Re: <br> as paragraph separator

Without a doubt, this has to be brought to the attention of Ian.

I intend to do so after bringing the issue to discussion at the W3C i18n WG
meeting this week. The HTML5 bug was filed under the WG's auspices, so it
would be best that the WG express an opinion on the matter.

Although I will abide by the majority decision, I still believe that <br> as
a paragraph separator is best. Although the HTML spec demands that <br> be
"used only for line breaks that are actually part of the content, as in
poems or addresses" and "not be used for separating thematic groups in a
paragraph", its "abuse" is rampant. That's because in certain scenarios it
is much easier to use than the recommended alternatives, and because it
works fine - until the applications has to support bidi text. Then,
suddenly, <br>'s bidi whitespace behavior starts to cause problems for bidi
users. Getting the app to stop using <br> at that point is an uphill battle,
since it is perceived as a great inconvenience for the benefit of a
"marginal use case" (bidi). I do not want the battle for HTML purism to be
waged entirely at the expense of bidi users.

Aharon





On Thu, Dec 16, 2010 at 4:51 AM, "Martin J. Dürst"
<duerst@it.aoyama.ac.jp>wrote:

> On 2010/12/16 8:23, Asmus Freytag wrote:
>
>  When the story was, "oh, nobody implements it correctly anyway" that
>> provided a lot more leeway to bring the standard inline with practice.
>> That leeway is now gone.
>>
>
> That wasn't the story. The story was something like "Mozilla does it
> correctly, but IE does it wrong, and won't change, so we have to change to
> IE's ways". So there is even less leeway.
>
> Anyway, we should bring all these things to the attention of the HTML
> folks, in particular Ian. They have their own ideas on <br> and they have
> their own guidelines on how to treat things with certain implementation
> conformance. Also, they will know how 'strict' is supposed to relate to
> HTML5.
>
> As for TR 20, I think we still need to change something. Because current
> browser implementation varies, we can at least for the current time not stay
> with the simple "LS is just a <br>".
>
> Regards,   Martin.
>
> --
> #-# Martin J. Dürst, Professor, Aoyama Gakuin University
> #-# http://www.sw.it.aoyama.ac.jp   mailto:duerst@it.aoyama.ac.jp
>