Problems Starting an Ordered List at Zero

[Rijk van Geijtenbeek]

Re: Problems Starting an Ordered List at Zero

On Sun, 26 Sep 2004 19:28:15 -0400, Stan Brown
<the_stan_brown @fastmail.fm> wrote:[color=blue]
> "Jukka K. Korpela" <jkorpela@cs.tu t.fi> wrote:[/color]
[color=blue][color=green]
>> The <ol> element is basically the same as <ul> but with different
>> default rendering. That is, their difference is presentational rather
>> than semantic.[/color]
>
> Hmm, I wonder.
>
> Never mind CSS for a minute; consider just generic documents. My
> take on a bulleted list is that the items are equal and could just
> as well appear in any order; my take on a numbered list is that theh
> numbers are steps in a procedure.
>
> I agree with you that the numbers themselves re not terribly
> important, but the fact that "microwave for 10 minutes" comes before
> "puree with a blender" _is_ important.[/color]

Yes, but you might note recipes often *don't* use numbered
(lists|paragrap hs), and rely on the presentation order to convey the
sequential relation between the steps.

I think in many, maybe even most cases, of UL on the web, people *do* find
the order of their list items important or at least non-arbitrary, but
they don't necessarily want the visually distracting numbering to show. It
is only when the items need to be referenced from elsewhere that numbering
becomes really useful.

--
Rijk van Geijtenbeek

The Web is a procrastination apparatus:
It can absorb as much time as is required to ensure that you
won't get any real work done. - J.Nielsen

[Neal]

Re: Problems Starting an Ordered List at Zero

On Sun, 26 Sep 2004 19:28:15 -0400, Stan Brown
<the_stan_brown @fastmail.fm> wrote:
[color=blue]
> "Jukka K. Korpela" <jkorpela@cs.tu t.fi> wrote in
> comp.infosystem s.www.authoring.html:[color=green]
>> The <ol> element is basically the same as <ul> but with different
>> default
>> rendering. That is, their difference is presentational rather than
>> semantic.[/color]
>
> Hmm, I wonder.
>
> Never mind CSS for a minute; consider just generic documents. My
> take on a bulleted list is that the items are equal and could just
> as well appear in any order; my take on a numbered list is that theh
> numbers are steps in a procedure.
>
> I agree with you that the numbers themselves re not terribly
> important, but the fact that "microwave for 10 minutes" comes before
> "puree with a blender" _is_ important.[/color]

Yet I know of no UA which will randomly display ul items but still put ol
items in the strict order they were put in by the author. In reality,
whether you use ul or ol, the items will be listed in the order you put
them. So the ol and ul markup gives a presentational difference to the
user.

If the ordering of the list items is critical, should we rely on HTML
rendering to add the right numbers? W3C says that "ordered and unordered
lists are rendered in an identical manner except that visual user agents
number ordered list items." It would seem a critical requirement of a
conforming UA to add numbering to ol element items. So it would seem
mandatory that ol items will have a numbering according to how we order
them in the source.

To me, it comes to this question: does ol and ul provide any different
semantic meaning in and of themselves, without considering the
presentation? It would seem there is an argument for either position. To
one, the specific markup defines the information as being inherently
ordered or not. To another, the content is in any event ordered as a rule,
and thus the reader will gather an ordering even if we specify ul.

I recall a discussion where it was posited that <l> would be a superior
element to ol and ul. One could use <l type="ordered"> or what have you to
deliver this information to the UA. Whether this eliminates any
presentational demands on the list elements is a matter of debate, but it
would certainly streamline the HTML element set.

Elements which require a specific rendering seem to be the red-headed
stepchildren of HTML, but some are useful. b and i have valid semantic
uses. hr can be useful in some instances. The list elements may be seen as
useful or not, but it remains that, for right or wrong, they are in HTML,
they do have mandated rendering rules, and authors can expect the UA to
respect this.

So, the real question isn't whether ol and ul are presentational or
semantic, as they clearly can be either. It's whether we're using these
elements more for their presentational traits or for their semantic
meaning to the document.

[Joel Shepherd]

Re: Problems Starting an Ordered List at Zero

In article <Xns95712C5CEC3 1jkorpelacstutf i@193.229.0.31> ,
"Jukka K. Korpela" <jkorpela@cs.tu t.fi> wrote:[color=blue]
>
> The <ol> element is basically the same as <ul> but with different default
> rendering. That is, their difference is presentational rather than
> semantic.[/color]

This is _very_ hard to swallow.

