[Bug 558] - paragraph tag needs formatting structure

[bugzilla-daemon@ecoinformatics.org]

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=558

jones at nceas.ucsb.edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|eml-cvs at ecoinformatics.org  |eml-dev at ecoinformatics.org



------- Additional Comments From jones at nceas.ucsb.edu  2002-08-26 14:59 -------
Tim, so I looked over your suggestions.  Good overview. Thanks.  Here's my
take...  Either we can go simple and include only a very few elements for
paragraph structuring, or we can go complex and let most of docbook in.  But a
middle ground is the worst case.  Here's why...

If we go simple, then we'll only need to define a text container type of tag
that can be used throughout EML, and we can make its few elements correspond to
docbook or XHTML (e.g., para, ol, ul, li, emphasis).  This is easy to implement,
and simple so easy to understand for EML authors and EML readers. And it would
be a subset of docbook so the translation would be 1:1 to docbook.

If we go complex, we could wholesale import tags from the docbook namespace. 
nobody will understand the tags (there are far too many), but we can cut and
paste from tools that output docbook format, and we can utilize already existing
stylesheets that know how to format & present docbook.  This is a far more
complex but also more flexible solution.

If we go the middle road, then there are more tags than a user can understand,
and we lose the ability to wholesale import an existing standard (so off the
shelf stylesheets and software won't be usable, and its a pain to maintain). 
Worst case scenario to me.

In looking at what you think is needed/not needed, it seems that the following
tags would be very useful:

ItemizedList, OrderedList, ListItem; (corresponds to ul,ol,li)
emphasis; (no italics in DocBook, predictably.)
procedure, step, substep; (Great!  substeps nestable to any depth!)

You also say that sections would be useful, but aren't allowed in paragraphs in
docbook:
section, title; (sections are great, but unfortunately cannot be in paragraphs.)

I would argue that the best course of action is for us to create a single
generic container for structured text that is mixed content and can be re-used
throughout EML, and is a direct subset of docbook.  We might call it 'textBlock'
or something like that, and it could contain titles, sections, and paragraphs. 
paragraphs could then in turn contain text, lists, and emphasis.

This would allow most of the formatting people would want, without importing the
whole docbook namespace.  The simplest case would devolve to
<textBlock><para>This is some text<para></textBlock>, and so is basically what
we have now.  This has the disadvantage of being non-standard formatting, so
requires us to have custom stylesheets for it (which would be a simple 1:1
mapping to docbook).  But then, almost all of EML is non-standard and requires
custom stylesheets, so that's not really an additional burden.  I'll go ahead
and create a candidate schema to demonstrate this proposal.  What do you think?