Re: Marking Up Poetry (Was: Conformance of DL Groups Missing DT or DD)

Dr. Olaf Hoffmann writes:

> [Ian wrote:]
> 
> > On Wed, 27 Feb 2008, Dr. Olaf Hoffmann wrote:
> >
> > > For historical reasons there are mainly to types of text - poetry
> > > and prose.
> >
> > ... terms that are defined by the specification do not take on any
> > meaning from the words used in their name.
> 
> > > It is the typical practice for HTML4 if you do not find an element with
> > > sufficient structure and semantical meaning, choose a closely related
> > > element to get the best what you can get.
> >
> > If that element is "close" but still not a fit, it's not good practice.
> 
> Well, if there is no better, then an author has just the choice
> between another language or to use an element with a related meaning
> and sufficient structure

Yes; it isn't currently possible to mark up poems as poems, so something
else has to be thosen.  There are lots of other things which can't be
marked up with that level of precision.  HTML includes some fairly
'generic' elements which don't impose much meaning (<div>, <br>, <p>)
and so can be used in many situations.

Where a precise element isn't available it makes sense to pick a generic
element (which is merely imprecise), rather than one which is specific
but intended for a different context (which is therefore wrong).

The whole point of <dl> is that it's for a list of pairs -- associating
some sort of 'names' with 'descriptions'.  It's reasonable to expand
this to encompass types of pairs not previously permitted, but it
doesn't make sense to co-opt it for lists of individual items (such as
lines in a poem) -- there are already several other ways of marking up
such lists, and using <dl> like this removes <dl>s very essence.

> - I do not think, this is good at all, it only shows, that the
> language in use is still rudimental.

Yes; HTML is too rudimentary to mark up poems precisely -- among many
other things.

> > > Therefore in HTML4 clearly dl/dt or dl/dd is already the best
> > > choice for most parts of poetry.
> >
> > It really isn't. :-)
> 
> Why it is, is already detailled in the wiki:
> http://esw.w3.org/topic/HTML/PoeticSemantics

Please could you explain that?  I saw that page mentions that <dl> can
be abused for this, but it doesn't read to me as an explanation as to
why it's the best.

Naïvely I'd probably put an entire poem in <blockquote>, with <br> for
line-breaks, and <p> around each verse (for poems with multiple verses),
and reading that page hasn't made it all clear why <dl> would be better.

> Some specific elements would be better, but before they exist, this
> fits best...

I don't think HTML should go to the granularity of specifying usage for
particular fields such as poetry.  But, if we do decide the language
should specifically cater for poems then it would make more sense to add
elements for that purpose.

Using other elements is what authors will have to do if the spec doesn't
cater for poetry, because authors aren't in a position to define new
elmements; but there's no point in the spec going out of its way to
re-purpose elements designed for other things when it could just as
easily create new ones.

> > > As detailled in the wiki the examples in the current HTML5 draft
> > > do not use elements with a sufficient structure and are therefore
> > > bad or poor examples to markup poetry.
> >
> > I don't understand why the examples in the spec are bad.
> 
> See:
> http://esw.w3.org/topic/HTML/PoeticSemantics

Again, I'm afraid that I don't follow.  The page states how poems could
be marked up, and which components a poem has, but I still haven't
understood _why_ poems need to be marked up at that level of
granularity, and in a way which user-agents know they are poems -- what
poetry-specific behaviour would user-agents have if poems were
explicitly marked up?

> > > > Poetry is no more important than stories, addresses, legal
> > > > documents, letters, and any number of other document types, none
> > > > of which have their own section either.
> > >
> > > It is not more important, but it has a richer microstructure.
> >
> > I do not buy that at all. Letter have a complex microstructure. So do
> > dictionaries, so do legal documents, so do addresses, etc... 
> 
> Sure, several of them have similar problems, that can be covered
> sometimes approximately with dl/dt/dd or some other elements and
> some crutches in HTML4 or XHTML1.x, if the meaning and usage of 
> such elements is not restricted by the specification.

But again, what would be done with that structure?  A letter typically
has the sender's address and phone number, the date sent, the
recipient's name, the recipient's address, a salutation, possibly a
heading, some paragraphs of content, a complimentary closing, and the
sender's name.

Currently the <p> element is about the best available for marking up
each of these (with <br>s for line-breaks in addresses).  That seems to
work.  If new elements were invented for each of the above, browsers
would still render them as they currently do <p>, and it's unlikely they
would provide additional functionality for any of them (as suggested for
<dfn> and <mark>).  So what would be the point?

> > Some of these are addressed by Microformats. ... Define a set of
> > class names if you really must deconstruct your poetry more than
> > this
> 
> Because the values of class have no predefined meaning,

Why do they need to be predefined?  You can do whatever you want with
poems you've marked up with your own class names, so the point of
predefined names is for others to do things with them; what do you have
in mind?

Smylers

Received on Wednesday, 27 February 2008 23:37:22 UTC