There is an enormous semantic difference between information with some
kind of ordering characteristic to it, and information without.

E.g., if I'm describing the sequence of events, the chronological
_order_ that they occurred in may be critical to understanding how they
are related. With an unordered list, the semantics are that the members
of the list may be re-arranged in any order presentationall y, without
changing the meaning or correctness of the information. An ordered list
-- even without presentational hints -- is indicating that the order of
the elements presented is important: that changing the order of the
elements with change the meaning of the overall list itself.

Your statement seems to me to say that <p> and <blockquote> should also
be considered to have no difference other than presentational, which, in
the default case is usually to be a little different. But they're both
just blocks of text, right?

I'd be surprised if you accepted that interpretation of <p> and
<blockquote>, so it also surprises me that you see nothing but
presentational differences between <ul> and <ol>.

--
Joel.

[Sam Hughes]

Re: Problems Starting an Ordered List at Zero

Lauri Raittila <lauri@raittila .cjb.net> wrote in
news:MPG.1bc168 0dcbf5e21798a02 7@news.individu al.net:
[color=blue]
> <list><li><mark er>0.</marker>First Item</li></list>[/color]

I was actually going to suggest something like this, before reading your
post. A document's textual content should not be hidden away in some
element. To be persnickety, I would use

<list><li><mark er>0</marker>. First Item</li></list>

because the dot is not part of the number, it is just punctuation. Of
course, dots _could_ be part of a number: in a list of rules, one could
have rule 2.3 and rule 2.4.

Alternately, in designing this hypothetical language, I might allow the
marker to contain non-numerical content and allow the author to put markup
_inside_ the marker to indicate the actual number. For example:

<list>
[...]
<li><marker>Rul e <num>2.3</num>:</marker> No blah blah may bladiddy blah
when the ball lands on the line.</li>
[...]
</list>

Of course, some "nums" could have letters in them, such as if one were
referring to "Rule <num>17-C</num>". Or if one used hexidecimal numbers in
the ordered list, for some reason. Perhaps the LIST element should have an
attribute which describes the meaning of the NUM elements' content. Such
an attribute would be inherited from a parent LIST element if not otherwise
specified, if such a parent exists.

[Toby Inkster]

Re: Problems Starting an Ordered List at Zero

Neal wrote:
[color=blue]
> Yet I know of no UA which will randomly display ul items but still put ol
> items in the strict order they were put in by the author.[/color]

No, but in doing so they would not be incorrect IMHO. A UA could, for
example, have an option to display all unordered lists in alphabetical
order instead of the author's order.

(Incidentally, I would like to see such a feature -- and also a feature
that allows you to right-click on TH elements to sort a table by that
column.)

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact

[Lauri Raittila]

Re: Problems Starting an Ordered List at Zero

In article <Xns9570DCAEE60 BBhughesrpiedu@ 130.133.1.4>, hughes@rpi.edu
says...[color=blue]
> Lauri Raittila <lauri@raittila .cjb.net> wrote in
> news:MPG.1bc168 0dcbf5e21798a02 7@news.individu al.net:
>[color=green]
> > <list><li><mark er>0.</marker>First Item</li></list>[/color]
>
> I was actually going to suggest something like this, before reading your
> post. A document's textual content should not be hidden away in some
> element. To be persnickety, I would use
>
> <list><li><mark er>0</marker>. First Item</li></list>
>
> because the dot is not part of the number, it is just punctuation.[/color]

No, it is not part of number, put it is part of list item marker. That
period doesn't belong to list item.
[color=blue]
> Of
> course, dots _could_ be part of a number: in a list of rules, one could
> have rule 2.3 and rule 2.4.[/color]

True. And these are currently hard to get using HTML/CSS.
[color=blue]
> Alternately, in designing this hypothetical language, I might allow the
> marker to contain non-numerical content and allow the author to put markup
> _inside_ the marker to indicate the actual number. For example:[/color]
[color=blue]
> Of course, some "nums" could have letters in them, such as if one were
> referring to "Rule <num>17-C</num>". Or if one used hexidecimal numbers in
> the ordered list, for some reason. Perhaps the LIST element should have an
> attribute which describes the meaning of the NUM elements' content. Such
> an attribute would be inherited from a parent LIST element if not otherwise
> specified, if such a parent exists.[/color]

I think that num element whitout attribute to tell it's type would be
useless, as people shoud understand what kind of number it is anyway (if
they don't, that is problem outside markup...), but computers could get
something from num element.



