Hey crit,

You're totally right when u say that the self-test should not alter the flash rom.
The thing is : it actually doesn't (and has never did) !
It is a mistake from the beginning as penguin already mentioned it on this thread (page 1) but also on tibank (He posted a comment on polypatch84 news page.).
Actually, penguin's fix removes the groups deletion from ptt (not only the groups in fact, my mistake there).

If this can help, i made some more accurate tests to see exactly which patch does what :

BPM

      PTT - [<]+[>]+[ON] - 2.40>2.55
X X locks the execution/edition of all prgms/apps
 XX deletes all variables from vat, archived or not, except for prgms/apps
   X creates empty predefined lists (L1-L6) in vat if not already existing
   X sets some system settings to the values inputed by the user

     TEACHER RESET - [8]+[5|2]+[ON] - 2.41>2.55
X   deletes all variables from vat, archived or not, except for singapore language apps if [2] used
X   creates empty predefined lists (L1-L6) in vat

     TESTGUARD2 - probably a ti program which has the same effect as ptt
X   locks the execution/edition of all prgms/apps

B = patched by brandon - "patchptt" - 2.40>2.53
P = patched by thepenguin77 - "nodel" - 2.43>2.53
M = patched by mapar - "ptterase" - 2.43>2.53

Note about mapar's patch :
It's the only one that fully cancels all ptt effects (including those nasty functions disabling on 2.53) but it has a little drawback (for students).
At the 1st ptt screen, the input of system settings works but has no effect on the actual values (part of the patch).
The problem is that, on the 2nd screen (ptt complete), the original ptt code displays the actual ones.
So, if the teacher sets different values than the actual ones, he may notice that his input has been ignored (if he takes enough time to read the 2nd screen).

Now, it's up to you to see what is better for polypatch84 =]