forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Johannes Schaefer <johannes.schae...@uidesign.de>
Subject semantic markup vs. WYSIWYG (was: [ANN] odt plugin)
Date Tue, 30 May 2006 08:51:35 GMT
Ross Gardler wrote:
> Ferdinand Soethe wrote:
>> Ross Gardler wrote:
>>
>>
>>> I'll certainly consider switching to a -0 if people want to see this,
>>> but I strongly caution against it. In my opinion it undermines the whole
>>> purpose of Forrest.
>>
>>
>> It is a slippery slope.

Yes, it is: Semantic/logical vs. presentational/physical markup.
See below and
http://en.wikipedia.org/wiki/Comparison_of_document_markup_languages#Characteristics

>>
>> Otoh there are many ways a developer can bend Forrest out of shape so
>> why not leave it up to them and their clients to decide.
>>
>> I guess if it defaults to _not_ carry over styles and the switch carries
>> a sufficient warning in place that should be ok.
> 
> Yes, it sounds reasonable. And Thorsten was careful to have a very clear
> warning in the contract description.
> 
> However, I'm still concerned of posts to the user list of the form:
> 
> "My section is not appearing in the TOC"
> "Everything looks fine in HTML but in PDF sections get all screwed up"
> "My headings have an inconsitent format"
> etc.

"Section" and "Heading" are semantic terms, i.e. we (the
users) talk about the meaning of the marked-up text.

Not appearing in the TOC e.g. means that they made them
*look* like section/heading but did not mark them up correctly.

X "looks fine" in Y but not in Z: may mean that Z does not
support the notion of X but Y does. What it usually means
is that the markup was fine-tuned for the *Look* not the
meaning, e.g. "My web page looks fine in MSIE but not in Z".

> The potential problems are massive, and since we claim to give a uniform
> output from inputs we are in danger of failng to deliver what we claim
> to do so. Sure the user should know what is going on, after all there is
> a warning in the contract description. But we all know users ignore
> things they do not understand and go ahead anyway. It works on the first
> test document so they spen time building a full site and then *bang*
> half the documents don't work.

We run a danger with WYSIWYG-input: users may use presentational
markup (bold, 24pt and these things) instead of semantic like
"section" and "heading".

I always thought of the Word/OOo (and now ODf) is to import
"legacy" documents, not to encourage users to write new content.
(OK, this is a radical point of view).

We have this in some of the plugins, e.g. OOo, Excel.
There was a discussion about carrying over "style"
from such plugins, see
  http://marc.theaimsgroup.com/?l=forrest-dev&m=110916000908143&w=2

If we allow "style" to carry through to the result this needs
to be implemented in all (!?) output formats Forrest supports
(at least users may request it; I requested it once for cell
coloring in PDF).

As for the Excel-Plugin: I ignored any styling done in Excel
(note, I introduced bg-color from the content of the cells).
And still, it's just a special-case input vehicle for not-XML-
savvy users. Others use a XML editor for "near WYSIWYG" (see
http://forrest.apache.org/docs_0_70/catalog.html).

To use these input-plugins "correctly" they need good
documentation, as e.g. for OOo-HowTo
  http://tinyurl.com/ooqcw
instructing them how to use OOo to write "proper" new documents
that work seamlessly with Forrest.

What do you think?
Johannes


> 
> Ross
> 

-- 
User Interface Design GmbH
Teinacher Str. 38, 71634 Ludwigsburg
Fon:      +49-7141-37700-0
Fax:      +49-7141-37700-99
Email:    jschaefer@uidesign.de
Internet: www.uidesign.de

Geschäftsstellen:
Teinacher Str. 38,    71634 Ludwigsburg
Truderinger Str. 330, 81825 München
Friedrichsring 46,    68161 Mannheim

Buch "User Interface Tuning" von Joachim Machate & Michael Burmester
www.user-interface-tuning.de

Attraktivität von interaktiven Produkten messen mit
www.attrakdiff.de

Mime
View raw message