Re: [q-lang-users] Special values or failing functions

[Tim H. <q...@st...>]

Alexander Nickolsky <AN...@sp...> writes:

[snip]
>   Then there comes another problem. Suppose I have a list of values
>   that I want to process using function like this, leaving only
>   results that make sense, that are 'proper' values.
>   It seems to me that I cannot make this procedure polymorphic,
>   because if for example 'false' means unwanted value for integer
>   function, 0 could mean unwanted value for boolean function, etc.
>   Another example is further processing of results. Say, I want
>   to find the longest path. It leads to the expression :
>
>   if ((isnum P) and (isnum Q)) then (max P Q) else if (isnum P) then P
>      else if (isnum Q) then Q where P = (dist (A,B) N ), Q = (dist (C,D) N );
>
>   I do not think it's very readable. Wouldn't it be nice to have a
>   common 'special' value like nil or NULL or whatever, to indicate
>   failure ?

Might () be appropriate?

Without going too far into the specifics of your algorithm, the two things
I would be doing are
a) giving your expression a name - specifically, use more helper functions.
You could invent "getmaxh" which takes all the (A,B) and (C,D) and dist
(A,B) N and all that, which pushes the conditional checks down a level
b) seek to use guards: write more rules for getmaxh that have the "when" on
the left side:

getmaxh (A,B) (C,D) () = ();
getmaxh (A,B) (C,D) Z:Integer = foo;
getmaxh (A,B) (C,D) Z = bar;    //general case

getmax  (A,B) (C,D) = something to do with getmaxh introducing the Z parameter;


As a bonus idea, can you seek to work with lists more, thereby allowing a
filter (<>()) expression as a tidy-up phase (before or after getmax-ing)?

~Tim
-- 
<http://spodzone.org.uk/>