Ordered lists, and how to break one up while still validating.

M

Mike Barnard

Yoohoo, it's me again.

How do you do large ordered lists, if you need to break them up with
headings yet still have it pass validation?

www.thermachek.com/temp/thelawdetail.php

This is an ordered list of 53 items, a breakdown of the 53 articles in
a UK law. The list is continious but broken by part headings. On
validation I get an error on each "Part1" "Part2" etc item;

Line 285, Column 3: document type does not allow element "H4" here;
assuming missing "LI" start-tag.
<h4>

I know what is wrong, I need to finish a list, do the heading then
start a new list. So I need the new list to start an a number other
than 1, where the previous list finished. The "start" property of the
<ol> has been deprecieated but no matching CSS has been created.

I have found a bodge for this at...
http://tb-one.se/2007/12/29/ol-start-number/
.... but it's a bit long winded. What would you do?

Leave it as it is? No validation but it works. (Must test in other
browsers than FF! Have not done so yet. Must get Safari to load too.)

Use the deprecieated 'start' property? No validation but *may* work in
more browsers?

Use the hack above? I will try it, but do you know of any problems
with it?

I look forward to having my eyes opened even wider! Happy Easter if
Easter is your thing.
 
B

BootNic

Yoohoo, it's me again.

How do you do large ordered lists, if you need to break them up with
headings yet still have it pass validation?

www.thermachek.com/temp/thelawdetail.php

This is an ordered list of 53 items, a breakdown of the 53 articles in
a UK law. The list is continious but broken by part headings. On
validation I get an error on each "Part1" "Part2" etc item;

Line 285, Column 3: document type does not allow element "H4" here;
assuming missing "LI" start-tag.
<h4>

I know what is wrong, I need to finish a list, do the heading then
start a new list. So I need the new list to start an a number other
than 1, where the previous list finished. The "start" property of the
<ol> has been deprecieated but no matching CSS has been created.

Are you sure you need to restart the list?

Let's see, what you have is something like this:

<li>text</li>
<h4>heading</h4>
<li>text</li>

I think you understand the error. However there is another way to go about
doing what you want other then closing the list, add header then restart
the list.

What you can do is:

<li>text
<h4>heading</h4>
</li>
<li>text</li>

[snip]
 
B

Bergamot

BootNic said:
What you can do is:

<li>text
<h4>heading</h4>
</li>
<li>text</li>

That doesn't look very logical.

If the list number really does need to resume on each list, it may be
better to use the deprecated start attribute on the <ol> than resort to
incorrect markup just to please the validator.

Validated code is not an end to itself, though some people seem to think
it is.

A nested list would be correct, too, I think. Like a table of contents.

<ol>
<li><hx>Section heading</hx>
<ol start="1">
<li>...

Looking at the page in question, though, I wonder what harm there is in
letting the lists restart numbering at 1.
 
M

Mike Barnard

Yoohoo, it's me again.

How do you do large ordered lists, if you need to break them up with
headings yet still have it pass validation?

www.thermachek.com/temp/thelawdetail.php

This is an ordered list of 53 items, a breakdown of the 53 articles in
a UK law. The list is continious but broken by part headings. On
validation I get an error on each "Part1" "Part2" etc item;

Line 285, Column 3: document type does not allow element "H4" here;
assuming missing "LI" start-tag.
<h4>

I know what is wrong, I need to finish a list, do the heading then
start a new list. So I need the new list to start an a number other
than 1, where the previous list finished. The "start" property of the
<ol> has been deprecieated but no matching CSS has been created.

Are you sure you need to restart the list?

Let's see, what you have is something like this:

<li>text</li>
<h4>heading</h4>
<li>text</li>

I think you understand the error. However there is another way to go about
doing what you want other then closing the list, add header then restart
the list.

What you can do is:

<li>text
<h4>heading</h4>
</li>
<li>text</li>

[snip]

Thanks Boot, I like the simple answers!
 
J

Jukka K. Korpela

Scripsit Mike Barnard:
I know what is wrong, I need to finish a list, do the heading then
start a new list. So I need the new list to start an a number other
than 1, where the previous list finished. The "start" property of the
<ol> has been deprecieated but no matching CSS has been created.

