Date: Fri, 15 Dec 2000 12:13:19 -0800 From: Bengt Richter <bokr@accessone.com> To: "Michael C . Wu" <keichii@peorth.iteration.net> Cc: freebsd-doc@freebsd.org Subject: URLs too transient? Re: www/en/security/security.sgml update Message-ID: <3.0.5.32.20001215121319.00900860@mail.accessone.com> In-Reply-To: <20001215021620.A37166@peorth.iteration.net>
next in thread | previous in thread | raw e-mail | index | archive | help
PMJI, but I had a possibly useful thought (?):
Using URLs/links in docs is great when there is assurance
that they can be followed, but for a long-lasting document,
is it wise to point to internet sites per se, unless there
is some kind of official guarantee that they will persist?
In a hardcopy manual, you wouldn't refer in a footnote to
other hard copy located at the library on 123 Main St, Anytown,
USA. You would refer by title, author, and ISBN number, etc.,
so it could be found in any similarly indexed repository.
That last is the key (pun) to what I am getting at: To have
reliable persistent distributed docs, the references have to
be in terms of a reference system, not locations. URLs serve
both purposes, so it's confusing, but it bears thinking about
(IMHO). Look at where that confusion led in the DOS/windows
file system and you realize the pain caused by including impermanent
physical/hardware references ('<drive letter>:') in what should be
pure information references.
Imagine seeing '#include "/dev/<hard disk id>/.../actual-file.h"'
in unix source! Gak. Thank goodness the unix founding fathers thought
to provide mount, to keep the information space separate and clean.
For a document that may exist live on the internet, on local disks,
on CD, and possibly bound paper, "http://" is a little like "/dev/",
so should non-relative references be strictly controlled/validated?
If you went with the unix philosophy, I guess you would have a mount
point (e.g., /http, like /cdrom), and keep the physical source and
file system format separate. Then we could have a document hierarchy
based on a mount point as unchanging root, instead of using site URLs
as root(s). Maybe not so much would have to change, except being very
careful about non-relative links.
What URL represents a reliable document root forever? I wouldn't matter,
if you could just plunk it into fstab.
Just my USD.02
Regards,
Bengt Richter
"We are all ignorant, just on different subjects." -- Will Rogers
At 02:16 2000-12-15 -0600 Michael C . Wu wrote:
>Well, here is another update. :)
>
>I removed the list of security advisories and simply pointed
>people at the url to fetch them, this should be lower maintainence
>for our security staff. IMHO, this is also a better way to save
>bandwidth for people in low-bandwidth countries. To list
>the whole set of advisories seems pointless.
>
>I am not sure how best to link to /auditors.html, so please make sure
>you do it the Right Way. :)
>
>A link and explanation is added to the auditing project and
>its homepage.
>
>
>--
>+------------------------------------------------------------------+
>| keichii@peorth.iteration.net | keichii@bsdconspiracy.net |
>| http://peorth.iteration.net/~keichii | Yes, BSD is a conspiracy. |
>+------------------------------------------------------------------+
>
>
>To Unsubscribe: send mail to majordomo@FreeBSD.org
>with "unsubscribe freebsd-doc" in the body of the message
>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-doc" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3.0.5.32.20001215121319.00900860>