Posts by Lionel Debroux 3 publicly visible posts • joined 23 Sep 2009
Above, my "there exist decent third-party OS which are based on TI's copyrighted OS at all (even if they aim at be compatible)." should read "there exist decent third-party OS which are *NOT* based on TI's copyrighted OS at all (even if they aim at be compatible).", of course.
There's also "Punix OS" ( http://sourceforge.net/projects/punix/ ), but that one doesn't aim at being (mostly) compatible with TI's OS.
Previewing posts isn't enough for me, it seems :P
Even before we factored the signing keys, the seven Flash-able models of the TI-Z80 and TI-68k series were, for years, programmable way more than enough for the purpose of adding loads of cheat sheets and cheat programs and using them in standardized tests.
(that is, unless the RAM + Flash of all calculators were forcefully reset before the test, which is far from being done 100% of the time, AFAIK).
All that was needed is the use of assembly programs and FlashApps, all of which can take advantage of officially-documented OS-level hooks TI put in their OS. There's been a while since someone made a program that catches, upon the pressing of ENTER, the commands entered by the user, and returns garbage answers to the user (e.g 2 + 2 [ENTER] gives 123).
Therefore, if cheating is /truly/ a concern of the standardized tests' regulation bodies, they should *never* have allowed programmable calculators in the first place (and TI calculators even less than most other, less-moddable models). They'd be full of fail if, in the light of the factoring of TI's public keys (and therefore, the deduction of TI's private keys), they decided to ban TI calculators, and only TI calculators.
All nine keys of TI-68k calculators, and all seven keys of TI-Z80 calculators (one key is common to both families), can be obtained as *clear text* (let me re-state that: we *didn't* break into any encrypted content), from e.g. OS upgrades, boot code dumps and developer certificates. The factors of one of the keys were released by TI themselves as part of the TI-Z80 SDK.
Developer certificates were formerly sent by TI to people who wanted to program FlashApps and test them on their own real calculators. TI's support going silent for a long while, both for developer certificate and FlashApp signing requests, thereby preventing progress, was one of the main reasons behind the factoring of the 4 TI-68k and 3 TI-Z80 FlashApp keys.
@Robert Hill: the algorithms in TI's OS are not necessarily good, and the toolchain they use to compile them sucks a lot. Plain and simple. Reimplementing various core OS functions can yield massive performance improvements: the "next_expression_index" routine of the TI-68k OS (get address of the next expression in the RPN stack of data) was reimplemented to be 3 to 4 times faster; the "DrawStr" routine (guess what, draw a string onto a plane) can be made *an order of magnitude* faster for the common case. Et caetera.
@Ed Blackshaw: +1, except that there exist decent third-party OS which are based on TI's copyrighted OS at all (even if they aim at be compatible). For TI-68k calculators, there's PedroM: http://www.ticalc.org/pub/89/os/ , featured on Slashdot years ago. PedroM 0.82, sporting a vastly expanded CAS (which handles arbitrary-precision arithmetic, unlike the builtin CAS - but the builtin CAS also can do things that PedroM cannot), should be released soon. The author posted screenshots at http://www.yaronet.com/posts.php?sl=&s=122534&p=1&h=13#13 .
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
