[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Nmh-workers] More fun with charset functions

From: Robert Elz
Subject: Re: [Nmh-workers] More fun with charset functions
Date: Tue, 13 May 2014 01:36:35 +0700

    Date:        Mon, 12 May 2014 14:12:12 -0400
    From:        Ken Hornstein <address@hidden>
    Message-ID:  <address@hidden>

  | The problem is right now, architecturally, it's hard (see previous
  | email).

Yes, I remember that.   But it isn't.   What is hard (I gather) is to
do it in a way that is architecturally clean.   Just printing a message
and exiting is trivial (ie: adios()), anywhere (and even if it leaves some
random state around for someone else to clean up, that's acceptable, as
a quick fix for a situation that should not really happen).

  | Returning "unknown-8bit" is at least valid according to the RFCs.  We
  | can always go back and fix that later (that DOES happen, although the
  | time frame can be large).

Unfortunately, that as a solution can also sometimes be impossible.  Once
code is distributed that acts in a certain way, some subset of the users
will grow to depend upon it, and then fixing it (or attempting to) can
generate lots of negative feedback (however stupid the "feature" the users
are depending on happens to be).

Given the choice, I'd go with the ugly code solution, that does the right
thing, over a seemingly cleaner "keep on going" solution, every time.
Ugly code (or code structure) does get fixed, well sometimes - if it is
ugly enough, and someone gets motivated.   Mis-features often never do (or can).


reply via email to

[Prev in Thread] Current Thread [Next in Thread]