[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: links in html within a page

From: Graham Percival
Subject: Re: links in html within a page
Date: Wed, 19 Sep 2007 17:10:41 -0700
User-agent: Icedove (X11/20070607)

Karl Berry wrote:
    but after much user consultation we decided to split the lilypond
    manual HTML into chapter/sections, but not subsections.

Far be it from me to question Lilypond users and developers, but this is
entirely baffling to me :).

To be honest, me too -- but approximately half the people who responded
said that they used the all-in-one HTML page instead of the split HTML
pages.  I was stunned that people hated the "one-subsection-per-page"
output so much that they downloaded a 5 megabytes of HTML+images every
time they used the docs.

The biggest complaint was loading time -- specifically, the loading time
of each doc lookup, not the loading time at the beginning.  People
preferred to spend X minutes downloading the entire one-page docs at the
beginning, instead of spending Y seconds to download each individual

This is somewhat a case of "the customer is never wrong" -- it doesn't
make sense to me, but the whole point of the docs is to be read.

Anyway, so what are you really trying to accomplish here?  Why are you
even using @unnumberedsubsubsec?  If you just want a heading within a
section, then use @subsubheading, which won't start a new sectional
unit.  On the other hand, if you do want a new sectional unit, doesn't
that go against your own self-imposed restriction of nothing below
"@section"?  I don't get it ...

The desire is this:

1 Notation
   1.1 Pitches
        1.1.1 Writing pitches
                normal pitches
                more accidentals
                notenames in different languages

where 1.1.1 has a web page which includes the four @unnumberedsubsubsec,
but all of the above is shown in the table of contents.

Perhaps the real answer is to use texi2html for our HTML docs; it claims
to be much more flexible.

I know the Lilypond manual has many exceptional needs.  I've always been
impressed and glad that it uses Texinfo at all.  Perhaps it's time for
the manual should be recast into another form and the PDF/HTML generated
from that, along with a "normal" Texinfo file (using @node!) for those
who actually read the Info output.  Just a thought ...

I may be biased because I've used texinfo so much by now (lilypond and
marsyas), but I think texinfo achieves a great balance between
simplicity and complexity.  I definitely don't want to stop using it; I
(or my users) just want the HTML to be a bit fancier.  I'll try to get
texi2html working with our build system and see what it produces.

- Graham

reply via email to

[Prev in Thread] Current Thread [Next in Thread]