More responses to slashdot comments

Tantek (and others)

As I have too much time on my hands  Another
draft response to  
some /. comments

[SDC=Slashdot comment, MFR=Micrformats response]

SDC: Mixing presentation and data - good... bad... good. But
it gets  
better a little, each time (maybe more of a spiral than a
wheel).
SDC: Ok, so this "microformats" thing is about
encoding extra data  
inside an HTML file by abusing CSS class names for markup,
isn't that  
completly unnecessary and nothing more than an ugly hack?

MFR: Several slashdot comments carry in essence the same
criticism as  
these - that microfomats, abuse the class attribute, and by
doing so  
mix presentation with document structure. This demonstrates
a  
misunderstanding of the class attribute of HTML. The class
attribute  
is very commonly used by web designers in conjunction with
CSS to  
style pages (which is one of the roles outlined for it in
the HTML  
specification [1]), and in truth, it is often overused in
this way.  
But despite that, class, according to the HTML specification
"has  
several roles in HTML", including "for general
purpose processing by  
user agents" [1].
Microformats make use of this second aspect of the class
(and id)  
attribute, and do so legitimately. It is not an abuse of the
class or  
id attribute to use it to add semantic context to a
document. Nor is  
the use of class in and of itself presentational - in fact,
it is an  
important mechanism for separating presentation from
structured content.

[1] http://www.w3.org/TR/html401/struct/global.html#h-7.5.2

john

John Allsopp

style master :: css editor :: http://westciv.com/st
yle_master
blog :: dog or higher :: http://blogs.w
estciv.com/dog_or_higher
WebPatterns :: http://webpatterns.org
Web Directions Conference :: Sydney September 28-29 :: http://wd06.com


_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On Jul 12, 2006, at 5:49 PM, John Allsopp wrote:

> Tantek (and others)
>
> As I have too much time on my hands  Another
draft response to  
> some /. comments
>
> [SDC=Slashdot comment, MFR=Micrformats response]
>
> SDC: Mixing presentation and data - good... bad...
good. But it  
> gets better a little, each time (maybe more of a spiral
than a wheel).
> SDC: Ok, so this "microformats" thing is
about encoding extra data  
> inside an HTML file by abusing CSS class names for
markup, isn't  
> that completly unnecessary and nothing more than an
ugly hack?

FWIW: I wrote a blog post about this: http://microformats.org
/blog/ 
2005/10/19/more-than-styling/

-ryan



_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

While I agree that the use of the class attribute in
microformats is
consistent with the intended use (i.e., to embed semantic
information
into HTML), I think there are some subtle gotchas to
consider.

My recollection is that when span and class were introduced,
there 
were a lot of people who had some unease with the very
existence of 
class, span and div, because it adds an orthogonal layer of
semantics 
onto HTML. 

For example, what is the difference between: 
   a) This is an <em>emphasized</em> point
vs.
   b) This is an <span
class="emphatic">emphasized</span>
point

Presumably, (a) would be preferred over (b), although I
personally
prefer (b). (I think strong and em are copouts, but that's
a whole other
story).

It gets even more confusing when you mix tags and classes
more
aggressively.

Example A:
--
<div class="hreview">
 <span><span
class="rating">5</span> out of 5
stars</span>
 <h4 class="summary">Crepes on Cole is
awesome</h4>
 ...
</div>

Is the h4 the "structure" that defines the
semantics of the document, or
is the "summary" class the structure that
defines the semantics? 

In other words, why not do this:

Example B:
--
<div class="hreview">
 <span><span
class="rating">5</span> out of 5
stars</span>
 <span class="summary">Crepes on Cole is
awesome</span>
 ...
</div>

Ostensibly, the reason for preferring example A over example
B is that
there are two systems of semantics for two different
reasons. The HTML
view of the document says that the header is an h4. The
microformats
view of the document says it's a "summary".
Example A allows both views
of the document to live together.

What's the problem with this? Well, you'll notice that the
HTML document
has become polluted by the div and span tags. Are the span
and div tags
part of the HTML semantics, or are they just scaffolding to
hold the
microformat information? In other words, should we ignore
the span and
div tags when interpreting our document as HTML? 