Right, so use the "start" attribute.
Use the deprecieated 'start' property? No validation but *may* work in
more browsers?

What do you mean by "No validation"? Validity is a formal thing, and a
document validates if you use document type definition (DTD) that
matches the actual markup. In this case, you can simply use HTML 4.01
Transitional. Even if this were not the case, "No validation" would not
be true; if you actually used an attribute not present in HTML
specifications, then you could say "Nonconforming markup" or even
(stretching the word "standard") "Nonstandard markup". But it would
still be valid if and only if you use a DTD that allows the markup.

Technically, the "start" attribute is presentational, so if it is
_essential_ that the items be numbered in a particular way, you should
put the numbers into actual content, e.g.
<li>1. Text of the item</li>
probably inside a <ul> element (since you dont want the
browser-generated numbers that <ol> produces at least when CSS is off)
with list-style-type: none.

But in practice, I think it is fairly safe the treat "start" as
semantic, i.e. to rely on the number being generated as specified by
this attribute.
 
M

Mike Barnard

That doesn't look very logical.

But, having tried it it works. It validates and works in all
browsers, so far.
If the list number really does need to resume on each list, it may be
better to use the deprecated start attribute on the <ol> than resort to
incorrect markup just to please the validator.

Are there any browsers that ignore this deprecieated term, or wil it
kick in to quirks mode? Dunno.
Validated code is not an end to itself, though some people seem to think
it is.

A nested list would be correct, too, I think. Like a table of contents.

<ol>
<li><hx>Section heading</hx>
<ol start="1">
<li>...

Looking at the page in question, though, I wonder what harm there is in
letting the lists restart numbering at 1.

It's a legal document that continues 1 to 53. The numbers match with
the articles. Look at the link to the original and you'll see how they
match.

Thanks.
 
B

Bergamot

Mike said:
Are there any browsers that ignore this deprecieated term,

If it's valid in quirks mode, it will be supported even in standards
mode. I don't know of any browsers that ignore deprecated stuff
altogether. FWIW, I've always wondered why they deprecated the start
attribute, since it's not really presentational. It's part of the
content, at least to me.
or wil it
kick in to quirks mode?

Quirks mode is only triggered by the doctype, not by anything else in
the markup.
It's a legal document that continues 1 to 53.

OK, then I suggest using what makes sense for your particular document,
not just what the validator wants.
 
B

Blinky the Shark

Are there any browsers that ignore this deprecieated term, or wil it
kick in to quirks mode? Dunno.

Deprecated, Mike. Common scanning error. Just a heads-up.

Appropriately for this holiday some celebrate, there are religious roots
for the word: "[L. deprecatus, p. p. of deprecari to avert by prayer To
pray against, as an evil; to seek to avert by prayer; to seek deliverance
from; to express deep regret for; to desire the removal of."

If prayer fails, you'll have to remove the offending markup by hand. :)
 
J

Jukka K. Korpela

Scripsit Bergamot:
If it's valid in quirks mode,

That's an absurd sentence, like "if 1 + 1 equals 2 when printed on
paper". Validity is a formal property of a document. Nothing that a
browser might do in displaying the document cannot change that.
Quirks mode is only triggered by the doctype, not by anything else in
the markup.

Wrong. The presence or absence of a comment, for example, before the
doctype declaration is crucial on IE.
OK, then I suggest using what makes sense for your particular
document, not just what the validator wants.

The last part "not just what the validator wants" is very strange. Are
you suggesting the use of invalid markup (which?), or are you just
saying that validity is not sufficient (which should be pretty obvious
anyway)?
 
D

dorayme

Mike Barnard said:
How do you do large ordered lists, if you need to break them up with
headings yet still have it pass validation?

First you make sure it is an ordered list. Just because a list has
numbers does not mean it is ordered in the meaning of "ordered list".
The meaning of order that is particularly relevant to an ordered list
can be a tricky matter. The folk who made up the css on this cannot be
blamed for not looking deeply into the matter, ordering and classifying
being a very complex linguistic human activity. They rested content on
probably no more than a couple of simple paradigms or examples. Thus:

If a shopping list, for example, consisted of dry groceries and were to
be purchased from a shop with a staff member behind a counter (as in the
good old days), then the order is unlikely to matter. But it might
matter if the items were to be purchased from different shops on a route
that minimised energy. If it matters which order the person is to get
things, then you would give an ordered list. If not, an unordered one.

