At 10:35 AM -0600 8/13/06, David Green wrote:
On 8/8/06, Darren Duncan wrote:

I thought your reasons made sense, and would be happy with a "Text" type, although I don't especially object to "Str" -- as you say, it's probably good enough given ordinary programming usage.

However the IRC excerpt contained some comments about not conflating or confusing "Str" with "strings". To me, that would mean they *should* have different names. If there's a distinction worth making, then it's worth using distinct terminology, and "Str" vs. "string" isn't distinct enough: you can tell people they're not really the same thing, but they ain't gonna believe you. There's nothing else "Str" could reasonably mean other than "string" (just as I don't think "regex" can very reasonably mean anything other than "regular expression").

Yes. And I agree that "Str" doesn't reasonably mean anything other than "string". My conclusion is that, given the context, I am satisfied with either "Text" or "Str" for the sequence of unicode characters.

While the issue could be revisited if enough other people prefer "Text", I decided (and announced) that I would focus my energies on pressing matters I found more important (frying bigger fish).

It's related to a principle of "crying wolf". If I keep pressing for every little thing to go my way, people won't want to listen when I raise more important issues.

An example that I did continue to press after initial indifference (or opposition) was a 'Blob' type addition, which ended up happening (thank you again, Larry). Other examples are ongoing, as I work to build a better case before re-raising them.

And besides, I have on several occasions raised suggestions where I was subsequently convinced that my proposal was wrong / would have made things worse. Eg, I once proposed that an undefined variable should never auto-vivify to a defined value, needing instead an explicit real value assignment, and suggested certain behaviour changes relative to Perl 5 on what should happen if you attempt to use an undefined value as a defined one (various types of failure). I later realized that this proposal would have made things worse in several ways, and in fact I would now oppose such a change, as unlikely as that is to be proposed.

But if you think more strongly about a 'Str' name change, good luck in raising a better case for it.

-- Darren Duncan

Reply via email to