W3C home > Mailing lists > Public > www-style@w3.org > January 2008

Re: flowing around both sides of a float

From: Peter Moulder <Peter.Moulder@infotech.monash.edu.au>
Date: Wed, 02 Jan 2008 02:39:22 +1100
To: www-style@w3.org
Message-id: <20080101153922.GA14009@bowman.infotech.monash.edu.au>

On Mon, Dec 31, 2007 at 02:13:46PM -0800, James Elmore wrote:
>
> On Dec 31, 2007, at 12:57 PM, Brad Kemper wrote:
>
>>
>> On Dec 31, 2007, at 10:29 AM, Ambrose Li wrote:
>>
>>> As in my involvement with Wikipedia, I am very much against this
>>> ["this" presumably meaning the citing of an expert].
>>> Especially in the case of Chinese (but including English), there is
>>> much typographic knowledge that, apparently, has NEVER been
>>> published in any book. Having to cite a reference for such things
>>> is the wrong approach IMHO, it just gives a false sense of authority.

Citing a typographer at least gives evidence of one published
typographer's opinion; it carries more weight than my own opinion about
typographic matters.  I agree that one published typographer's opinion
isn't the final word.

> <rant>
> I have been arguing for a more complete set of abilities in CSS for months, 
> including expanding the use of 'float's. How can we as a group say some 
> effect or control would not be useful in the future?

We make our best guess as to how useful it will be, and let this (with
other things such as implementation difficulty) influence what features
we put the most work into getting to implementation first.

In order to keep CSS reasonably implementable, one might even choose not
to include some features in the spec, or to include features but somehow
privilege some features over others (e.g. by choice of what CSS3 module
for the feature to be in, as has been done with some features being
placed in css3-gcpm).

(The <rant/> is directed at CSS development generally; the rest of my
response is specific to the context of flowing a single block of text
around both sides of a float.)

> Don't argue that no one uses a proposed feature -- no one can until it is 
> documented and implemented.

This is one reason I mentioned print publications: because designers can
already achieve this effect in magazines etc.  (See below.)

> Saying 'no one needs it' or [...] only limits your thinking and slows down
> the process.
> </rant>

Establishing that a feature isn't needed allows redirecting efforts to
other features, thereby speeding up the process.

I agree that it can be useful to give designers free reign in order to
find out what features are useful, but adding things to CSS isn't
necessarily the best way of allowing such experimentation.  Adding the
feature to a drawing program or DTP software (or animation software or
whatever) can be better because e.g.: one needn't consider adaptation;
such software already allows pixel-level control over lots of things;
the editor for the format is the sole renderer; there are many good
typographers who can justify doing lots of experimentation; it's easier
to combine different editing software to produce the final product (e.g.
using a generic graphics editor to modify the output of a DTP program
before re-importing into that DTP program).  And also because in some
cases (including the current case) the behaviour is already available
outside of CSS.

I thought it useful to look at what magazines and high-value newspapers
and advertizing brochures do for the following reasons:

  - Print is almost the ideal medium for using flowing on both sides:
    the designer knows exactly what the sizes are and how big the text
    is.  To the extent that print is the ideal medium, we can say that
    if the feature isn't useful in print then it isn't useful anywhere.

      OTOH, as already noted, print isn't the medium to benefit most
      from the feature in every way:

	- We don't yet know how well columns will work in desktop-style
	  browsers.  Without columns, there aren't as good alternatives
	  for how to float something to the middle of a page, so maybe
	  the feature could be useful in a desktop-style browser even if
	  the greater availability of columns in print media makes the
	  feature not as likely to be used in print.
	
	- Interactivity or animation (especially if the flowed content
	  is pictorial rather than textual) may provide use cases not
	  available in print.

    I believe that both InDesign and QuarkXPress (and Scribus) allow
    flowing a single text block both sides of a float; can anyone
    confirm?  If there are print publications with good typography that
    are already using software that allows doing this, and if those
    publications choose not to use this feature, then it is evidence
    that the feature is not useful or is harmful (at least for print
    media and the content/readership of those publications).
    
  - Looking at print publications gives evidence of the opinion of good
    typographers.

    (That's not to say that all print publications are controlled by
    good typographers, but I think one can choose publications that are
    likely to have good typographical decisions: for example, glossy
    fashion magazines are more likely to evidence good typography than a
    local newspaper or club newsletter.)

> Sorry about the rant, but I hear the same arguments every time a new
> (or even an old) feature is proposed and it gets to me. Didn't anyone
> learn about logic and creative thinking in school?

If I'm sometimes negative about adding features to standards then it's
because I like standards to be, well, standard, i.e. fully implemented.
Adding features to CSS is useless except to the extent that they are
widely implemented.

As a programmer, I know that it takes a long time to implement even very
little things once one includes writing test cases and checking that the
implementation doesn't allow remote exploits (given that CSS is
precisely intended for reading files created by possibly-malicious
strangers).  How many programmer-years does it take to implement even
CSS 2.1 ?  According to
http://en.wikipedia.org/wiki/Comparison_of_layout_engines_%28CSS%29#CSS_version_support,
not a single one of the world's major browser renderers completely
support it, despite the efforts of some of the largest software
companies.  What hope is there for smaller companies to implement even
CSS 2.1, let alone CSS3 ?

pjrm.
Received on Tuesday, 1 January 2008 15:39:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:57 GMT