However there might be within the one list a couple of categories of
things to be fetched, one in an order, the other in no particular order.
The logical way to list the items here would depend very much on the
exact circumstances. The main list might be an ordered one with suome
sub lists within the order which are unordered.

For example, your son should first take the dress item to be altered,
and while it is being altered should second go fetch the dry groceries
and third collect the dress item and take to dry cleaners for pressing.
This would be a 3 item ordered list with the second item having a sub
unordered list.

You can easily imagine and and find cases far more complex. Sometimes it
is quite a challenge to mark up. But the first thing is to settle what
it is your material is really meaning.

There are traps. The list might have numbers on them without being an
ordered list, where it would be actually wrong to use an OL. An example,
you send your son to the parts shop for some spare parts for something
and the items are in fact labelled with product numbers. You happen to
want consecutive numbers. It does not matter that it would be practical
for mnemonic purposes to list them in order. It is still an unordered
list. It does not matter what order the counter staff go off and fetch,
how they pack them in your bag, etc. The point is that all the items are
to be fetched in any order and they are all to be as specified. The
number may be a way to identify the items. This is not an ordering
matter but a matching matter. If you are wanting to use an html list,
you would use a ul but put in the numbers as part of the text in the
list item.

Or you would use a table! I have argued on more than one occasion at
some length that an ordered list is not more than a two column table in
the sense that each is as semantically as appropriate as the other. An
ordered list is not more than a 2 col table where the order is in one
col and the item in the other, the headings being either specified to
make clear what the number numbering means or else to be understood.

Normally I would say an unordered list has no exact semantic table
equivalent. But I make an exception for a special type of unordered
list. One in which the numbers are an identifying feature of the item
(rather than a call for action in the real world in a particular order -
as in an algorithm). The spare parts example is appropriate here. The
table would certainly be appropriate where the part number is on the
left and the item description on the right, row by row a very tabular
affair.

Be careful not to take on trust what is said to you on this Mike, think
through it all. Ask yourself what do the numbers mean?

You might conclude that the items you are dealing with *is* an ordered
list. (I even have some considerations to lend weight to that, I have
others, and more of them against it. But I won't trouble you with these
specific things). I have not discussed this. But the mere presence of
numbers is no guarantee. This does not mean you cannot lay things out as
you want. You can still have numbers and letters and whatever you want
as I have mentioned above, either in an unordered list or in a table).

I mention all this stuff so you can pick the most intelligent tool for
the job.
 
A

Andy Dingley

It's a legal document that continues 1 to 53. The numbers match with
the articles. Look at the link to the original and you'll see how they
match.

Then hard-code the numbers. Especially if they're references to paras
in some other cited document. Especially so in this case because you
know the numbers _are_ stable, i.e. skipping or duplicating a list
element in your document should preserve the numbers uncanged, not re-
number in sequence.

Numbering of <ol> in HTML just doesn't work particularly flexibly.
Attempts to make it work are amusing exercises for coders, but it's no
way to get the job done.
 
B

Ben C

First you make sure it is an ordered list. Just because a list has
numbers does not mean it is ordered in the meaning of "ordered list".
The meaning of order that is particularly relevant to an ordered list
can be a tricky matter. The folk who made up the css on this cannot be
blamed for not looking deeply into the matter, ordering and classifying
being a very complex linguistic human activity. They rested content on
probably no more than a couple of simple paradigms or examples. Thus:

If a shopping list, for example, consisted of dry groceries and were to
be purchased from a shop with a staff member behind a counter (as in the
good old days), then the order is unlikely to matter. But it might
matter if the items were to be purchased from different shops on a route
that minimised energy. If it matters which order the person is to get
things, then you would give an ordered list. If not, an unordered one.

Putting numbers on a list doesn't usually mean anything much about order
because people number them in the order they're in anyway.