This is essentially the same issue that people have with
hacks like
adding <div class="topleft"> into a
document, which is clearly
presentational. While the microformat tags are not
presentational, they
are a different, orthogonal set of semantics that is being
embedded into
the document, which can make it hard to know how to
interpret the
semantics of the HTML tags.

Anyway, I think the above arguments are not a reason to
discount
microformats -- the benefits outweigh the costs. But I do
think there is
a valid point to consider here, which is that mashing two
(or more)
systems of semantics together comes at a cost.

-Sho


 

-----Original Message-----
From: microformats-discuss-bouncesmicroformats.org
[mailto:microformats-discuss-bouncesmicroformats.org] On Behalf
Of John
Allsopp
Sent: Wednesday, July 12, 2006 5:50 PM
To: microformats-discussmicroformats.org
Subject: [uf-discuss] More responses to slashdot comments

Tantek (and others)

As I have too much time on my hands  Another
draft response to  
some /. comments

[SDC=Slashdot comment, MFR=Micrformats response]

SDC: Mixing presentation and data - good... bad... good. But
it gets  
better a little, each time (maybe more of a spiral than a
wheel).
SDC: Ok, so this "microformats" thing is about
encoding extra data  
inside an HTML file by abusing CSS class names for markup,
isn't that  
completly unnecessary and nothing more than an ugly hack?

MFR: Several slashdot comments carry in essence the same
criticism as  
these - that microfomats, abuse the class attribute, and by
doing so  
mix presentation with document structure. This demonstrates
a  
misunderstanding of the class attribute of HTML. The class
attribute  
is very commonly used by web designers in conjunction with
CSS to  
style pages (which is one of the roles outlined for it in
the HTML  
specification [1]), and in truth, it is often overused in
this way.  
But despite that, class, according to the HTML specification
"has  
several roles in HTML", including "for general
purpose processing by  
user agents" [1].
Microformats make use of this second aspect of the class
(and id)  
attribute, and do so legitimately. It is not an abuse of the
class or  
id attribute to use it to add semantic context to a
document. Nor is  
the use of class in and of itself presentational - in fact,
it is an  
important mechanism for separating presentation from
structured content.

[1] http://www.w3.org/TR/html401/struct/global.html#h-7.5.2

john

John Allsopp

style master :: css editor :: http://westciv.com/st
yle_master
blog :: dog or higher :: http://blogs.w
estciv.com/dog_or_higher
WebPatterns :: http://webpatterns.org
Web Directions Conference :: Sydney September 28-29 :: http://wd06.com


_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

--- Sho Kuwamoto <skuwamotadobe.com> wrote:

> Example A:
> --
> <div class="hreview">
>  <span><span
class="rating">5</span> out of 5
> stars</span>
>  <h4 class="summary">Crepes on Cole
is awesome</h4>
>  ...
> </div>

[snip]

> Example B:
> --
> <div class="hreview">
>  <span><span
class="rating">5</span> out of 5
> stars</span>
>  <span class="summary">Crepes on Cole
is awesome</span>
>  ...
> </div>
> 
> Ostensibly, the reason for preferring example A over
> example B is that
> there are two systems of semantics for two different
> reasons. The HTML
> view of the document says that the header is an h4. The
> microformats
> view of the document says it's a
"summary". Example A
> allows both views
> of the document to live together.


<span>s and <h4>s are not structurally
equivalent.  span
and div tags are general structural markup, while heading
tags are specifically defined in relation to other heading
tags.  Collectively they define an outline for the page,
while the set of spans on a page defines nothing.

Sure you can create a stylesheet that instructs a browser
to render the two in the same way, but you are sending a
document with a different structure.  What if the user has
other ideas on rendering heading tags, wants to extract a
page outline programatically, or isn't using a CSS
compliant browser?  In those cases, the structual
differences make a pragmatic, presentational difference
that can affect usability.


> This is essentially the same issue that people have
with
> hacks like
> adding <div class="topleft"> into a
document, which is
> clearly presentational. 

That may be an unfortunate choice of class name, but one of
the uses of class names is as a style sheet selector[1] so
I hardly see that as a hack.  Web pages have an inherent
presentational aspect and coding specifically to address
that isn't something to discourage.

