From owner-freebsd-doc Thu Feb 14 23:16:40 2002 Delivered-To: freebsd-doc@freebsd.org Received: from bazooka.trit.org (bazooka.trit.org [63.198.170.138]) by hub.freebsd.org (Postfix) with ESMTP id 8EB9D37B404; Thu, 14 Feb 2002 23:16:37 -0800 (PST) Received: by bazooka.trit.org (Postfix, from userid 1000) id 584AD3E2F; Fri, 15 Feb 2002 07:16:37 +0000 (UTC) Received: from bazooka (localhost [127.0.0.1]) by bazooka.trit.org (Postfix) with ESMTP id 544443C12E; Fri, 15 Feb 2002 07:16:37 +0000 (UTC) To: Nik Clayton Cc: doc@freebsd.org Subject: Re: replacement In-Reply-To: <20020213230809.I92878@canyon.nothing-going-on.org>; from nik@freebsd.org on "Wed, 13 Feb 2002 23:08:09 +0000" Date: Fri, 15 Feb 2002 07:16:32 +0000 From: Dima Dorfman Message-Id: <20020215071637.584AD3E2F@bazooka.trit.org> Sender: owner-freebsd-doc@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org [ I'll try not to repeat what others have already said. ] Nik Clayton wrote: > On Sun, Feb 10, 2002 at 10:05:49AM +0000, Dima Dorfman wrote: > > I propose to replace the tag with and a "role" > > attribute as follows: > > > > > > 1. Definitely 'package', not 'port'. Ports are just the infrastructure > that produce packages. Indeed; my change from "port" to "package" was deliberate. People have already given reasons for and against this, but my main motivation (which has only been mentioned in an off-hand manner) was that "port" is ambiguous; I can think of at least three different ways it can be interpreted in this context (TCP/UDP type ports, i386/alpha type ports, and cvsup/procmail type ports). Given that, as you [Nik] point out, FreeBSD is the only place where the definition we want is used, it probably isn't a good idea. If people are concerned with length (I haven't seen this brought up), we can use "pkg", but SGML isn't exactly known for being pithy. > 2. I have a nagging feeling that we should make sure the package's > category is included somewhere that DSSSL/XSLT stylesheets can > access it. > > net/cvsup > > is less easy to parse (in a stylesheet) than something like > > cvsup I don't really see how the category would be useful without the port name. Actually, the "net/" part of "net/cvsup" isn't exactly the category; it's the directory in which the port resides. net/cvsup is in `net' and `devel', but only `net/cvsup' would work in this case. Either way, I don't see how we could use the category. I guess I'm not really objecting to this, but rather I'm not sure it's worth the trouble. I can't see how it would help, and it adds more verbosity (as if there wasn't enough (I know SGML is verbose by nature, but I don't have to like it)). If [most] people disagree with me, I'll do the (trivial) work to make it happen. The second construct above is also ugly in that it adds a `role'-sensitive `category' attribute. It isn't really invalid, but it certainly isn't elegant. > We can always make these entities, something like > > &pkg.net.cvsup; I don't really have anything to add on this point except to point out that if we eventually decide to do this people will want to replace all occurances of '' with entities at once, I'd rather only have to do one mass-commit (i.e., get it done right the first time). If, on the other hand, people would be content with having some documents (or parts of documents) use entities while others use directly, I'd rather get everything converted to the latter for now and leave this discussion (and decision) for later. Thanks, everyone, for your comments. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-doc" in the body of the message