discuss: Thread: output tree wrangling (website content overhaul); archive proposal


[<<] [<] Page 1 of 1 [>] [>>]
Subject: output tree wrangling (website content overhaul); archive proposal
From: "Martin A. Brown" ####@####.####
Date: 19 Mar 2016 00:41:44 +0000
Message-Id: <alpine.LSU.2.11.1603181637280.12423@znpeba.jbaqresebt.arg>

Hello all,

I have been working through our output directory (the filesystem 
underneath http://www.tldp.org/) to try to understand what is 
present and what changes would be necessitated by automated 
publication of HOWTOs and guides (and the refs and faqs), from our 
git repository.

More findings and thoughts....

  * There are many renamings, for example, 'ppp-ssh' has become 
    'VPN-PPP-SSH-HOWTO' and (embarassingly, as I contacted the 
    author claiming it wasn't on our site) 'intro-linux' should be
    named 'Intro-Linux' to match the name in the source tree.

    I plan on providing some polite HTTP user-agent redirection to 
    the newest name of these older documents.  This will be 
    accomplished with this familiar trick and a bit of text.  (I 
    tried it out and 2 seconds is long enough for a human to see the 
    text, but short enough to not be unusually intrusive.)

      <meta http-equiv="refresh" content="2;URL='$NEWURL'" /> 

    My suspicion?  Thoroughly uncontroversial.

  * Some of the documents in the output tree have been retired 
    and/or moved many years ago.  Where there's a redirection 
    already in place for a document, I plan to drop the redirect.

       * Acer-Laptop-HOWTO
       * Linux-From-Scratch-HOWTO  (see more on LFS below)
       * Distributions-HOWTO
       * MIPS-HOWTO
       * 3Dfx-HOWTO
       * PostgreSQL-HOWTO.html

    My suspicion?  Thoroughly uncontroversial.

    (The newest of the above redirects was installed in 2005, so I 
    don't think there should be much dissent, at least, not on the 
    part of those who had written the original documents and moved 
    their content to a separate location more than ten years ago.)

  * Archive candidates:  there are some documents which are very, 
    very old, for example one on Sendmail+UUCP and another on an old 
    serial card, the GTEK-BBS-550.  These deserve to be held in our 
    archive forever.  Some of these are so old that we do not have 
    the sources in our version control system.

    I hope this is uncontroversial.

    See proposal.

  * Archived data:  we already have an archived set of data, stored
    under http://tldp.org/pub/Linux/docs/.  To my knowledge, there 
    are no links directly to this documentation and these elements 
    of the history of Linux.

    See also below proposal. 


Archive proposal:
-----------------
I propose (and will volunteer to undertake some of the effort) to 
write a DocBook XML document that provides a human-curated index to 
a partial store of archived documents we maintain in our document 
root [0].  The goal would be to provide some context around what 
this stuff is, who created it and why, and provide links into the 
historic matter.

I would welcome any others who wished to help examine some of this 
older content.

If nobody objects in a week or so, I will probably start on this 
project, while continuing on the output tree reorganization and 
cleanup.


Future ideas:
-------------
Documentation of various sources could be hosted on TLDP even if 
not in our version control system.

For example, the Linux From Scratch project is wonderful and 
interesting.  The content is great stuff.  Our source repository has 
an old version (5.0) of the LFS book (which was first seen in the 
TLDP repo on 2000-04-26 as the Linux-From-Scratch-HOWTO).  The 
current version of the LFS book, though, is 7.9.

I'm not doing anything with the idea yet.  This brings up the idea 
(yet again; this is not new) that it would be nice if TLDP could 
provide a facility for checking out remotely maintained 
documentation sources furnished by third parties, in this case LFS, 
and build that into our hosted and mirrored output tree.  A laudable 
goal!

If we were to do this, I'd think we would want to be able to be able 
to checkout (or clone) the documentation from whatever remote 
version control system was in play, put the files into an 
appropriately-named directory and fire-off a job to process the 
documentation automatically into a separate area of the output tree.

  Stuff from our VCS                     Would end up here:
  https://github.com/tLDP/LDP/...   ->   http://www.tldp.org/en/

  Stuff from another VCS                 Might end up here:
  hg://another.id/ERF/ERF-HOWTO     ->   http://www.tldp.org/hosted/

There would clearly be some work and, especially coordination 
involved here, but for anything that was Linuxdoc, Asciidoc 
or one of the DocBook formats popular in the last 15 years, we could 
probably handle.


Anyawy, I'm going to retain focus on cleaning up our existing output 
tree first, engage the metadata management question, and then maybe 
work on the hosted documentation concept.  Unless, of course, 
somebody else beats me to it!

That's all from here,

-Martin

 [0] http://tldp.org/ls-lR.txt

-- 
Martin A. Brown
http://linux-ip.net/
[<<] [<] Page 1 of 1 [>] [>>]


  ©The Linux Documentation Project, 2014. Listserver maintained by dr Serge Victor on ibiblio.org servers. See current spam statz.