<br><div class="gmail_quote">On Thu, Oct 7, 2010 at 2:26 PM, Greg Hendershott <span dir="ltr"><<a href="mailto:greghendershott@gmail.com">greghendershott@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<br>On the server side, it's tricky because I obviously wouldn't want to<br>
rewrite the Racket web server, dispatching, continuations, etc. I'm<br>
just talking about the "edge" of it that is the user-supplied<br>
dispatchee procs. Have them take a request this way and provide<br>
response this way, instead. Maybe this is where the idea gets stuck in<br>
mud, if were to come down to two bad choices: A bolt-on with too<br>
inefficient translation, or, going too far into rewriting server guts.<br></blockquote><div><br></div><div>The different level of difficulty comes from that you are generating a request on the client, but trying to parse a request on the server. If you focus on just generating a response on the server (or also try to parse the response on the client), then the two will be roughly on par. </div>
<div><br></div><div>Basically it is harder to parse than to generate. So I would say having a translation layer will be the way to go for now, until in the future you are ready to parse the request yourself, that's when the ol' good RFCs come in handy. </div>
<div><br></div><div>Cheers,</div><div>yc</div><div><br></div><div><br></div><div><br></div></div>