[plt-dev] `rico' and ".rkt"

From: Carl Eastlund (carl.eastlund at gmail.com)
Date: Mon Feb 15 15:13:49 EST 2010

On Mon, Feb 15, 2010 at 2:00 PM, Carl Eastlund <carl.eastlund at gmail.com> wrote:
> On Mon, Feb 15, 2010 at 1:50 PM, Shriram Krishnamurthi <sk at cs.brown.edu> wrote:
>> I agree.  I simply don't see the point to moving past the three-letter
>> suffix space.
> I think Jon Rafkind made a good point on IRC, which is that as our
> extensions proliferate we don't want to end up with a set like
> OCaml's: .ml, mli, .cmo, .cma, .cmi, and .cmxa.  Given a directory
> with files of all those extensions, an outsider will have no idea what
> language they correspond to -- except possibly the .ml one, because ML
> is the full langauge name -- let alone which aspect of OCaml each
> corresponds to.
> If we have ".ss" and ".rkt" and ".scrbl" and ".zo", it won't be clear
> what language any of them correspond to, nor that they are all
> related.  If we have ".racket", at least that one extension will be
> readable.  If they were ".plt-scheme", ".plt-racket", ".plt-scribble",
> and ".plt-compiled", their purpose and relationship would be much
> clearer.  Even without the "plt-" would give half that benefit
> (purpose, but not relationship).

To be clear, I was illustrating the point of why I want longer names:
for clarity of purpose of files.  Going to long names all starting
with ".plt-" would have lots of clarity.  This is not, however, the
time to change all our conventions at once.  The ".plt-" and
"scribble" and "compiled" parts are all purely hypothetical, as far as
I am concerned.  My concrete proposal is still to name our Racket
files ".racket" and leave all existing extensions alone.


Posted on the dev mailing list.