From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266128AbUIALP1 (ORCPT ); Wed, 1 Sep 2004 07:15:27 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266136AbUIALP1 (ORCPT ); Wed, 1 Sep 2004 07:15:27 -0400 Received: from bay18-f27.bay18.hotmail.com ([65.54.187.77]:14354 "EHLO hotmail.com") by vger.kernel.org with ESMTP id S266128AbUIALPZ (ORCPT ); Wed, 1 Sep 2004 07:15:25 -0400 X-Originating-IP: [67.0.155.180] X-Originating-Email: [codermattie@hotmail.com] From: "michael mattie" To: linux-kernel@vger.kernel.org Subject: Re: silent semantic changes with reiser4 Date: Wed, 01 Sep 2004 11:15:23 +0000 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 01 Sep 2004 11:15:24.0360 (UTC) FILETIME=[FA239C80:01C49014] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org In reading this thread I have not found one example *app* that cannot be implemented in XML. From my understanding the big win of hybrid files is that they expose the structure of files to the Unix toolset. Assume that this is possible and you can cd through the structure, you have acheived zero code sharing, only a never ending job of intergrating a infinite list of format interpreters. With XML you have a concesus on syntax (structure) that is independant of the semantics and a significant and error prone chunk of code can be shoved into a userspace library. I don't see anything like this in hybrid files. So far the vision I see in this thread for hybrid files is icons, mime types, character encoding, thumb nails, and meta data ? This is hetrogenous content ! I see no win in a VFS interface to a ad-hoc poor re-implementation of w3.org. I do admit it can be exciting to cd through the internal structure of a file, I do it all the time with Xpath and XSH [ see http://xsh.sourceforge.net/ ] from CPAN. The big win of hybrid files can be had with a install. Transparent traversals of archive files should be shoved into the shell with all of the other ugly hacks where it can be caged and easily killed with a signal. Even windows programmer's reached this conclusion for some of their cruft. My conclusion from reading this thread is that a proposal for hybrid files should demonstrate a significant win over the work of w3.org before such a radical semantic alteration is made. I hope the comments of a humble and grateful user can be of some use and please accept my apology if I have mis-understood this discussion or squandered your time. _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today - it's FREE! hthttp://messenger.msn.click-url.com/go/onm00200471ave/direct/01/