Re: [Evolution] Dealing with broken mailers

[Dan Winship <danw ximian com>]

> In my opinion, optimal interoperability and compatibility with existing,
> if broken, MUAs should be a priority, especially if the MUAs in question
> are used by (I suspect) millions of people.

But where do you stop? There's no end to the number of broken things
mailers can do.

Early web browsers were very lax about the sort of HTML they accepted,
and as a result, now any browser that wants to succeed must accept all
manner of broken HTML, and render it the same way everyone else does, or
people will consider that browser to be broken and not want to use it.
This makes implementing an HTML rendering engine incredibly painful.
Just ask the GtkHTML hackers.

Most other areas of the internet haven't reached that level of
brokenness, and that's a good thing. But if people add hacks to mail
clients to work around problems in other clients and servers, then the
people shipping the buggy software have less pressure on them to fix the
problems, and worse yet, people implementing new clients and servers may
make the same mistakes and never realize it, because the clients they
test against are trying to work around their bugginess. And then after a
while, anyone who wants to write a successful mail client will need to
write thousands of lines of compatibility crap, and do interoperability
testing with dozens of other programs to make sure that it's doing all
the necessary kludges.

-- Dan