IDE versus MSG Box - Field Tabstops

[Geoff Canyon]

On Wed, Aug 14, 2013 at 8:09 PM, Peter Haworth <pete at lcsql.com> wrote:

> I'm leery of
> anything that tries to interpret my input based on arbitrary logic and
> would much rather set either tabstops or tabwidths and have each of them
> retain the values I set them to without trying to interpret them.
>
> Also, I've always found the automatic generation of tabstops under Geoff's
> rule 3 to be arbitrary but no doubt some folks prefer it that way.
>


I understand "arbitrary" even less than "sometimes." This behavior does
something useful with input that otherwise does not meet the spec. The
alternative would be to say that the order of the tabstops is irrelevant
and the values are always absolute, but presumably then this would happen:

  set the tabstops of fld 1 to "200,100,50,400,300"
  put the tabstops of fld 1 -- puts "50,100,200,300,400"

Which of course means that (as has been decried in other posts) the output
is not identical to the input.

I'm puzzled by the concern regarding this particular "interpretation."
LiveCode is overflowing with cases where atypical input is interpreted
sensibly, from negative character references, to type inference/automatic
casting, to storing the names of properties in variables for future use as
properties, to this little gem:

  put "test3this" into x
  add 14 to char 5 of x
  put x -- puts "test17this"

I wonder if all of those seem good and useful just because we're all used
to them.