-laden particles get blown everywhere and we know they affect bees and any other insect. So it could very well be that widespread use of vacuum planters is a part of the problem. Unfortunately air drills don't work very well for row crops that do best with rows of singulated seeds.

The Alberta Bee Keepers Commission refuses to back any attempt to completely ban neonic use in Canada as it would decimate their industry. Fewer crops means fewer bees are required by farmers.

The reason neonics are used is that when the plant is young, the neonics are taken up through the plant and make the plant toxic to pests that would eat the little leaves, killing the plant. On one of my dry bean fields last year was seeded without neonic seed treatment, and we did see some yield reduction from pests eating the plants at an early stage, including from works eating the shoots underground. If there's a chance neonics can be used safely, then for sure they are a huge benefit.

There is the other issue of neonics present in the pollen, leading to bees getting a bit of a buzz. It's not clear to me how much neonic there is in the flower at that late stage of the plant's growth, or what the consequences of that are. Bees around here are heavily used to pollinate hybrid canola, all of which was treated with neonics. So it's really hard to say what the consequences are.

It's true we can control the use of pesticides, and we should and do. This doesn't have to mean an outright ban. A complete ban would mean the return to more toxic insecticides being sprayed at more regular intervals on a crop, which none of us wants.