[racket] `def' ?

From: Neil Van Dyke (neil at neilvandyke.org)
Date: Thu May 10 13:46:00 EDT 2012

Luke Vilnis wrote at 05/10/2012 12:58 PM:
> [...] along with (imo) sexier looking code samples to show to 
> prospective Racketeers is worth it.[...]

Separate from whether "def" is a good idea, I think we might be better 
off forgetting about sexy examples for a while.  We tend to get our 
'sexy' code examples wrong anyway, and I suspect that we will continue 
to get them wrong even if some of the identifiers are made shorter.

I think Racket evolution should have a guiding principle of doing things 
that its developers find useful or otherwise think are good practice, 
and *not* get derailed changing the language to try to appeal to 
real/imagined other people who can't see that or won't take the 
initiative and try it.

(Tangential anecdote: Before I started learning Scheme, one of the 
reasons I liked Lisps was that I could get all terse with lots of 
tricks.  Scheme taught me that terse-ness is not as important as I 
thought it was, and indeed I was doing it more than optimal.  This is 
something I came to appreciate because I made an effort to try what 
seemed to be the idiomatic ways of doing things.  Of course, I'm still 
biased by prior experiences, just like everyone else, but I did make the 
effort.  Given that, these mythical prospective Racketeers, on some 
Python/Java/Arc/somecrud site, pontificating on how Racket would be so 
much more practical if only there were such-and-such changes, without 
actually knowing Racket-- they can eat my shorts.)

I'm not criticizing Luke's suggestion.  This is just an opportunity to 
point out what I consider an ongoing/recurring problem.  I think I am at 
least as guilty of this as anyone.

Neil V.


Posted on the users mailing list.