On Fri, 14 Dec 2001, Roth, Kevin P. wrote:

> > Based on not so very much input, I puy DICT on the list of things
> > to get dumped in v8 unless someone stands up for it.
>
> I use it every once in a while - it's kind of a fun thing to have around. I
> have to wonder - if it's already in and already works, is there any point
> in ripping it out? Unless bugs are found of course, in which case it
> becomes work to keep it in...

In general, there's no point in ripping out working code, no. But there are
reasons why we might do this anyway:

 o If hardly anyone is using the code, it just adds to the footprint with
   very little gain. We already get accused for being "bloated" every now
   and then... The again, the DICT support code is ridiculously small! ;-)

 o We don't have any test cases for the protocol, which means that when we
   rewrite internal designs (like for the upcoming multi interface), we
   must modify the implementation without being able to fully verify its
   functionality.

 o As Bjorn Reese said, the curl implementation is not fully DICT RFC-
   compliant, since that document contains some mistakes that Bjorn pointed
   out to the RFC author(s). They never responded and the document remains.
   This is not a big problem, it just proves that DICT might not be a very
   thriving protocol.

> Having said that, I could simply surf to http://dict.org/ for any of my
> personal dictionary needs, so don't let me keep you from removing this
> protocol if that's what you want to do. ;-)

This wouldn't happen until version 8 if we should decide to do this. It gives
us time to ventilate ideas and to turn things over a couple of more times
before we actually make anything set in stone. :-)

-- 
    Daniel Stenberg -- curl groks URLs -- http://curl.haxx.se/