this is something of my own meandering. Although other people have probably figured it out. I'm sure builder boy and weregoose have figured out this trick and possibly used it once or twice. It has limited application outside of complicated implicit stacks and even then it's far easier to use sub routines. The applications for which it gains you anything are not easy to determine at all. I used it once (other then messing around with it to confuse people) in a program that shaded an arbitrary area.
because For( behaves differently from While when exiting you can do some really crazy things In practice this trick is extremely hard to pull off but you can do things like inline recursion and complex stack insanity rather quickly without the use of extra programs. Normally doing this is like playing with fire but it's occasionally useful.
I like the 89 It fixes a lot of things that were wrong with the TI basic language. For example something I found yesterday matâ–¶list() and listâ–¶mat() functionality was fixed to A) make sense and B) be useful. Although I don't like having a difference between end statements for if while and for I liked the single end statement It allow for some useful exploits of the interpreter. It is much different from 84 basic though
Edit: I think you should disregard this cause I think I'm getting my calculators confused
scratch that it's a kernel and api's meaning graphic display libraries (we should try to find though for Ndless and complicated drivers for delayed write i/o, network protocols, USB and all that) We have really got to find thous libraries for ndless because they will open ended for development, fast as a bat out of hell and already there
Oh yah I've neglected to mention that nucleus is not an OS like TI OS it's more like the lynx kernel except with out the command line the only way to interface to nucleus directly is through compiled binary programs. So TI did write everything we see and all the math functions. just not the stuff that makes everything work. nucleus is a kernel only to the best of my knowledge.
agreed I haven't really used Ndless because my programs exploit a lot of the changes in behavior since OS 1.4 in order to work. little things that make me able to do more with less in N-spire basic.
Altough, my main worry with the future of the TI-Nspire is that TI won't start releasing hardware revisions of their machine that breaks compatibility with old programs like what happened to TI-Boy SE on the 84+
I say port the computer emulator for the GBA for pc by recompiling it for the N-spire
I did not know TI used someone else's OS (Nucleus) for the TI-Nspire, though. I was sure it was their own . Not only they try to hinder us with DMCA notices and can't fix their old calc OSes bugs, but now they can't even write their own OS for the Nspire
nope they not only use someone else's OS they also use: an open source project for their encryption (openSSL) another open source project for their compression of tns files (zlib) a third for xml parsing (eXpat) and their gui concepts are based on Fathom Dynamic Data. (at least that is the impression i get)
goto settings->handheld status->about and it tells you all this stuff
I personally think its a gift for them to be using libraries we can see the source code of because it can give us a better understanding of how the calculator works and how to crack it. because we know what has to be in the OS so we don't have to guess how they did things.
In the screenie from the post, the "memory" shown on the Nspire is the addition of RAM+ROM (I have no idea why they added them together, but never mind).
I'm not sure that you are looking at the memory hierarchy the right way although this seems to make sense because of how TI has done there previous calculators but none the less the N-spire is a huge departure from previous calculators as TI advertises it's more like a computer and that goes for both inside and out. The N-spire believe it or not has a boot strap,a kernel, hardware abstraction, and the works. In fact I'm not even sure that it is single core any more (it may have multiple processors) but I digress. I do not have time to post a full description of how the N-spire works from a hard where stand point but I will say that there is 32MB of flash memory (used like a hard disk) and you are seeing that minus the partitioning and file system overhead. (likely FAT) To understand fully the gigantic departure from TI's norms you can compare a cell phone to a desktop computer. the N-spire is the OS it runs is called Nucleus RTOS. heres the wikipedia http://en.wikipedia.org/wiki/Nucleus_RTOS and here's a useful page of features http://www.mentor.com/products/embedded_software/nucleus_rtos/
My point was that we are messing with a difference of <.001 seconds when the only responsiveness that is important is that of updating the screen in <.01666 of a second so though builder boy's programs wall time is low wall time is not always the most important. And actually after a review of the data that builder boy has supplied and a few of my own test (inaccurate as they may be) I think builder boy's got me beat on this one his code runs fast enough that all of this is irrelevant. I just now have been able to get the code over to my calculator and it runs faster than I think mine could. It just goes to show me and everyone else the simplest solution is often the best. Although I think I have a couple of small improvements on his code but I won't post them because I'm not at all sure they are actually any faster.
Show Posts
. Not only they try to hinder us with DMCA notices and can't fix their old calc OSes bugs, but now they can't even write their own OS for the Nspire