Should we aspire to publishing elegant HTML?  Sure, but
it's not the only goal or, I'd say, even the primary one.

-ml

[1] http://www.w3.org/TR/html401/struct/global.html#h-7.5.2
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On Jul 13, 2006, at 11:49 AM, Sho Kuwamoto wrote:

> For example, what is the difference between:
>    a) This is an <em>emphasized</em> point
> vs.
>    b) This is an <span
class="emphatic">emphasized</span>
point
>
> Presumably, (a) would be preferred over (b), although I
personally
> prefer (b). (I think strong and em are copouts, but
that's a whole  
> other
> story).

No need to presume here.  It's directly stated on the
process page:

http://microformats.org/wiki/process#Propose_a_Microfo
rmat

----
There are other things to try before developing a
microformat. First,  
ask yourself these questions:

    1. Is there a standard element in XHTML that would work?
    2. Is there a compound of XHTML elements that would
work?
    3. Ok, if the answer to the above two is 'no,' we can
talk about  
a microformat.
----

XHTML semantics take precedence over microformat semantics.

> Ostensibly, the reason for preferring example A over
example B is that
> there are two systems of semantics for two different
reasons. The HTML
> view of the document says that the header is an h4. The
microformats
> view of the document says it's a
"summary". Example A allows both  
> views
> of the document to live together.

I think we prefer XHTML semantics over microformat semantics
because  
semantics only become useful when understood and XHTML is
widely  
understood.

> Are the span and div tags
> part of the HTML semantics, or are they just
scaffolding to hold the
> microformat information?

They're part of HTML.  They're just parts that don't mean
much.

> In other words, should we ignore the span and
> div tags when interpreting our document as HTML?

Until we have some need to communicate that something spans
or  
divides, yes, we should ignore those semantics.  If they
become  
useful, we should use them.

> This is essentially the same issue that people have
with hacks like
> adding <div class="topleft"> into a
document, which is clearly
> presentational.

Not necessarily.  Consider the following:

<h1>For sale: used monitor</h1>
<div class="description">
	<div class="bottom">The bottom of the
monitor is in great  
condition.</div>
	<div class="topleft">There is a slight
discoloration in the top-left  
corner.</div>
</div>

The meaning of "topleft" here has nothing to do
with the presentation  
of the HTML document. The meanings of class attribute values
are not  
pre-defined. The meanings of XHTML tags are pre-defined.

Peace,
Scott

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Michael Leikam wrote:
> <span>s and <h4>s are not structurally
equivalent.  span
> and div tags are general structural markup, while
heading
> tags are specifically defined in relation to other
heading
> tags.  Collectively they define an outline for the
page,
> while the set of spans on a page defines nothing.

Exactly my point. There are two competing schema living in
the same
document: the world of HTML (semantically poor and
unextensible), and
the world of microformats. While this works out OK usually,
I believe
there are cases where the two worlds combine in
uncomfortable ways.

BTW, I don't know if saying that spans "define
nothing" in the HTML
world fits with my view of the things. To me, spans and divs
in HTML are
an escape mechanism that allows me to add my own meaning
above and
beyond what HTML provides. For example, I might have an
"appendix"
section, which I think of as "containing" other
elements, such as h3,
etc. 

> > This is essentially the same issue that people
have with
> > hacks like
> > adding <div class="topleft">
into a document, which is
> > clearly presentational. 
> 
> That may be an unfortunate choice of class name, but
one of
> the uses of class names is as a style sheet selector[1]
so
> I hardly see that as a hack.  Web pages have an
inherent
> presentational aspect and coding specifically to
address
> that isn't something to discourage.

Well, I guess my viewpoint is that adding span tags with
classes into
your HTML purely for the sake of having a place for CSS to
attach itself
is, in a way, adding presentation to your document. In an
ideal world,
the divs and spans would be added to your document to
represent
structure, and CSS would be applied later to make it look
right.

For example, I might have a page that has four sections:
"mainNav",
"adSection", "links", and
"main". Creating divs for these sections seems
fine. I'm just marking out parts of my document that have
meaning. 

