On 2005-01-26 22:34:34 -0500, Derek Martin wrote: > I personally never use Usenet, but to me, NNTP and e-mail seem > like (from the perspective of the user) exactly the same thing, > so I see no reason why mutt shouldn't do both. Both involve the > sending and receiving of messages, and the nature and > construction of those messages are essentially identical. The > only difference is on the back end, which is not at all of > concern to the user. What is of concern is having a uniform > interface to those things which are done the same way. I think > mutt needs NNTP support. That argument has a lot in favor of it. But, even when I was still reading Usenet news, requirements for tools tend to be different. With e-mail, I have to read and answer as many messages as efficiently as possible, and I have to make sure I can quickly find old messages. With Usenet, I had to make sure my user agent recognizes and deletes as many irrelevant articles as possible, and it wasn't a problem when some article was ignored. So, basically, I'd leave it for programs like slrn to be good news readers, and would be happy with mutt remaining a good mail reader. > > And I'm skeptic about header-caching maildirs. > What is the nature of your skepticism? the performance > enhancement is substantial, and there has been no complaint of a > down side that I've seen... What's to be skeptical of? What I'm skeptical about is the point that maildir is built very carefully not to require standard Unix locking functions. That makes maildir the most robust folder format around, even in broken NFS environments. The hcache patch, as I read it, requires locking mechanisms to work. Hence, it will take some of maildir's stability away. Regards, -- Thomas Roessler · Personal soap box at <http://log.does-not-exist.org/>.
Attachment:
pgpnTwWyi5OGC.pgp
Description: PGP signature