Stylesheets - Text flow (wasRe: How to control flow of dynamic content with CSS?)

Author

Text flow (wasRe: How to control flow of dynamic content with CSS?)

Mark Johnson

2004-05-14, 12:16 am

Johannes Koch <koch@w3development.de> wrote:

>Robert Bowen wrote:

[color=darkred]
[color=darkred]
>No, it is not.

I wouldn't think so, either. Sounds like server side programming,
rather. With xslt I supposed you could use count and keep a recursive
counter to half of that, and then break to new <td> or whatever,
assuming you don't want to group subcategories or blocks.

>There is no automatic column flow control in CSS now. Why
>should it be?

Float is kind of like that. And NN used to have <multicol>. It seems
the logical thing, either for proprietary tags or some future CSS.
Styles are supposed to be about the presentation side and appearance,
after all, positioning, box models and all that? Having text processor
column support, as it were, seems very consistent with that. And I
mean if NN could support multicol, with gutter widths, as far back as
NN 3 in the mid 90s, it's not like it can't be done. CSS3 specifies
that options be applied to boxes in the case of page breaks, for ex.

Harlan Messinger

2004-05-14, 12:16 am

"Mark Johnson" <102334.12@compuserve.com> wrote in message
news:05p6a057omfqttd16jt42u9hfarlshng7j@4ax.com...
> Johannes Koch <koch@w3development.de> wrote:
>
>
>
>
>
> I wouldn't think so, either. Sounds like server side programming,
> rather. With xslt I supposed you could use count and keep a recursive
> counter to half of that, and then break to new <td> or whatever,
> assuming you don't want to group subcategories or blocks.

The server doesn't know where the lines are going to break in the client, so
it has nothing to count.

>
>
>
> Float is kind of like that.

Not at all. There is no flow from one block to another.

> And NN used to have <multicol>. It seems
> the logical thing, either for proprietary tags or some future CSS.
> Styles are supposed to be about the presentation side and appearance,
> after all, positioning, box models and all that? Having text processor
> column support, as it were, seems very consistent with that. And I
> mean if NN could support multicol, with gutter widths, as far back as
> NN 3 in the mid 90s, it's not like it can't be done. CSS3 specifies
> that options be applied to boxes in the case of page breaks, for ex.

Eric Jarvis

2004-05-14, 12:16 am

Mark Johnson 102334.12@compuserve.com wrote:
> Johannes Koch <koch@w3development.de> wrote:
>
>
>
>
>
> I wouldn't think so, either. Sounds like server side programming,
> rather. With xslt I supposed you could use count and keep a recursive
> counter to half of that, and then break to new <td> or whatever,
> assuming you don't want to group subcategories or blocks.
>
>
>
> Float is kind of like that. And NN used to have <multicol>. It seems
> the logical thing, either for proprietary tags or some future CSS.
> Styles are supposed to be about the presentation side and appearance,
> after all, positioning, box models and all that? Having text processor
> column support, as it were, seems very consistent with that. And I
> mean if NN could support multicol, with gutter widths, as far back as
> NN 3 in the mid 90s, it's not like it can't be done. CSS3 specifies
> that options be applied to boxes in the case of page breaks, for ex.
>

the reason it's not a smart idea is that newspaper style columns of text
don't work for screen display in the way they do on paper...the last thing
you want for screen display it to have to scroll all the way down a column
and then have to scroll back to the top and back down the next column...it
is far more effort than simply scrolling down

it could be done...it would largely be totally pointless

it's the sort of thing people often think is a cute idea after looking at
some paper layouts...it's also the sort of thing that site users complain
about when trying to use it on the web...the idea isn't new...people HAVE
come up with ways of emulating it...they tend not to bother with it for
very long

the only way such columns work is with small amounts of text...at anything
over 15 lines or so it's best avoided...at anything less it's largely
pointless

--
eric
www.ericjarvis.co.uk
"live fast, die only if strictly necessary"

Lauri Raittila

2004-05-14, 12:16 am

Eric Jarvis wrote:
> the only way such columns work is with small amounts of text...at anything
> over 15 lines or so it's best avoided...at anything less it's largely
> pointless

Multicolumn could work. If the browser would detect window height, and
add more columns, when content would not fit. So that users would only
need to scroll right. It wouldn't cause problems, if pageup/down, and
mouse wheel, and all other means to scroll down would then work to scroll
right.

It might be sencible aproach on some text intensive thing, like ebooks,
but I doubt it would be good for webpages. But the gain against normal
scrolling down would be quite minimal.

