Inconsistency when using previously defined variables as arguments to functions

ev really is that evil -- take a look at my old note in
http://www.math.utexas.edu/pipermail/maxima/2007/006737.html for some
examples (certainly not exhaustive!).

It is easy enough to create a pseudo-function that will quote its argument
and 'protect' it from simplification:

:lisp (defprop $nosimp (lambda(x) (list '(mlabox simp) (cadr x) '$nosimp))
mfexpr*)

ex: nosimp(2+3);

                nosimp"
                "2 + 3"
                """""""

rembox(ex,'nosimp);

            5

block([simp:false],print(rembox(%o44,'nosimp)));

2 + 3        <<< printed result from within the block

            5    <<< gets simplified when returned from block


On Mon, Jan 31, 2011 at 18:45, thomas <thomas at geogebra.org> wrote:

> On 02/01/2011 12:30 AM, Stavros Macrakis wrote:
>
>> I don't know what a "Hold" statement is.
>>
>> To simplify an expression that was created with simp:false, you can use
>> simply
>>
>>      block([simp:true], expr )
>>
>> This avoids all the perverse semantics of 'ev'.
>>
>> If the expression has been simplified (or partially simplified) under
>> different conditions than the current conditions (e.g. different
>> assumptions, different global flags, etc.), you can do:
>>
>>         block([simp:true], expand(expr,0,0) )
>>
>> which forces Maxima to rescan the whole expression, including parts that
>> are already tagged as simplified.
>>
>>            -s
>>
>
> Hi!
> Sorry for not explaining what I meant with "hold": I meant a command
> available in other CAS that is basically a way to enter an expression
> without evaluating or simplifying it. Since to my knowledge maxima doesn't
> have any such thing, so we emulate it by using simp:false as default and use
> ev(....., simp) every time we really do want to have somethign
> evaluated/simplified.
>
> Is "ev" (and using simp:false) really that evil? (Like I said, I've yet to
> encounter any "perverse" semantics of it, so far it worked pretty well)?
> (In any case, thank you for your suggestions and the tip with 'block' :)
> )
>
> Cheers
>
> Thomas
>