--
Lauri Raittila <http://www.iki.fi/lr> <http://www.iki.fi/zwak/fonts>

[Jukka K. Korpela]

Re: Problems Starting an Ordered List at Zero

Joel Shepherd <joelshep@ix.ne tcom.com> wrote:
[color=blue]
> In article <Xns95712C5CEC3 1jkorpelacstutf i@193.229.0.31> ,
> "Jukka K. Korpela" <jkorpela@cs.tu t.fi> wrote:[color=green]
>>
>> The <ol> element is basically the same as <ul> but with different
>> default rendering. That is, their difference is presentational
>> rather than semantic.[/color]
>
> This is _very_ hard to swallow.[/color]

Try it with some spicy CSS! :-)
[color=blue]
> There is an enormous semantic difference between information with
> some kind of ordering characteristic to it, and information without.[/color]

But both <ul> and <ol> are inherently ordered - the order of the <li>
elements is significant in both of them. As usual, the HTML
specifications are not explicit in such semantics, but this is surely the
intended and generally understood semantics. It would not be appropriate
for a browser to render (at default settings) <li> elements in a
randomized order, or in any order different from the order in the HTML
source.
[color=blue]
> E.g., if I'm describing the sequence of events, the chronological
> _order_ that they occurred in may be critical to understanding how
> they are related.[/color]

Surely. But the only difference between <ul> and <ol> is that by using
<ol> you are asking the browser to add numbers to the items (as you could
do in CSS: ul { list-style-type: decimal; }, for example). It's
comparable to adding numbers to headings - often a desirable thing to do,
but not possible in practice without actually changing the content of
heading elements.
[color=blue]
> With an unordered list, the semantics are that the
> members of the list may be re-arranged in any order presentationall y,
> without changing the meaning or correctness of the information.[/color]

That might be a theoretically possible idea, but it's not the intention,
and not a common opinion, and not reflected in any browser's behavior.

If it had been the intention, at least HTML 2.0 specification would have
said it. Instead, the specifications describe <ul> and <ol> as _lists_
(and _sequences_), which implies that the order of subelements is
significant and not mutable. The poor wordings "unordered list" and
"ordered list" cannot change this; logically, an "unordered list" is an
oxymoron, since "list" (as opposite to "collection " or "set") implies an
order; and the simple solution is that the idea always was that <ol> is a
list where the ordering is explicitly expressed using some kind of
numbering.
[color=blue]
> Your statement seems to me to say that <p> and <blockquote> should
> also be considered to have no difference other than presentational,[/color]

In actual practice on Web pages, what _is_ the difference if not
presentational? But in this case, the specifications are fairly clear.
The problem is that browsers started using a fixed (and unimaginative)
method of rendering <blockquote> elements, so it was too easy to start
abusing the markup. There's no such phenomenon for <ul> and <ol>; they
were presentational variants from the beginning.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[Alan J. Flavell]

Re: Problems Starting an Ordered List at Zero

On Mon, 27 Sep 2004, Jukka K. Korpela wrote:
[color=blue]
> But both <ul> and <ol> are inherently ordered[/color]

In so far as the items of a <ul> have to be presented in the source
code in one particular ordering, I agree. However, the HTML
specification -says- that the items are "unordered" , whatever that may
exactly mean.
[color=blue]
> - the order of the <li> elements is significant in both of them.[/color]

I'd have to disagree with that. By using <ul> you are supposed to
be telling HTML that the ordering in the source code is not
substantive to the meaning. Taking their example:

* 100 g. flour
* 10 g. sugar
* 1 cup water
* 2 eggs
* salt, pepper

These ingredients could still make the same recipe if the salt were
listed first, the water last etc.

Whereas it would likely make an entirely different result if these
procedural steps were re-sequenced:

1. Mix dry ingredients thoroughly.
2. Pour in wet ingredients.
3. Mix for 10 minutes.
4. Bake for one hour at 300 degrees.
[color=blue]
> As usual, the HTML specifications are not explicit in such
> semantics,[/color]

True enough, and it even goes on to say:

Ordered and unordered lists are rendered in an identical manner
except that visual user agents number ordered list items.

although - as so often - it's not at all clear whether this is
intended as merely an observation about the behaviour of existing
browsers, or a specification of how browsers are required to behave
for all time.

