CBFalconer said:
> Richard Heathfield wrote:
>>
<snip>
>>
>> Since my home unaccountably lacks a resident Pascal expert, I
>> would appreciate feedback on whether these are good answers.
>>
>> All eleven of them may be found here:
>>
>>
http://www.cpax.org.uk/prg/****table/pascal/pwp/exercises/ch03/index.php
>
> ex08 is a mistake. The standard procedure write(real : field :
> precision) covers the complete mechanism. You want precision = 2
> here.
Um, this "mistake" is the spec, Chuck. I understand what you're saying,
but I think this is one of those silly exercises along the same lines
as "write a macro to find the size of an object without using sizeof"
or "write a sorty thing without using qsort". I didn't choose the
exercise. I merely solved it.
> * The single input data line contains a real number r and an
> * integer p. Note that p indicates the position of the digit at
> * which r is to be rounded, relative to the units column (so 1
> * means round to nearest 10, 2 means round to nearest hundred, -1
> * means round to nearest hundredth, etc).
>
> This is out and out wrong.
It's the spec! How can it possibly be wrong? :-)
(Let me count the ways...)
Seriously, it seems to me that the problem is well-defined and
meaningful. The fact that Pascal has a different mechanism for
achieving the same goal is of little im****t, surely?
> Apart from that, on a quick scan, I see
> nothing wrong.
Delighted to hear it, but I am concerned at the apparent lack of
robustness of read and readln. For example, I tried running ex08 and
giving an input of 'frog', and got this output:
Runtime error 106 at $080480E4
$080480E4
$0804809F
If this were C, I'd have added a lot more validation, but I don't quite
see how to do that with readln - at least, not yet.
--
Richard Heathfield
"Usenet is a strange place" - dmr 29/7/1999
http://www.cpax.org.uk
email: rjh at the above domain, - www.
|
