Hi Raffael,<div><br></div><div>I am having a second thought about side-effects</div><div><br></div><div>If you record all side-effects it would be possible to revert back the side-effects.</div><div><br></div><div>I am looking at a sequential program as a sequence of actions; </div>
<div><br></div><div>Any sequence of revertable actions is revertable.</div><div><br></div><div>P1 = a1 a2 ... an</div><div><br></div><div>(inverse P1) = (inverse an) (inverse an-1) ... (inverse a1)</div><div><br></div><div>
If for each action you record the context in which it is executed</div><div>then you can deterministically revert the whole program.</div><div><br></div><div>But it is not realistic to record the context of all actions, thus</div>
<div>I just expect that the (inverse P1) will return all possibilities (as </div><div>in the square/sqrt example).</div><div><br></div><div>Kind regards</div><div>Taoufik</div>