From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.io/gmane.emacs.gnus.general/38788 Path: main.gmane.org!not-for-mail From: prj@po.cwru.edu (Paul Jarc) Newsgroups: gmane.emacs.gnus.general Subject: Re: IMAP vs nnml ? (was: Palm and Gnus) Date: Sun, 16 Sep 2001 01:00:45 -0400 Organization: What did you have in mind? A short, blunt, human pyramid? Message-ID: References: <87d74u4r9g.fsf@inanna.rimspace.net> <87zo7wsqm8.fsf@inanna.rimspace.net> NNTP-Posting-Host: coloc-standby.netfonds.no Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable X-Trace: main.gmane.org 1035174595 24311 80.91.224.250 (21 Oct 2002 04:29:55 GMT) X-Complaints-To: usenet@main.gmane.org NNTP-Posting-Date: Mon, 21 Oct 2002 04:29:55 +0000 (UTC) Return-Path: Return-Path: Original-Received: (qmail 7975 invoked from network); 16 Sep 2001 05:00:47 -0000 Original-Received: from multivac.student.cwru.edu (HELO multivac.cwru.edu) (261@129.22.96.25) by gnus.org with SMTP; 16 Sep 2001 05:00:47 -0000 Original-Received: (qmail 11472 invoked by uid 500); 16 Sep 2001 05:01:07 -0000 Mail-Followup-To: ding@gnus.org Original-To: ding@gnus.org Mail-Copies-To: never In-Reply-To: (Kai.Grossjohann@CS.Uni-Dortmund.DE's message of "Sun, 16 Sep 2001 01:36:58 +0200") User-Agent: Gnus/5.090004 (Oort Gnus v0.04) Emacs/20.7 Original-Lines: 16 Xref: main.gmane.org gmane.emacs.gnus.general:38788 X-Report-Spam: http://spam.gmane.org/gmane.emacs.gnus.general:38788 Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Gro=DFjohann) wrote: > prj@po.cwru.edu (Paul Jarc) writes: >> I would not be surprised to learn that nnmaildir is much slower on a >> block-structured filesystem. ReiserFS (which is not block-structured) >> handles lots of small files quickly and space-efficiently, and >> nnmaildir takes advantage of that. > > Well, I don't expect nnml to be better than nnmaildir in this case. > IIUC, both use one file per message. Yes, but nnmaildir also uses one NOV file per message. So startup would probably be significantly slower. (Afterwards, the NOV data is typically cached in memory.) paul