<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=UTF-8" http-equiv="Content-Type"> <title></title> </head> <body bgcolor="#ffffff" text="#000000"> Robby Findler wrote at 02/15/2010 12:52 PM: <blockquote cite="mid:932b2f1f1002150952m33720a7j6d4545aff0c91540@mail.gmail.com" type="cite"> <pre wrap="">On Mon, Feb 15, 2010 at 11:38 AM, Neil Van Dyke <a class="moz-txt-link-rfc2396E" href="mailto:neil@neilvandyke.org"><neil@neilvandyke.org></a> wrote: </pre> <blockquote type="cite"> <pre wrap="">Would it work to make the automagical translation happen *only* if ".ss" is specified (which is what all the legacy code has)? Then, if ".rkt" or no extension is specified, then assume the filename has ".rkt", and never assume ".ss". </pre> </blockquote> <pre wrap=""> There may be legacy code providers, too. Ie a planet package that has a "main.ss" file intead of a "main.tekcar" (or whatever the extension ends up being). </pre> </blockquote> <br> I'd say that people would continue to say "main.ss" in this case. Then, the automagical rule for ".ss" in references will keep their code running if that package ever renames the file to "main.rkt".<br> <br> Basically, seeing ".ss" in code is a sign that legacy code is involved one one or both ends, so that's when the automagical backward-compatible stuff kicks in.<br> <br> When code instead says ".rkt" or (my proposal) has no extension at all then we know legacy code is *not* on either end, so we don't pollute the behavior of new code with the backward-compatibility kludge.<br> <br> <div class="moz-signature">-- <br> <a class="moz-txt-link-freetext" href="http://www.neilvandyke.org/">http://www.neilvandyke.org/</a> </div> </body> </html>