On Wednesday 01 February 2006 22:37, doug00 / gmail.com wrote:
> What would be even better would be to allow optional labels after end
> statements, such as "end class", "end def", so the parser can catch
> more errors.
> I've implemented this as well in a separate project.

Holy mother of god in the heavens, for the love of all that's good and sacred, 
please NO! I've had more than enough XML and VHDL already, thankyouverymuch. 
Next thing you see we're coding in palindromes to doublecheck we know what 
order we wrote the characters in...

I want to see a single half-serious programmer that actually makes errors 
because he can't remember what scope he's in. That's what the indentation 
conventions are there anyway. If you keep your code well-structured, It's 
impossible to have that cause a show-stopping bug that can survive five 
minutes in the codebase.

That said, I prefer indentation as a convention, not as syntax. I made certain 
sporadic forays into Python; but I like how Ruby tries to resemble natural 
language with the likes of statement modifiers, and do/begin and end for code 
blocks. It feels more natural to me to type "end" to end a code block, and I 
usually associate the backspace key probably the most commonly used in 
editors to unindent code manually with correcting typos - ergo not a natural 
component in the flow of typing.

Pretty much my $.02

David Vallner
If I had two cents for everytime I say "Just my two cents"...