discuss: Man pages converted to HTML
Subject:
Re: [discuss] Man pages converted to HTML
From:
David Lawyer ####@####.####
Date:
28 May 2007 23:03:10 -0000
Message-Id: <20070528230324.GA2227@davespc>
On Mon, May 28, 2007 at 09:28:26AM -0400, Sam Varshavchik wrote:
> David Lawyer writes:
>
> >Would you be willing to help LDP by setting up and maintaining a
> >distribution of manpages on the LDP website? I think that LDP should
> >provide them in both HTML and groff (troff) just like we supply HOWTOs
> >in plain text and html. HTML doesn't help much with the existing man
> >pages since they don't contain links (but should!).
>
> Of course they do. For starters, they have internal links. I did not
> convert all the man pages to HTML by hand. That would've been insane. I
> use a script, and right now it adds up to about ten thousand lines of Perl
> code. The script has some heuristics, which includes taking what looks
> like references to other man pages, and turning them into a link. Of
> course, that only works if it's a reference to another page that I also
> have, but I'll get back to this in a moment. This also means that
> refreshing the HTML against newer man page releases requires very little
> work. When I began the conversion, the current release was 2.41. Updating
> to 2.46 took about a week. I started on 2.48 yesterday, and will probably
> finish it today. So, there's really nothing here for me to
> maintain.
I didn't realize you had done all this. So HTML does help for
following links, but the majority of the time I use the man pages I
don't need to follow any links. But it's sure nice to have the links
when one does need them.
> My HTML compilation includes not just LDP's man pages, but they also
Actually, they are not LDP's man pages. The latest manifesto says
nothing about LDP maintaining any man pages and when it did it was
mostly a mistake to say this. These pages may be found at their main
site:
ftp://ftp.win.tue.nl/pub/linux-local/manpages
> include man pages from several other sources. This brings up a problem
> with sourcing man page distribution as HTML. It's also link related: with
> HTML, you'd obviously want all the links to work. That, of course, is not
> a problem for internal links to other pages in your collection. But what
> about links to man pages that come from other sources? There are several
> references in LDP's man pages to other pages, specifically coreutils and
> util-linux. Because I include them in my HTML compilation, those links
> work.
>
> I'd certainly be willing to assemble a tarball for you that contains only
> the LDP pages, I'll just need a little bit of time to modify my script to
> make a second tarball with LDP's subset. But it will not have any links to
> non-LDP man pages, which makes it less useful, of course. For that reason,
> I'd still want to keep my multi-sourced collection.
I'd like to see the whole multi-sourced collection available on LDP
but we are short of volunteers so if you are willing, perhaps you
could be given write access to the part of the ldp site that would
contain the man pages in both html and groff, etc. This likely would
be the page (and directory tree) you get to by selecting "man pages"
on LDP's homepage. Thus you would be the LDP volunteer maintaining
LDP's expanded distribution of man pages, but not actually maintaining
any particular man-page (unless there is to be a man page telling
people about the availability of html man-pages).
Perhaps the other stuff referenced on the page you get to by clicking
"man-page" should be put into another directory also referenced from
the homepage --or what amounts to the same thing: create a new
directory for only man-pages.
David Lawyer
