Switch www.mozilla.org to HTML5 doctype
Samuel Sidler
All discussion should happen here, however.)
Hey everyone,
I'd like to propose that we change the doctype we use on
www.mozilla.org to the HTML5 doctype. Specifically:
<!DOCTYPE html>
There's quite a bit of rationale for this:
1. In the old design (and for years prior) we used and conformed to
HTML4 Transitional. The closest modern equivalent is the upcoming
HTML5 doctype. After switching to the new design, we changed our
doctype to XHTML 1.0 Strict. Because of this, most of our site doesn't
validate. We can fix up the entire site, but seems like the practical
thing to do is to use something closer to the previous HTML4 doctype.
2. As we promote standards, it's pretty clear that HTML5 is a highly
strategic technology that we should be supporting. Doing so on www.mozilla.org
is just a start and we should push for most of our *.mozilla.org
sites to use the HTML5 doctype when possible in the future. There's no
reason www.mozilla.org shouldn't be at the cutting edge of this
transition.
3. We've talked a lot about using more modern HTML5-only tags like the
<video> tag. These are really HTML5-only tags.
I'm not sure of any negative issues with making this change except
that some pages might not display correctly. But I actually don't
believe that's true in 99% of our pages and we can fix any issues we
notice along the way.
If anyone has any comments about this, it'd be great, otherwise, I
think we should make this change in a week or two at most so those who
want to continue to make www.mozilla.org valid can do so on a doctype
we'll be using long term.
-Sam
davidwboswell
> that some pages might not display correctly.
I'd be interested to hear from other people if they know of any
downsides we're overlooking. Unless there are some significant
negatives though, I think the upside of promoting HTML5 makes this
switch a good idea.
David
David E. Ross
According to W3C's Philippe Le H�garet, HTML 5 is not yet ready for
prime time. It is still in discussion and flux. Thus, a browser that
implements HTML 5 may have to change its implementation again and again
until the specification matures. A Web page that is coded in HTML 5
will also have to be repeatedly updated. Would not the effort be better
expended in removing HTML 4.01 and XHMTL 1 errors?
See <http://www.w3.org/QA/2009/05/_watching_the_google_io.html>.
--
David E. Ross
<http://www.rossde.com/>
Go to Mozdev at <http://www.mozdev.org/> for quick access to
extensions for Firefox, Thunderbird, SeaMonkey, and other
Mozilla-related applications. You can access Mozdev much
more quickly than you can Mozilla Add-Ons.
Samuel Sidler
> According to W3C's Philippe Le Hégaret, HTML 5 is not yet ready for
> prime time. It is still in discussion and flux. Thus, a browser that
> implements HTML 5 may have to change its implementation again and
> again
> until the specification matures. A Web page that is coded in HTML 5
> will also have to be repeatedly updated. Would not the effort be
> better
> expended in removing HTML 4.01 and XHMTL 1 errors?
Sure, it's clear that HTML5 is still in flux and we very well may have
to make changes as needed to mozilla.org. But the parts of the spec
we'd actually be using are very, very close to HTML 4.01 and aren't
likely to change at all.
Switching to HTML5 also allows us to use new features that some
browsers have implemented (like <video>) and still validate against a
doctype.
There aren't many efforts needed to validate against HTML 4.01 since
many people have devoted effort to keeping the previous version of the
site validating against that spec and we haven't change that much.
Regardless, there's hardly any effort needed to switch to HTML5. It's
merely a change in doctype in our header include. There's a lot of
effort in fixing errors.
-Sam
Robert Kaiser
> I'm not sure of any negative issues with making this change except that
> some pages might not display correctly. But I actually don't believe
> that's true in 99% of our pages and we can fix any issues we notice
> along the way.
I don't think anything will display worse, as we render HTML5 in the
same render mode as strict HTML4, and I believe in terms of rendering
it's the same as XHTML strict as well.
As I said in the bug, I think it's a major mistake that we even did
switch to XHTML, as our content was never made ready for that, and we're
serving a lie rather than a standard right now, as XHTML served as
text/html is no XHTML. If I realized that switch before we went public,
I would have raised that before, but it flew under the radar.
The only sensible thing to do is to go HTML5 as we are promoting that
technology in all kinds of places and might want to use <video> et al.
in some places - and in addition, our HTML4 content is still valid as
HTML5 and using text/html as a MIME type is correct as well.
"Correcting" our content for /> which is unneeded in HTML is just a
waste of time, and have we ever intended as a group to really go XHTML
then we should have really prepared for it and not gone with the lie we
have gone with.
In summary: Let's go HTML5 ASAP, ideally yesterday.
Robert Kaiser
David E. Ross
> On Sep 6, 2009, at 10:25 PM, David E. Ross wrote:
>
One goal of the Web site should be to attract potential users of Mozilla
applications, users who are not yet using Firefox or other Gecko
browsers. Would this change cause problems for users of IE 6 or IE 7?
How about users of Safari or Opera? Given that SeaMonkey 2 is not yet
released for end-users, what impact would this have on users of
SeaMonkey 1.1.x?
Samuel Sidler
> One goal of the Web site should be to attract potential users of
> Mozilla
> applications, users who are not yet using Firefox or other Gecko
> browsers. Would this change cause problems for users of IE 6 or IE 7?
> How about users of Safari or Opera? Given that SeaMonkey 2 is not yet
> released for end-users, what impact would this have on users of
> SeaMonkey 1.1.x?
I believe all browsers without HTML5 rendering engines simply use
quirks mode to display anything with the "html" doctype. They won't
have any issues.
-Sam
Sander
> I believe all browsers without HTML5 rendering engines simply use
> quirks mode to display anything with the "html" doctype. They won't
> have any issues.
<!doctype html> will actually trigger standards mode (not quirks mode)
in all browsers. It was chosen as the doctype in HTML5 because it was
the shortest possible string which fulfills this requirement.
Sander
Samuel Sidler
That's even better since we'd then be conforming to standards mode in
all browsers, something we'd been doing with the HTML 4.01 doctype.
-Sam
davidwboswell
> How about users of Safari or Opera? Given that SeaMonkey 2 is not yet
> released for end-users, what impact would this have on users of
> SeaMonkey 1.1.x?
Why don't we change the doctype on the home page for a couple of days
and test? It's probably worth testing on a small sample of pages
before switching everything over anyway.
David
Samuel Sidler
That sounds like a good plan to me.
-Sam
davidwboswell
> > and test? It's probably worth testing on a small sample of pages
> > before switching everything over anyway.
>
> That sounds like a good plan to me.
We just made the doctype switch on the site. If it breaks things or
doesn't work for other reasons, we can revert this change.
David
Robert Kaiser
> We just made the doctype switch on the site.
\o/ \o/ \o/ \o/ \o/ \o/ \o/ \o/ \o/
Robert Kaiser