On Thu, 21 Nov 2002 09:17:17 -0800 (PST), "Jeremy C. Reed"
<reed@reedmedia.net> wrote:
> On Thu, 21 Nov 2002, James K. Lowden wrote:
> 
> > I don't like /usr/pkg/share/doc/html/<package> because I generally
> > don't like naming directories according to the format of its files'
> > contents.  I look for documentation for <package>; the format of that
> > documentation is not in my search criteria.
> 
> One reason could be that you could easily point a webserver's document
> root to that single /usr/pkg/share/doc/html directory to serve up
> documentation for various applications. 

I hadn't thought of that.  I can imagine a time, maybe 5 years ago, when
html was special that way, and people thought html docs would be served up
centrally instead of locally installed.   Of course, it would be more
compelling if there were a  /usr/pkg/share/index.htm.  :)

It didn't turn out that way.  For every /usr/pkg/share/doc/html available
via a web server, there must be 1000 that aren't.  

> (But, I don't know of any security
> impact of these docs for the local system or end-users, especially when
> I hear there are many ways to compromise Windows systems through MSIE.)

The day /usr/pkg/share/ poses a threat to Windows users, I'll be very,
very concerned!  I think it will pose a threat to Microsoft first.  ;-)

As Greg says, browsers understand more than html.  Neither /usr/share nor
/usr/pkg/share is sensitive.  hier(7) doesn't talk about /usr/pkg, but
says /usr/share has "architecture-independent text files".  The FHS 
http://www.pathname.com/fhs/2.2/fhs-4.11.html says much the same, "for all
read-only architecture independent data files".  

If I were going to publish the docs, I don't think I'd stop at html.  I'd
probably

	ln -s /usr/pkg/share /usr/share/pkg

and publish /usr/share.

Regards, 

--jkl