No-one writes:

3. Go home
1. Go to shop
2. Buy apples

So you might just as well use bullets or nothing as write:

1. Go to shop
2. Buy apples
3. Go home

The main point of numbers is so you can refer to them. "In section 2,
rotten apples are deprecated", etc.

[...]
Normally I would say an unordered list has no exact semantic table
equivalent. But I make an exception for a special type of unordered
list. One in which the numbers are an identifying feature of the item
(rather than a call for action in the real world in a particular order -
as in an algorithm). The spare parts example is appropriate here. The
table would certainly be appropriate where the part number is on the
left and the item description on the right, row by row a very tabular
affair.

I think they're all like that really-- the numbers are just labels to
refer to the items on the list.
 
B

Ben C

Scripsit Mike Barnard:


Right, so use the "start" attribute.


What do you mean by "No validation"? Validity is a formal thing, and a
document validates if you use document type definition (DTD) that
matches the actual markup. In this case, you can simply use HTML 4.01
Transitional. Even if this were not the case, "No validation" would not
be true; if you actually used an attribute not present in HTML
specifications, then you could say "Nonconforming markup" or even
(stretching the word "standard") "Nonstandard markup". But it would
still be valid if and only if you use a DTD that allows the markup.

Technically, the "start" attribute is presentational, so if it is
_essential_ that the items be numbered in a particular way, you should
put the numbers into actual content, e.g.
<li>1. Text of the item</li>
probably inside a <ul> element (since you dont want the
browser-generated numbers that <ol> produces at least when CSS is off)
with list-style-type: none.

You should be able to use CSS to reset the counters (using counter-reset
and counter-increment), but it isn't supported very widely. Opera
manages it.

Easier just to put the numbers in the <li>s as you suggest.
 
D

dorayme

Ben C said:
Putting numbers on a list doesn't usually mean anything much about order
because people number them in the order they're in anyway.

Putting the numbers on either unordered or ordered lists is one thing.
The list being ordered or unordered in its nature is another.

No-one writes:

3. Go home
1. Go to shop
2. Buy apples

So you might just as well use bullets or nothing as write:

1. Go to shop
2. Buy apples
3. Go home

Not necessarily. If the numbers are labels that have a meaning that is
not a mere (and mostly not particularly useful) confirmation of the
order on the page, then they might well be so written as in your first
example. You can see this better if you *always* think in terms of a 2
col table and ask yourself what would be in the head of the table over
the number col. If "Shop number in the mall" was over it and you asked
your son to fetch various things, you might well give him an ordered
list - yes, an ordered one - but with "3" showing in the first row
because shop 3 was in fact the shop you wanted him to go to first. The
table rows, as they proceeded down would reflect the order you wanted or
recommended your son to go in, the numbers being merely to identify
something to do with the list item. Think my case of the dress
alteration/dry cleaner example.


The main point of numbers is so you can refer to them. "In section 2,
rotten apples are deprecated", etc.

I can't really agree with this as a general statement.

[...]
Normally I would say an unordered list has no exact semantic table
equivalent. But I make an exception for a special type of unordered
list. One in which the numbers are an identifying feature of the item
(rather than a call for action in the real world in a particular order -
as in an algorithm). The spare parts example is appropriate here. The
table would certainly be appropriate where the part number is on the
left and the item description on the right, row by row a very tabular
affair.

I think they're all like that really-- the numbers are just labels to
refer to the items on the list.

If you mean by "they" in "they're all like that really" to refer to all
lists that have numbers on them, then I disagree. In an ordered list the
numbers are not "just" labels. They are a reflection of an important
real world situation (actual, required or imagined). In an algorithm,
for example, the order in which the list items are arranged is
important. The actual numbers are not as important, they do have a
labelling function.

Consider how an ordered list might have no numbers

<ol style="list-style: none;">


(A rare sight because so many authors simply do not take the real
distinction between an ordered and an unordered list seriously, most
authors, I suspect (perhaps this is unfair? I do not refer to regulars
here.) just think of an ordered list as a presentational device to get
numbers up on the page.

Just as an unordered one might have no numbers or even bullets, so too
might an ordered one. They are different in meaning and this meaning
might be very important. In a critical situation, it could mean the
difference between life and death. Want examples?
 
