Please wait...
interesting read. Intermediate language so you can efficiently generate machine code portably.
Till the last sentence, who's the sponsor of this; The company known for the portability of its OS-Products over bazillions of processor & hardware architectures:
Microsoft.
That having been said, what's the difference between this and any other bytecode?
These sound like features that are already available in the Java virtual machine or in the .Net allegedly-portable runtime.
Java is a pretty high-level bytecode, not suitable as a "common intermediate langauge" for compilers. LLVM is better suited for that role.
.NET, I think, is somewhere in the middle of those two. I believe it assumes GC? That is not true of LLVM.
Seafile - a cloud storage thing.
Done in python + django; with a pretty slink json api:
https://github.com/haiwen/seafile/wiki/Seafile-web-API
(in comparison to owncloud java monsters)
Just as many careers were launched during the boom years of the 1990's when quite a lot of people just looked through the want ads to see what paid well, and went ahead and learned to code because it was in demand at the time. (Many of these people left the industry after the crash -- and good riddance to them.)
But back in the 8-bit days, anyone who had a computer did learn a bit of BASIC sooner or later.
Young folks who take up programming don't seem to grasp certain things about the computer that we seem to understand.
At least, it seems like that from the few I've met.
I don't know if it's because they never had to suffer through the command line interface, or just farted around with the old technology enough to understand how truly stupid a computer is. Just something... missing.
I find they sometimes have unrealistic expectations from the machine. Which, perversely, is useful when creating a GUI, since you just program to those expectations, but not as good if you're trying to understand why the machine isn't doing what you want.
maybe the arduino or the pi is a good startingpoint for hacking about these days...
its hard to find computers which you can "understand" in a lifetime, but they still exist.
why to shell out a fortune for a proprietary embedded development environment:
http://linuxgizmos.com/why-use-commercial-embedded-linux-development-tools/
"For example, when a suspected tool bug is spotted, the commercial tool product’s own engineering staff is generally prepared and ready to diagnose and fix the defects."
seems like a funny joke... to read that from a company, where my ex employer in the 90'ies would insist on developing an own IC-CAD software, because of mentor graphics wouldn't even listen to them for the million dollar they paid them in annual royalities...
its hard to find computers which you can "understand" in a lifetime,
but they still exist.
There's something to be said about that. I understood my Commodore 64, stem to stern, every IC, every register, every bit of how the built in ROM's drove the custom IC's. It can't be done on modern computers; everyone involved in the design of one hands off to a black box on the other side of the bus, maybe knowing the interface well but not knowing what it really does inside.
But it's worth it because we can spend our Saturdays watching funny cat videos.
http://www.youtube.com/watch?v=oQ5v9Qo6XzA
Speaking of cat videos - Nyan Cat looks best on the C64 version, but I am partial to the spectrum version with the more vibrant colours :-)
oh, demoscene all but dead - squeezing the most out of a c64 is still a challenge.
on modern pc the resulting size of the binary is the challenge- what can you get done on a modern pc with 1k code?
on modern pc the resulting size of the binary is the challenge- what
can you get done on a modern pc with 1k code?
Lots of truth there. "Back in the day" we were limited to a maximum of 128K in executable code size (after you pushed it through the compiler).
Citadel:K2NE managed to get it all done, including landline-style Citadel networking, *and* the Citadel:K2NE mail> networking with its internal "optimum path" routing code, in 126 to 127k. It was strictly single user and strictly dialup. I was working on making it multi-user back in '92 or so and then the internet hit big time in '93 and there was simply no looking back. For those that might be interested, the method I was employing involved room "locking." Meaning, if a user was "in" a room, then another user's session would treat that room as non-existent until the first user left that room. I am very happy that I was able to use the Internet as a built-in excuse to stop trying to make MSDOS do something it was absolutely not capable of doing - multiuser/multitasking/multithreading. And the thought of doing that in 128k of executable code *still* gives me headaches at the memory... ;)