(For example, there might -hypothetically- be a future stylesheet
option for sorting unordered list items into alphabetical order, or by
string length, or whatever...)
[color=blue]
> It would not be appropriate for a browser to render (at default
> settings) <li> elements in a randomized order, or in any order
> different from the order in the HTML source.[/color]

The exact presentation of the three list types depends on the user
agent. [...] This is a stylistic issue and is properly handled by
style sheets. - (HTML4.01 section 10.1).
[color=blue]
> Surely. But the only difference between <ul> and <ol> is that by
> using <ol> you are asking the browser to add numbers to the items[/color]

Semantically, with <ol> you are declaring to HTML that the ordering of
the items is significant. The numbering is an external representation
of that semantic structure.
[color=blue][color=green]
> > With an unordered list, the semantics are that the members of the
> > list may be re-arranged in any order presentationall y, without
> > changing the meaning or correctness of the information.[/color]
>
> That might be a theoretically possible idea, but it's not the
> intention, and not a common opinion, and not reflected in any
> browser's behavior.[/color]

I wouldn't expect any browser to resequence list items by default,
indeed.

But consider a compact list presentation.

item item item item
item item item item

There are two entirely plausible ways to sequence these items
(by columns or by rows, I mean).
[color=blue]
> If it had been the intention, at least HTML 2.0 specification would
> have said it. Instead, the specifications describe <ul> and <ol> as
> _lists_ (and _sequences_), which implies that the order of
> subelements is significant and not mutable. The poor wordings
> "unordered list" and "ordered list" cannot change this; logically,
> an "unordered list" is an oxymoron, since "list" (as opposite to
> "collection " or "set") implies an order;[/color]

There's something in what you say; and often the W3C specs read like
the work of a committee who couldn't quite agree, and who then levered
several competing opinions into place without resolving the
contradictions.

all the best

[Chris Morris]

Re: Problems Starting an Ordered List at Zero

Stan Brown <the_stan_brown @fastmail.fm> writes:[color=blue]
> "Jukka K. Korpela" <jkorpela@cs.tu t.fi> wrote in
> comp.infosystem s.www.authoring.html:[color=green]
> >Indeed, because there's no CSS equivalent to the start attribute.[/color]
>
> I wonder why, given that the attribute is deprecated in HTML 4.01
> Strict.[/color]

Which was a mistake, I think - it seems to me to be entirely part of
the content. Consider a list of steps:

11. Prepare foo ...
12. Prepare bar ...
....
17. Take the preparation from step 11 and ...

(Slightly contrived example, I know rewording step 17 would help in
*this* case, and for that matter this list would start at 1 anyway)

The number '11' on the list item is clearly part of the essential
content, so a stylesheet solution is unacceptable. So why deprecate
'start'?

--
Chris

[Stan Brown]

Re: Problems Starting an Ordered List at Zero

"Rijk van Geijtenbeek" <rijk@operaremo vethiz.com> wrote in
comp.infosystem s.www.authoring.html:

(reacting to my example of a recipe as needing sequential steps, as
difference between <ol> and <ul> not being presentational)
[color=blue]
>Yes, but you might note recipes often *don't* use numbered
>(lists|paragra phs), and rely on the presentation order to convey the
>sequential relation between the steps.[/color]

Good point. I think my position falls to the ground. It's the
arrangement itself, not the fact of the numbers -- which, as you
say, aren't usually use in recipes.

--
Stan Brown, Oak Road Systems, Tompkins County, New York, USA

http://OakRoadSystems.com/

HTML 4.01 spec: http://www.w3.org/TR/html401/
validator: http://validator.w3.org/
CSS 2.1 spec: http://www.w3.org/TR/CSS21/
validator: http://jigsaw.w3.org/css-validator/

[Brian]

Re: Problems Starting an Ordered List at Zero

Stan Brown wrote:[color=blue]
> "Rijk van Geijtenbeek" wrote:
>
> (reacting to my example of a recipe as needing sequential steps, as
> difference between <ol> and <ul> not being presentational)
>
>[color=green]
>> Yes, but you might note recipes often *don't* use numbered
>> (lists|paragrap hs), and rely on the presentation order to convey
>> the sequential relation between the steps.[/color]
>
>
> Good point. I think my position falls to the ground.[/color]

Stick to your guns.
[color=blue]
> It's the arrangement itself, not the fact of the numbers -- which, as
> you say, aren't usually use in recipes.[/color]

As is often the case, Alan Flavell has hit the nail on the head:
Whether ordered steps in a recipe are prepended by numbers is
presentational, but the fact that the steps are ordered is semantic. By
contrast, ingredients in a <ul> can be listed in any order without a
change in the semantics.