B

Ben C

Putting the numbers on either unordered or ordered lists is one thing.
The list being ordered or unordered in its nature is another.
OK.


Not necessarily. If the numbers are labels that have a meaning that is
not a mere (and mostly not particularly useful) confirmation of the
order on the page, then they might well be so written as in your first
example.

Certainly if the numbers aren't just in ascending order then they
probably do mean something else. But it's rare that they aren't just in
ascending order (and quite hard to get OLs not to just go up in
ascending order).

That's why I think that in most cases the only difference between
numbers and bullets is that you can use the numbers to refer to the
items in other text in the document. The list is usually ordered (even
if arbitrarily) in only one way.
You can see this better if you *always* think in terms of a 2
col table and ask yourself what would be in the head of the table over
the number col. If "Shop number in the mall" was over it and you asked
your son to fetch various things, you might well give him an ordered
list - yes, an ordered one - but with "3" showing in the first row
because shop 3 was in fact the shop you wanted him to go to first. The
table rows, as they proceeded down would reflect the order you wanted or
recommended your son to go in, the numbers being merely to identify
something to do with the list item. Think my case of the dress
alteration/dry cleaner example.

Yes, any time you have some data for which you have two orderings. For
example:

3. Tom
1. Dick
2. Harry

might mean Tom won at Boggle but came 3rd at Scrabble, etc.

[...]
If you mean by "they" in "they're all like that really" to refer to all
lists that have numbers on them, then I disagree.

Well if the numbers are in the same order that the items are arranged on
the page, then they aren't really saying anything.

A list is either ordered or unordered in the sense of whether the order
matters, and either kind of list can be written with numbers or bullets.

If an ordered list is written with numbers, and the numbers are
in-order, then surely they are just labels? The list is in order anyway.

1. Light blue touchpaper
2. Stand back

doesn't mean anything different from:

* Light blue touchpaper
* Stand back

They're both obviously ordered lists. The numbers are just labels.

If the numbers aren't in the same order as the items' vertical positions
then they aren't just labels, but then perhaps it's really a table and
not just a list at all.
In an ordered list the numbers are not "just" labels. They are a
reflection of an important real world situation (actual, required or
imagined). In an algorithm, for example, the order in which the list
items are arranged is important. The actual numbers are not as
important, they do have a labelling function.

Consider how an ordered list might have no numbers

<ol style="list-style: none;">

Yes I can see an ordered list might have no numbers. But this just makes
me more convinced than ever that the numbers are mere labels.

If an ordered list can be written just as well with bullets, which it
can, then using in-order numbers instead doesn't seem to be adding
anything. Perhaps it emphasizes the order a bit? I don't even think it
does, especially as one often numbers unordered lists.

The only point I can see to in-order numbers is cross-referencing.
(A rare sight because so many authors simply do not take the real
distinction between an ordered and an unordered list seriously, most
authors, I suspect (perhaps this is unfair? I do not refer to regulars
here.) just think of an ordered list as a presentational device to get
numbers up on the page.

That is my cynical view of OLs. But in that case they are misnamed--
they should be called numbered and unnumbered lists, not ordered and
unordered. Perhaps you're right, people should use OL/UL for
ordered/unordered lists and decide whether they're numbered or get
bullets with CSS.
Just as an unordered one might have no numbers or even bullets, so too
might an ordered one. They are different in meaning and this meaning
might be very important. In a critical situation, it could mean the
difference between life and death. Want examples?

I can think of examples where doing things in the wrong order results in
misadventure and death. Can you give an example where using in-order
numbers rather than bullets (or the other way round) on either an
unordered or ordered list could mean the difference between life and
death?
 
D

dorayme

Ben C said:
Ben C <[email protected]> wrote: [...]
....
Well if the numbers are in the same order that the items are arranged on
the page, then they aren't really saying anything.

A list is either ordered or unordered in the sense of whether the order
matters, and either kind of list can be written with numbers or bullets.

If an ordered list is written with numbers, and the numbers are
in-order, then surely they are just labels? The list is in order anyway.