--
Lauri Raittila <http://www.iki.fi/lr> <http://www.iki.fi/zwak/fonts>
I'm looking for work | Etsin työtä

Mark Johnson

2004-05-14, 9:28 am

"Harlan Messinger" <h.messinger@comcast.net> wrote:

>"Mark Johnson" <102334.12@compuserve.com> wrote in message
>news:05p6a057omfqttd16jt42u9hfarlshng7j@4ax.com...
[color=darkred]
[color=darkred]
[color=darkred]
[color=darkred]
[color=darkred]
[color=darkred]
>The server doesn't know where the lines are going to break in the client

Yes it does. He's describing a list of items. You simply split the
half of them into table cells, divs, what have you, as the page is
created. This isn't about flowing text. You're just putting the half
one place, and the other half somewhere else.

[color=darkred]
[color=darkred]
>Not at all. There is no flow from one block to another.

I believe there is:
[color=darkred]

For example, at a page break, a block could be left open at the border
or closed.

I'm not saying it's flowing text. But multicol was sort of that. And
it doesn't seem to violate any rule regarding 'anonymous content'.

Mark Johnson

2004-05-14, 9:28 am

Eric Jarvis <web@ericjarvis.co.uk> wrote:

>the reason it's not a smart idea is that newspaper style columns of text
>don't work for screen display in the way they do on paper...the last thing
>you want for screen display it to have to scroll all the way down a column

That's the point that's used. You have only a window on the page -
like someone nearsighted using a magnifying sheet.

>it could be done...it would largely be totally pointless

It would be just perfect for indices, for example, if you don't want
to insert hard column breaks, in whatever way. But you would like the
option of keeping a sub-block together.

>it's the sort of thing people often think is a cute idea after looking at
>some paper layouts...it's also the sort of thing that site users complain
>about when trying to use it on the web...the idea isn't new...people HAVE
>come up with ways of emulating it...they tend not to bother with it for
>very long

It would be too much of a workaround. Workarounds are the thing you
jettison as soon as the feature is properly implemented. There's no
substitute. This has to be system level programming.

Harlan Messinger

2004-05-14, 5:34 pm

"Mark Johnson" <102334.12@compuserve.com> wrote in message
news:o3c9a0hj8khqllkgv6s3ruuppmicfik672@4ax.com...
> "Harlan Messinger" <h.messinger@comcast.net> wrote:
>
>
>
and[color=darkred]
>
>
>
>
>
> Yes it does. He's describing a list of items. You simply split the
> half of them into table cells, divs, what have you, as the page is
> created.

Even though he could have meant that, he didn't specify that at all.

> This isn't about flowing text.

I guess I was confused by the part where he wrote "how can I make my text
'flow' in two columns?".

> You're just putting the half
> one place, and the other half somewhere else.

>
>
>
>
>
> I believe there is:

I've never seen text leak out of the float block that contains it and into
another arbitrary block on the page.

>
>
> For example, at a page break, a block could be left open at the border
> or closed.

HTML doesn't know about page breaks. In the HTML code, an end tag is either
in a particular place or not.

>
> I'm not saying it's flowing text.

You said above that you believe text does flow from one block to another,
but now you're *not* saying it's flowing text.

If you're not saying it's flowing text, then why are you presenting it as an
evidence that you can have flowing text?

> But multicol was sort of that.

That's nice, but what bearing does that have on the current question with
respect to current browsers and standard HTML and CSS?

> And
> it doesn't seem to violate any rule regarding 'anonymous content'.

I don't know what that means or how it relates to anything I said.

Mark Johnson

2004-05-15, 5:30 pm

"Harlan Messinger" <h.messinger@comcast.net> wrote:

>"Mark Johnson" <102334.12@compuserve.com> wrote in message
>news:o3c9a0hj8khqllkgv6s3ruuppmicfik672@4ax.com...
[color=darkred]
[color=darkred]
>I guess I was confused by the part where he wrote "how can I make my text
>'flow' in two columns?".

And:

[color=darkred]
[color=darkred]
[color=darkred]
[color=darkred]
>I've never seen text leak out of the float block

But the blocks themselves do. Float left, for example. If there's
room, the block is off to the right. If not, it's below.

And:

[color=darkred]
[color=darkred]
>HTML doesn't know about page breaks.

CSS does, or what I read of CSS 3, at any rate.

[color=darkred]
>I don't know what that means

I believe it's a term you'll find in the W3C spec for 2 or 3.

ginowhitaker

2004-06-16, 4:23 pm

Take a look at this example of flowing column text:
http://www.iht.com

ginowhitaker

2004-06-16, 4:27 pm

quote:
Originally posted by ginowhitaker
Take a look at this example of flowing column text:
http://www.iht.com

Click on any article.