'''[Tcl Learning Exercises]: Answer: Number Sign Class'''


** Description **

Write a procedure that takes one argument, a number, and returns `-1`, `0`, or
`1`, indicating that the number is negative, zero, or positive, respectively.
The procedure may contain exactly one command, `[expr]`, which may use the
''if-then-else'' (`?...:`) operator at most one time.  '''Bonus Round:''' Add
an additional constraint:  The ''if-then-else'' operator may not be used.

Here is one answer:

======
proc numsign n {expr {$n == 0 ? 0 : $n/abs($n)}}
======

Scroll down for the bonus round answer...
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>
<<br>>

Here is a solution by [rmax] to the bonus round:

======
proc numsign n {expr {($n > 0) - ($n < 0)}}
======

----
[AMG]: While this may be an interesting beginner problem, I don't think it succeeds in demonstrating anything special about Tcl other than that math must be done using [[[expr]]].  Text manipulation problems will likely do better to motivate interest in Tcl.

----
[RKZn] 2017-05-27: Actually there is a subtle difference in the two procs. The return value of first one (with the `[?:]` operator) is represented as an integer or a double, depending on the input's [representation]. The second proc always returns a integer representation. 
======
tcl> info patch
8.6.1
tcl> proc numsign n {expr {$n == 0 ? 0 : $n/abs($n)}}
tcl> tcl::unsupported::representation [numsign -5]
value is a int with a refcount of 1, object pointer at 0xb507f0, internal representation 0xffffffffffffffff:0xb53250, no string representation
tcl> tcl::unsupported::representation [numsign -5.5]
value is a double with a refcount of 1, object pointer at 0x91a210, internal representation 0xbff0000000000000:0xb50ee0, no string representation
tcl> proc numsign n {expr {($n > 0) - ($n < 0)}}
tcl> tcl::unsupported::representation [numsign -5]
value is a int with a refcount of 1, object pointer at 0x919c40, internal representation 0xffffffffffffffff:0xb50880, no string representation
tcl> tcl::unsupported::representation [numsign -5.5]
value is a int with a refcount of 1, object pointer at 0xb50ee0, internal representation 0xffffffffffffffff:0xb53130, no string representation
======

-----

Other ways to do it from my own post on https://codegolf.stackexchange.com/a/106044/29325%|%Code Golf - Output the sign%|% :

======
puts [expr $n<0?-1:$n>0]
======


======
puts [expr !!$n|$n>>31]
======




<<categories>> Tcl Learning Exercises]