1. Light blue touchpaper
2. Stand back

doesn't mean anything different from:

* Light blue touchpaper
* Stand back

They're both obviously ordered lists. The numbers are just labels.

If the numbers aren't in the same order as the items' vertical positions
then they aren't just labels, but then perhaps it's really a table and
not just a list at all.

What is *really* a table and what is *really* a list is not a simple
matter. An ordered list, and even some unordered lists are semantically
not that different from 2 col tables with special headings. I have
argued this a number of times. But please read on first.
Yes I can see an ordered list might have no numbers. But this just makes
me more convinced than ever that the numbers are mere labels.

If an ordered list can be written just as well with bullets, which it
can, then using in-order numbers instead doesn't seem to be adding
anything. Perhaps it emphasizes the order a bit? I don't even think it
does, especially as one often numbers unordered lists.

It all depends on quite what is meant by "mere labels". I don't think we
are as far apart on this question of the numbers as might appear. We
both agree that the presence of numbers or bullets do not alter the
fundamental difference between two quite different types of list, the ol
and the ul. You appear to lean toward the idea that they are almost
gratuitous or else useful for later cross reference.

The boy you sent to get the cheese with the unpronounceable and
impossible to spell name will be relieved to tell his parent that the
grocer said he did not have the item with the 15 in front of it. Even if
the list was ordered!

Whereas I am more wary of quite so downplaying them. The numbers are
information to the user that the list is probably an ordered one, that
the order is important. The numbers are *not* always mere labels. They
are doing a semantic job that goes beyond simply linking the item to a
label.

In an ideal world in which ordered list were seen to be such, the
numbers would not be needed except as mere labels (as they are in
unordered ones).

The only point I can see to in-order numbers is cross-referencing.


That is my cynical view of OLs. But in that case they are misnamed--
they should be called numbered and unnumbered lists, not ordered and
unordered. Perhaps you're right, people should use OL/UL for
ordered/unordered lists and decide whether they're numbered or get
bullets with CSS.


I can think of examples where doing things in the wrong order results in
misadventure and death.

And so you will agree that it is sometimes crucial to know if a list is
ordered or not. One indication (it is not 100% reliable because we both
can see how unordered lists are sometimes also labelled) is the
numbering rather than bulleting. This function is not for mere cross
referencing.
 
D

dorayme

dorayme said:
... The numbers are *not* always mere labels. They
are doing a semantic job that goes beyond simply linking the item to a
label.

Perhaps, Ben, you might be a little more convinced that the numbers in
an ordered list are not *mere labels* by the following consideration:
imagine an ordered list that got printed out, damaged and broken up.
You know from the context or from memory that the list critically
ordered. You cannot reconstruct the order if the labels did not have an
ordering meaning. If they were 1, 2, 3 etc this would immediately give
you the order, first, second, third. If they were mere labels with no
meaning at all (beyond a handy cross referencing device) you would be
lost and it could result in a critical accident.

This is why I say that the numbers in an ordered list, while not being
necessary, are nevertheless more than mere labels as might be the case
in an unordered list. One can deduce from the n-format label which is
nth item in a list in a way one cannot deduce from an arbitrary-label
which is the nth item in a list.
 
B

Ben C

Perhaps, Ben, you might be a little more convinced that the numbers in
an ordered list are not *mere labels* by the following consideration:
imagine an ordered list that got printed out, damaged and broken up.

You know from the context or from memory that the list critically
ordered. You cannot reconstruct the order if the labels did not have an
ordering meaning. If they were 1, 2, 3 etc this would immediately give
you the order, first, second, third. If they were mere labels with no
meaning at all (beyond a handy cross referencing device) you would be
lost and it could result in a critical accident.

This is why I say that the numbers in an ordered list, while not being
necessary, are nevertheless more than mere labels as might be the case
in an unordered list. One can deduce from the n-format label which is
nth item in a list in a way one cannot deduce from an arbitrary-label
which is the nth item in a list.

I looked at the HTML spec and they do say UL is for data where the order
isn't important and OL for when it is. Their example is the ingredients
in a cake (unordered) and the instructions to make it (ordered).

It then goes on to say that ordered lists are rendered with numbers,
which are supposed to emphasize the order, like you're saying. So you
aren't the only person who thinks this.

