Re: [WSG] Tag clouds spit /

2006-03-14 Thread Ben Ward
On 3/14/06, James Ellis [EMAIL PROTECTED] wrote:
 I've often wondered why these so-called tag clouds don't use big and
 small tags (maybe they do ?)

 they are inline tags. they convey meaning : 2 bigs is more important
 than 1 big etc.

The problem there is that big and small tags _don't_ mean
anything. They're semantically void just like span or font.
They're purely presentational [1]

On 3/14/06, Ben Buchanan [EMAIL PROTECTED] wrote:
 Technorati's multiple-emphasis method is interesting, but I'm not sure
 if the semantics are entirely sound (can you really just keep adding
 more emphasis?). Given that current user agents don't really do
 anything with the tag combinations you are ultimately left with a
 visual-only cue to importance.

I don't see why you _can't_ nest them, but it is unspecified
behaviour. What do you mean by 'User Agents doing something with the
tag combinations'? Is there anything you want a UA to do that isn't
visual or aural that can't be described in a stylesheet?

I'd be inclined to either nest em and strong phrase elements, or to
group the tag importance into just three levels for markup, and add
additional granularity via classes.

So:
* Normal Importance: span, span class=important1, span
class=important2
* Higher Importance: em, em class=important1, em class=important2
* Highest Important: strong, strong class=important1, strong
class=important2

As such, with raw HTML you're describing three different levels of
importance, and with CSS you can enhance that extra levels as required
with classes. Ultimately, so long as the mark-up is in some way
logical then you've probably already achieved 'optimal' mark-up within
the limitations of HTML.

Ben

[1] http://www.w3.org/TR/html4/present/graphics.html#h-15.2.1


Re: [WSG] Tag clouds spit /

2006-03-14 Thread Micky Mourelo
Tag clouds are, almost always, a list of links ordered alphabetically, in which the more whatever, the bigger in font size. So it seems logical to use a simple UL with As in every LI.You are right about semantic meaning on presentation, but you could add the frequency to the tag {Japan (14798) Kyoto (12585)}; you could even hide it through css (I do not sse the problem here, after all font size would be representing the frequency) You're example is not bad as it is really a table what you are coding, but you are showing less information trough presentation, why hide data? as I understand it you have 3 columns but through css you only show info regarding 2 columns.
If your data consisted of only 2 columns, would you still use a table?My point is, tags usually only carry two data, tag and frequency, so a UL or DL often do. In your case a table might seem more correct only if you really need those 3 columns.
Please no BIG or SMALL, we are going forward not backwards.If you show the frequency number you have more semantics than strongs or ems will ever have, so using classes would be the way to go, and, as someone pointed out, you need more than two levels of magnitude on tag clouds; otherwise they would not look like clouds,



[WSG] Tag clouds spit /

2006-03-13 Thread Jude Robinson

Leaving aside arguments about whether or not tag clouds are the new
mullets (http://www.zeldman.com/daily/0405d.shtml), I've been
pondering how to make one that is accessible, given the questionable
accessibility of most of the ones out there.

Tag clouds suffer from a number of problems, the biggest one being that
they rely on style to provide meaning - shade (or colour) and
font-size are used to provide information about frequency and freshness.
This is obviously *bad* as it does not account for non-CSS user
agents.

There are problems even for users of CSS supporting user agents - whilst 
it is easy to infer some meaning from the size of text (ie: bigger = 
more frequent), it is not easy in this case to infer meaning from a 
change in colour (which equals more fresh: red or blue?)


Wah wah wah. Check out this (very) basic demo (content borrowed from 
http://www.connotea.org/): http://www.dotcode.com/accessible_tag_cloud.html


Presenting it in a table is as far as I'm aware the only accessible way 
to do it, as using the title attribute or enclosing the information in 
the link text itself (then hiding it using CSS) both have serious drawbacks.


Which means there are (surprisingly) problems for IE.

Thoughts? Mild agreement? Sly refutations?

**
The discussion list for  http://webstandardsgroup.org/

See http://webstandardsgroup.org/mail/guidelines.cfm
for some hints on posting to the list  getting help
**



Re: [WSG] Tag clouds spit /

2006-03-13 Thread Christian Montoya
On 3/13/06, Jude Robinson [EMAIL PROTECTED] wrote:
 Leaving aside arguments about whether or not tag clouds are the new
 mullets (http://www.zeldman.com/daily/0405d.shtml), I've been
 pondering how to make one that is accessible, given the questionable
 accessibility of most of the ones out there.
...
 Wah wah wah. Check out this (very) basic demo (content borrowed from
 http://www.connotea.org/): http://www.dotcode.com/accessible_tag_cloud.html

 Presenting it in a table is as far as I'm aware the only accessible way
 to do it, as using the title attribute or enclosing the information in
 the link text itself (then hiding it using CSS) both have serious drawbacks.
...
 Thoughts? Mild agreement? Sly refutations?

I think this still suffers from some of the other problems with tag
clouds. For one thing, there's still too many links, many of which
lead to duplicate content (people usually tag the same entry with
multiple tags). Also, it would make more sense in regards to
accessibility  if the most frequent tags were first, and the (ordered)
list descended by frequency. I don't see a table as a good use since
the tabular data is all hidden. The tag cloud is still sorted
alphabetically which isn't too useful as far as tag clouds go.

So, I think a better option would be an ordered list starting with the
most frequent tag and descending from there. The title of the list
could summarize this information and the class could be something like
tag_frequency_list. It wouldn't look like a typical tag cloud but
those are ugly anyway (just like mullets).

--
--
Christian Montoya
christianmontoya.com ... rdpdesign.com ... cssliquid.com
**
The discussion list for  http://webstandardsgroup.org/

 See http://webstandardsgroup.org/mail/guidelines.cfm
 for some hints on posting to the list  getting help
**



Re: [WSG] Tag clouds spit /

2006-03-13 Thread Ben Buchanan
[snip- tag clouds]
 Thoughts? Mild agreement? Sly refutations?

For a simple tag cloud I'm not sure how a table would work - the cloud
isn't really tabular, in the way I would describe tabular. If a
second rating was applied; eg. popularity + freshness (ie. how many
times does the tag appear, and how recently) then you could use a
table to create a matrix. But for your garden variety tag cloud, this
probably isn't applicable.

Technorati's multiple-emphasis method is interesting, but I'm not sure
if the semantics are entirely sound (can you really just keep adding
more emphasis?). Given that current user agents don't really do
anything with the tag combinations you are ultimately left with a
visual-only cue to importance.

I'm with Christian on this one - I'd go with an ordered list with the
most popular tags first. You could use a nested list to group together
bands or popularity levels; or use an ordered list for the bands
with an unordered list for all the equal weight items in that band.

cheers,

Ben

--
--- http://www.200ok.com.au/
--- The future has arrived; it's just not
--- evenly distributed. - William Gibson
**
The discussion list for  http://webstandardsgroup.org/

 See http://webstandardsgroup.org/mail/guidelines.cfm
 for some hints on posting to the list  getting help
**



Re: [WSG] Tag clouds spit /

2006-03-13 Thread jacinta
This message has been returned to you because I will be on leave from Wed 
15/3/2006 to Friday 17/3/2006. During my absence, please contact Judy Hageman 
on 72224
([EMAIL PROTECTED]).


**
The discussion list for  http://webstandardsgroup.org/

 See http://webstandardsgroup.org/mail/guidelines.cfm
 for some hints on posting to the list  getting help
**