--
Brian (remove "invalid" to email me)

http://www.tsmchughs.com/

[Darin McGrew]

Re: Problems Starting an Ordered List at Zero

Brian <usenet3@juliet remblay.com.inv alid> wrote:[color=blue]
> Whether ordered steps in a recipe are prepended by numbers is
> presentational, but the fact that the steps are ordered is semantic. By
> contrast, ingredients in a <ul> can be listed in any order without a
> change in the semantics.[/color]

Many recipes include language like, "combine the first 5 ingredients" or
"fold in the next 3 ingredients". Yes, the instructions could be edited to
remove the dependency upon the order of the ingredients, but in the
original recipe, the order of the ingredients has meaning.
--
Darin McGrew, mcgrew@stanford alumni.org, http://www.rahul.net/mcgrew/
Web Design Group, darin@htmlhelp. com, http://www.HTMLHelp.com/

"Do not look where you fell, but where you slipped." - African Proverb

[Jukka K. Korpela]

Re: Problems Starting an Ordered List at Zero

Neal <neal413@yahoo. com> wrote:
[color=blue]
> If the ordering of the list items is critical, should we rely on HTML
> rendering to add the right numbers? W3C says that "ordered and
> unordered lists are rendered in an identical manner except that
> visual user agents number ordered list items."[/color]

I hadn't realized they say that. It's very odd, because it sounds like a
requirement. In fact, some people have said that <ol> should generally be
preferred over <ul>, since numbering the items makes the presentation
more understandable on a _speech_ browser. And this sounds natural,
doesn't it? I even used to favor the idea but then realized I went
overboard; it's ultimately user agents' responsibility to render simple
lists properly - and to allow the user select a rendering mode where all
lists are read as numbered, I think.

Before that, they say, near the start of

http://www.w3.org/TR/html4/struct/lists.html

as follows: "An ordered list, created using the OL element, should
contain information where order should be emphasized, as in a recipe". So
the point is in _emphasizing_ the order. Or, as I would say, making it
more _explicit_. The numbers give no extra information; the just
emphasize an order and allow references by number.

By the way, I noticed that the HTML 4.01 specification contains, in the
Lists section discussed, an example of using the start="..." attribute,
indicated as deprecated. It is followed by a note "Details about
numbers", which mentions that the list item counter can be reset using
the value="..." attribute in a <li> element, without emphasizing, and
without even mentioning, that this attribute has been deprecated, too.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

[Alan J. Flavell]

Re: Problems Starting an Ordered List at Zero

On Mon, 27 Sep 2004, Darin McGrew wrote:
[color=blue]
> Many recipes include language like, "combine the first 5 ingredients" or
> "fold in the next 3 ingredients". Yes, the instructions could be edited to
> remove the dependency upon the order of the ingredients, but in the
> original recipe, the order of the ingredients has meaning.[/color]

And this is exactly where <ol start="blah"> falls down; because the
text is liable to refer to the steps by their intended *presentation*
names, and will work fine on client agents which honour the start
attribute; but support for the start attribute is not mandated, with
the results that you can certainly predict.

This is what I call "leakage": the assumption that one can refer to
step 3, or the "green bullet", or any other presentation detail that's
not mandated by the spec, from the running text. A careful author
will avoid falling into that kind of trap, but it's sometimes hard to
avoid.

Now, in an ancient publishing markup language that I was using a
considerable number of years ago, it was possible to assign labels to
the steps, and to refer to those &flour; and &eggs; and &water; items
in the running text, whereupon the renderer would plug-in whatever
list marker had been used. But HTML is too simple for that.

cheers

[Sam Hughes]

Re: Problems Starting an Ordered List at Zero

Lauri Raittila <lauri@raittila .cjb.net> wrote in
news:MPG.1bc1fc 3b565ab57198a02 8@news.individu al.net:
[color=blue]
> In article <Xns9570DCAEE60 BBhughesrpiedu@ 130.133.1.4>, hughes@rpi.edu
> says...
>[color=green]
>> Of course, some "nums" could have letters in them, such as if one
>> were referring to "Rule <num>17-C</num>". Or if one used hexidecimal
>> numbers in the ordered list, for some reason. Perhaps the LIST
>> element should have an attribute which describes the meaning of the
>> NUM elements' content. Such an attribute would be inherited from a
>> parent LIST element if not otherwise specified, if such a parent
>> exists.[/color]
>
> I think that num element whitout attribute to tell it's type would be
> useless, as people shoud understand what kind of number it is anyway
> (if they don't, that is problem outside markup...), but computers
> could get something from num element.[/color]

This would work fine with the attribute in the LIST element as mentioned
above. If different numbering schemes are used in a list, then NUM is
meaningless anyway.