Re: [idn] phasing out ACE

["D. J. Bernstein" <djb@cr.yp.to>]

Adam M. Costello writes:
> There has been much debate about whether ACE should be phased out. 
> Even though people have different views, I don't think we really need
> to debate this question.

There are two issues here.


1. Should applications be required to handle UTF-8 properly, in
preparation for a UTF-8 world?

The lessons of history are helpful at this point. If Keith Moore and his
buddies had---with or without Quoted-Printable---required 8-bit-clean
mail software in 1991, we wouldn't have sendmail's 8-bit bugs today. If
they had thought a little more broadly and fixed other protocols, we
would have been able to deploy UTF-8 years ago.

Unfortunately, they, like you, didn't consider the long term. I don't
know if they were as blatant as you in claiming that shortsightedness
was a good thing; but the bottom line is that they, like you, made a
decision destined to look really stupid to future users.


2. Should short-term plans be considered in the context of the desired
transition to UTF-8?

Common sense is helpful at this point. The cost of the long-term move to
UTF-8 obviously depends on the choice of short-term plan. Consquently,
ignoring the UTF-8 transition will not produce the same cost-benefit
analysis as taking it into account.

Here is one example of the variation in costs. Suppose we do a massive
redeployment of mail-displaying programs etc. to present IDNA names as
non-ASCII glyphs. Moving to UTF-8 then requires _another_ massive
redeployment of those programs.

In contrast, if IDNA were modified to also require proper display of
UTF-8, programs dedicated to displays would require only one upgrade.
Notice how a small change in the short-term plan provides huge long-term
benefits---benefits that you are refusing to take into account.

Even better, if we scrap this 7-bit garbage and move directly to UTF-8,
_all_ programs will require at most one upgrade---maybe 0.5 on average.
But you don't need to think about this possibility to understand why
ignoring the long-term plan is stupid.


---D. J. Bernstein, Associate Professor, Department of Mathematics,
Statistics, and Computer Science, University of Illinois at Chicago