Depending on the look I wanted to achieve, I might find
myself needing
to surround, say, the first three divs by another div
(let's call it
"leftColumn" because there is no semantic
relationship between these
three sections). That, to me, feels like adding markup to
your document
for the sake of presentation.


Anyway, getting back to my earlier point... I think the pros
of
microformats outweigh the cons, so I'm not arguing against
microformats.
I'm just saying that when people look at all the spans and
divs with
classes on them and wonder if that will complicate things...
well... I
think it's a fair point. It probably does slightly
complicate things.

-Sho
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On Jul 13, 2006, at 3:17 PM, Sho Kuwamoto wrote:

> Depending on the look I wanted to achieve, I might find
myself needing
> to surround, say, the first three divs by another div
(let's call it
> "leftColumn" because there is no semantic
relationship between these
> three sections).

Why isn't "leftColumn" a semantic relationship?
 It means something,  
doesn't it?  There's no reason a bunch of designers
couldn't get  
together and agree on what "leftColumn" means
just like the W3C  
agreed on what "h1" means, and the microformats
community agreed on  
what "vevent" means.  Until someone find a
babblefish, that shared  
meaning is all we have to communicate.  That's as semantic
as it gets.

Peace,
Scott

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

On 7/13/06, Sho Kuwamoto <skuwamotadobe.com> wrote:
> Exactly my point. There are two competing schema living
in the same
> document: the world of HTML (semantically poor and
unextensible), and
> the world of microformats. While this works out OK
usually, I believe
> there are cases where the two worlds combine in
uncomfortable ways.

There have to be two types of schema because there are, as
you rightly
said, two orthogonal sets of semantics in HTML.

The first is the tag/attribute based semantics, which are
very
strictly defined by the W3C spec.  These are mainly to do
with
document structure and so on, and everyone understands what
they mean.

The second set of semantics are class/id based, and are
completely
'unregulated', that is to say the specific meanings
aren't specified
by the HTML spec.  If I want id="shopping-list" then there's no
reason I shouldn't mark my pages up that way, and there's
some
semantic value in doing so over something like id="centre-column".

Microformats form conventions for how the *both* sets of
semantics
should be used.  Microformats will, by preference, use the
first set
as far as possible (i.e. using ADDR in hCard) and then
define sensible
semantic ids/classes for stuff that isn't covered.

Microformats differ from schemas like the W3C's HTML spec,
because
pages don't have a mechanism for declaring that they
conform to a
specific microformat.  I don't think this is so much a
weakness as a
strength!

Take rel="tag" for instance.  The microformat
for this declares very
specifically what semantics we can read from the
relationship between
the current page and the URL being linked. However, there's
nothing to
stop someone who's never heard of microformats deciding to
use
rel="tag" on one of their pages, because
it seems a sensible value to
use, and you can't tell my looking at a page whether the
author had
the microformat in mind or not.

I believe that the strength of microformats is that they are
always
sensible markup, so it doesn't matter if someone knows
about the
microformat being used or not, the markup still makes sense
to them -
if I'm looking at a link and see rel="tag" in
there then that's not
cryptic - I can understand what the link is saying even if I
haven't
heard of the microformat.

The converse of this is that if I build a parser that
understands
rel="tag" into my search engine, then I
have a spec that tells me a
sensible way to parse and understand the semantics of the
link.  When
my search engine finds the hypothetical page above, that
uses
rel="tag" without knowing the
microformat, then because the spec
defines a sensible way of parsing it, my search engine will
have a
good chance of correctly understanding what the link
relationship
means.

-Ciaran
_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss

Scott Reynen wrote:
> Why isn't "leftColumn" a semantic
relationship?  It means something, 
> doesn't it? 

It has meaning, but doesn't give meaning to anything else. 
"Semantic" 
would indicate that it told you something 'about' the
meaning the 
content, not where it is located.

You might say that term has a "Geographic"
relationship to the rest of 
the content.  But given that each user agent displays
content 
differently, and that the CSS might actually contain
.leftColumn (position:absolute;bottom:0;right:0;}
it doesn't necessarily indicate that.


Atamido

_______________________________________________
microformats-discuss mailing list
microformats-discussmicroformats.org
http://microformats.org/mailman/listinfo/microforma
ts-discuss