This is disturbing!

Mark Waddingham mark at livecode.com
Thu Sep 6 12:43:12 EDT 2018


On 2018-09-06 18:21, Mark Wieder via use-livecode wrote:
> Yeah. IMO automatic type conversion is one of the failure points of
> the xtalk paradigm, but it's always been thus.

To be fair, in the days of HyperCard when everything was strings (and 
numbers were decimal strings) the rules worked absolutely fine I think.

However, the use of doubles as the internal rep for numbers, and 
introduction of arrays broke a few invariants a consistent 
implementation of the above view relies upon - hence the annoying points 
of friction.

I don't think implicit type conversion is the problem per-se - just the 
precise details of what gets converted to what, and the inability to say 
'at this point, this needs to actually be a <foo>'.

If you want to be abstract about it then you can view a programming 
language as a compression algorithm - it is a way to express a set of 
possible outcomes in a linear sequence of text. With that point of view, 
they suffer exactly the same problem as any compression algorithm 
suffers - all compression algorithms will expand some input.

i.e. What you might gain in some places in terms of ease / clarity / 
ability; you will lose elsewhere - the hard bit is making sure that such 
cases are 'edge' cases and easily avoided.

Warmest Regards,

Mark.

-- 
Mark Waddingham ~ mark at livecode.com ~ http://www.livecode.com/
LiveCode: Everyone can create apps




More information about the use-livecode mailing list