> > Thanks for the great detail. So, if I am reading between the
> > lines, I would tell the teachers in my department (who are users,
> > but not technical users) of the NSpire that:
> >
> > 1) Students can load a program into the Nspire which,
> > 2) Even when they go into PTT mode after arriving for a test,
> > 3) allows them to access documents.
> >
> Correct if students know about ndless, know about this hack and also
> find a way to actually download the files somewhere (as said before,
> they get censored rapidly), and know how to enable it
Or can make such files by themselves and share them across a group of
students. The programs that we've seen are simple in concept and
implementation, showing that tampering with the PTT mode is easy.
> > To avoid this we would have to have them essentially reload the OS
> > when they arrive.
> >
> Correct.
> PTT hacks are generally done (well, the ones I've been able to look
> at, since they diseappear quite fast) effective by putting them into
> the (Ndless') "startup" folder so that, as Lionel said, they can copy
> the actual documents to the other virtual PTT root. Before going in
> PTT, one could check of there is a PTT-hack in such a folder. But
> maybe there have been some other hacks that I don't know of that act
> differently.
>
>
> > ---------
> > Also, how would a teacher (easily?) identify a calculator with
> > this ability or in this mode, or is that not obvious?
> Pretty much by going to the document browser and see if any files
> are here ?
Yeah, that's probably the easiest way to detect most of the programs
that we've seen so far. But there could be programs aiming at being more
stealthy. For instance, patching out the tests + branches which disable
some math functionality while in PTT mode wouldn't be detectable through
the document browser.
Trustworthiness of a given calculator model in a given standardized test
is a problem, which will only get worse with the increasingly complex
pieces of hardware and software that calculators are. TI, as the market
leader in many countries, is more impacted than the other manufacturers,
but all manufacturers are hit by the problem.
One way to solve the problem would be to make teaching and exams closer
to the real world, simplify the rules, and save everyone money, by:
* nuking silly artificial restrictions such as "no CAS", "no trig
functions", etc. The French school system is far from being perfect, but
that's something we get right, IMO.
* letting users use auxiliary programs and documents, of course. We
can do that in the main standardized test in France.
* as a consequence, eliminating the need for protection modes, for
separated non-CAS and CAS OS and models, for reflashing calculators in
the exam room, etc.
In the real world, when solving real-world problems, getting the result
right, and providing a proper explanation, usually requires knowing the
basics (the theory behind the result) anyway. Calculators are great
tools, but one needs to check the orders of magnitude and whether a
result / solution feels good, on various criteria (when attempting to do
things correctly, at least).
If the exam's questions are done the right way, calculators help those
who know their stuff, and don't help (or even hamper, by slowing them
down) those who don't.
Alternatively, it's possible to persist in error, in exams disconnected
from real-world problem solving, in failed attempts to restrict
functionality, and nevertheless have to resort to high-cost and
heavy-handed measures such as erasing the entire contents of the NAND
Flash memory through the maintenance menu, and reflash an OS upgrade
tailored to the exam's needs (forbidden functionality merely compiled
out), right before the exam is taken, just in case.
I strongly doubt that the undermined models can be banned outright,
especially if they're in widespread use, because it would severely
distort the exam's fairness: an allowed calculator model is used for
most of the school year, so students get used to it; but several weeks
before the exam, it becomes forbidden ???
Note that this in this direction as well, special testing modes and
suchlike half-measures are eliminated as well, by treating anything as
untrustworthy (like the Nspire series is, even if TI keeps trying to
achieve this aim, nLaunch shows that they have royally failed) and
therefore simply reflashing the calculators with a tailor-made OS.
This will bear high, recurrent costs, obviously.
Ramblings over :)
Lionel.