Either way, we all hold GPL (version and other offers?) on sunjava (todo trademark as constant definition in public domain?) and hope for working together, but if my programs running on GPL'ed derivativeprogram of sunjava, which we all branched just before Oracle bought java and various other things which do not occur in GPL'ed java (by direct permission of GPL for everyone to use modify and distribute and not allow distribution without patent rights recursively as staying GPL, and since sun did distribute many times before a later version was bought, and because Oracle gave everyone GPL on some later versions of OpenJDK (which I hope continues, as Oracle is free to never make another OpenJDK version, but at that moment OpenJDK would be forked and not called java or openjdk or sunjava, some new name I'd guess. I'm confident that wont happen at least not so directly, but if none of our phones run these great (any)java programs, only through some system of misunderstood paths through tech outside the phones, then in what sense is the phone a computer (turing complete by its surface controls or handheld controls)? Google risks an opensource fork, not likely any time soon but if backward compatibility is not reached many people lose huge value in our history of (any)java development when not able to instantly compile on the computer itself. I realize its not meant to be a safe system from inside in the normal sense of programmers restart servers instead of tracking down the wrong bit somewhere deep inside, and I'd like to help put a stop to that. If rules could be defined in math of what the (any)java program can do, and extending those rules down to the bits (or does wave timing affect it? Give us a really good clock, and an array of signal processing stuff too, I'd use it especially if java hooked in, but we cant have this delay stuff with the speakers and screen and... not really the subject here, just somebody's branch of (any)java we thought would continue to free us from platform dependence. Nobody downloads my desktop sunjava or java progs on phones because the executable jar file (containing its own source code and classes) does not come with an executable function to run such file.
Now that I think of it, if Google does not provide an execute function in Android, for any turing complete logic, then legally I cant call it any kind of "java" which is turingcomplete, unless there is such a function but only reaching hooks into installed progs such as I saw some java runner prog in app store, but it didnt work well. Headless I could accept if it reached progs like at a "localhost" port, if there is such a thing, or even better if AndroidJava wrapped its local details of screen and addressing inside the sunjava math definition (TODO trademark for public domain to refer to that unchangeable concept), and from that stronger point of negotiation from GPL java programmers and people who use our tools, I suggest many of us would ask for Google and Oracle and any others involved to make our old programs work like they were on a desktop that just had a smaller screen (Maybe Google would buy any relevant patents and give to GPL or public domain? Or whatever solution maybe could be reached.
Should we all seek trademark for sunjava unchangeable math definition in (or that may be a problem if the definition was not pure math) trademark sunjava to public domain or GPL. Until then I informally will say first is this or that a sunjava program, and vaguely imply some versions changed after that. Also the word alljava, maybe we should trademark, to refer to both of these and any other variations, such as Google wants their own name of their java-like system and there was a patent conflict and legally some (any)java forked.
Also, Sun Machines, which existed long ago, ran some variation of (or equal) to earlier version of sunjava (or is there evidence such machines contained any code not duplicated in any sunjava or at least OpenJDK GPL (thanks for the irrevokable patent (if any apply) permission to use modify and distribute recursively under GPL, many may put it to good use some day, like Google I heard considered but had to have some part of Oracle proprietary patented code, which I'm not sure which parts, anything GPL'ed and distributed by its patent owner is ordinary GPL, and no further limiting conditions does GPL allow, only extra permissions multilicensing) Just so we all know the most basic thing of GPLv2 at least, if distributed its either not obeying the GPL2 (and therefore no GPL2 prog has been distributed since it was never GPL2 to begin with if such patent preceded it and was not granted, or if was granted after such program was publicly widely known then, as its supposed to work, patents are not granted for things publicly known). Sometimes things are confusing to keep straight. I'm not claiming any patent violation. I am claiming everyone in the world can fork OpenJDK version today under plain GPL which says no extra conditions can be added, as long as it stays GPL and its called something else.
Can we have a platform independent language that works for part of our global infrastructure, not low like hardware routers, but there are network expansions and we're not using them much. The network isnt even turing complete. We dont know what it is, but cracks are so obvious and many that its thought normal. Is this how we want to have simulated minds roaming a cracked dangerous unknown tech thats not derived from any theory of computing logic, such as turing machines, except as it was designed, but after designing some other people got ahold of it and had their own problems to fix in it. Its redundant for safety, but its nonexistant in the space between computers. I cant reach a cloud without passing a non-turing-complete layer. Most are so scared of touching it they say viruses or hackers may attack. Let them attack a computer proven by basic math and expanded within those rules. What is a virus but a number that some tech reacts to? If we define the reactions precisely, as they do in antivirus research they have viruses all over their hard drive, not running, because a file is just a number sitting there, unless a file is part of a buggy system thats secretly running in some driver level and things on occasion go wrong almost anywhere. Theres too much code, and not enough Occams Razor. I use Windows7 often, not because its the best, but because of incompatibility with other things across the turing-incomplete internet space, and whatever is in the computers past them. They fear their stuff will not behave predictably in Linux because its a different computer, and Windows has some parts shared across the network or in files or something. Its getting better, but isnt it insane that you cant simulate a pocket calculator almost anywhere without admitting the possibility that you dont know how its internals work. If we had such a math layer at the core, I would often make the legal statement that I guarantee against certain behaviors, sometimes called bugs, if I could prove it, as long as the network wires are as reliable as the chip insides which are huge numbers of cycles, billions per second, for at least days? or years without a single bit in the chip wrong unless its fixed within some cycles, or unless we didnt see it because software just runs it again after finding a bad hashcode? I dont know, but I'm not going to run a minimalist computing grid for AI games and science, to be explored eventually, derived from anything that is undefined its chance of differing from what the math says, and considering how many mutations it could recover from statistically. They censor the Internet in part because if it were used as a computer all hell would break loose, until it may become unfixable or cost ever more to maintain the patches upon patches. I trust sunjava and OpenJDK and some basics of math, enough to run that if defined in pure math excluding any libraries or native hooks. An opcode is math. A compiler is math of unicode bits to opcode bits. Math never breaks or changes at all. Opensource is my way toward that, and hopefully many others will try eventually. Can we stop protecting the ancient cracks in the infrastructure and admit nothing can safely be run without a full emulation on top of it or replacing the hardware or hopefully both? Anyjava is emulated, virtual machine, at least the opcodes and class datastructs and size limits. A computer is a computer nomatter how many complexities its grown, and can still simulate simple things as math if it is itself still math.
