discuss: wanted: a really good case for LinuxDoc
Subject:
Re: wanted: a really good case for LinuxDoc
From:
David Lawyer ####@####.####
Date:
13 Dec 2003 22:13:13 -0000
Message-Id: <20031213221056.GB507@lafn.org>
On Sat, Dec 13, 2003 at 01:54:35PM +0100, Stein Gjoen wrote:
[skip]
>
> Finally a number of semantic web arguments have been put forward
> to somehow prove the superiorness of DocBook. Metadate, we hear,
> should allegedly make us a better future. My arguments against
> are as follows:
> - we do not appear to move measurable towards this future
> - we have no current use for these extra futures
> - I have already shown that the tags in LinuxDoc SGML
> can be used to create NetHelp documents that offer real
> advantages here and today.
I don't fully agree with the above, but I still think that it's usually
better for new writers to use LinuxDoc. With LinuxDoc, semantic analysis
will be more difficult than for DocBook. So the writer of the semantic
analysis software will need to do more work at the benefit of less work
by the authors using LinuxDoc. So one person does more work in writing
the software but thousands of writers do less work in writing. Even if
the extra work done by the software programmer exceeds the effort saved
by a hundred writers, it's still less person-hours if LinuxDoc is used
instead of DocBook.
Another point. The tags only help perhaps less than 1% in a full semantic
analysis. So the benefit is not very much. When AI is smart enough to
understand the meaning (based in part on world knowledge), it may be
almost smart enough to write the documentation so that then there would
be no need for markup languages anymore.
> The NetHelp approach offers
> - better overview of document
> http://www.nyx.net/~sgjoen/nethelp-c.html
> - better keyword facility
> http://www.nyx.net/~sgjoen/nethelp-i.html
> - document interlinking (ref. header in both URLs above)
> - and I do have a working prototype
> http://www.nyx.net/~sgjoen/home.html
>
> Conclusion: LinuxDoc SGML offers all we need and allows further
> advances as is. It is easy to start with and it is easy to
> continue with. The tool chain is simple and it is mature.
I agree.
David Lawyer