Perhaps using bullets for ordered information is bad style if nothing
else, since it could cause critical accidents.

If someone pedantic is using numbers just as labels perhaps they should
technically use an OL or even a DL and restyle it?
 
D

dorayme

Ben C said:
I looked at the HTML spec and they do say UL is for data where the order
isn't important and OL for when it is. Their example is the ingredients
in a cake (unordered) and the instructions to make it (ordered).

It then goes on to say that ordered lists are rendered with numbers,
which are supposed to emphasize the order, like you're saying. So you
aren't the only person who thinks this.

Perhaps using bullets for ordered information is bad style if nothing
else, since it could cause critical accidents.

If someone pedantic is using numbers just as labels perhaps they should
technically use an OL or even a DL and restyle it?

Is your last "OL" a typo?

Anyway, just talking to you about this has made me even more uneasy
about the whole distinction as implemented in html and css. I am rather
tempted by the view that there should never have been the distinction in
the first place. I repeat that a couple of paradigm cases (see your
reference to the examples used in the html specs) are just too thin a
ground to rest this machinery on. (I say repeat because I have made the
point before with other examples)

Imagine no "ol" or "ul" in the first place in html, just "list". Keep it
simple! Leave it to the context and the author to make it clear how
important the order is.

If you actually think about it, there are surely gradations of
importance. By this I mean specifically that a list's order might be
partly important and this will never ever be captured by meanings built
into html elements. You could have a list that was ordered to an extent!
And I am *not* meaning one that is so clear that it can be split into an
ol with ul sub lists or vice versa. There may well be contexts where the
author is simply unsure which order to put something in but knows that
the order or rough order might be important.

As in an algorithmic proposal to be experimented with by the reader.
Sure the author might ol it and add a rider that the reader transpose
the order to make the procedure do what he wants. Or he might ul it and
add a rider that the order they are seeing on the page is roughly right
(eg. you do not light the match first but you might put in the detonator
before you add the right side panel to the box depending on the tools
you have available or your dexterity). Never mind the details, I am
saying that the distinction between ol and ul is really not all that
useful in practice. It is a distracting straight jacket and impedes
development. Authors should really be freed of having to worry about it.

We would never be having this discussion, others would never be worried
which to use where and when and so on. The context is often the best
indicator of what is meant. I am saying that the html meanings here are
nowhere near the power of contexts in their capacity to transmit
information.
 
B

Ben C

Is your last "OL" a typo?

Yes, I meant UL.
Anyway, just talking to you about this has made me even more uneasy
about the whole distinction as implemented in html and css. I am
rather tempted by the view that there should never have been the
distinction in the first place. I repeat that a couple of paradigm
cases (see your reference to the examples used in the html specs) are
just too thin a ground to rest this machinery on. (I say repeat
because I have made the point before with other examples)

Imagine no "ol" or "ul" in the first place in html, just "list". Keep
it simple! Leave it to the context and the author to make it clear how
important the order is.

The reality is many documents contain lists with either numbers or
bullets. Word processors have a button for whether you want numbers or
bullets. People expect to choose numbers or bullets. So that's what you
got in HTML and they were called UL and OL, even if that's not what
they're supposed to mean.

Now that you can style either how you like perhaps just one kind of list
would be better, although it's more convenient just to leave it how it
is.
If you actually think about it, there are surely gradations of
importance. By this I mean specifically that a list's order might be
partly important and this will never ever be captured by meanings
built into html elements. You could have a list that was ordered to an
extent! And I am *not* meaning one that is so clear that it can be
split into an ol with ul sub lists or vice versa. There may well be
contexts where the author is simply unsure which order to put
something in but knows that the order or rough order might be
important.
[...]
We would never be having this discussion, others would never be
worried which to use where and when and so on. The context is often
the best indicator of what is meant. I am saying that the html
meanings here are nowhere near the power of contexts in their capacity
to transmit information.

They never are. The criterion for choosing a tag can't reasonably be
anything more strict than: is there a better tag I could have used from
the choice available?
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,744
Messages
2,569,482
Members
44,901
Latest member
Noble71S45

Latest Threads

Top