Safari 3.1 & Firefox 3 trumped by Webkit Nightly?

On 19-Mar-2008, at 07:24, Bill Rowe wrote:
> On 3/18/08 at 2:01 PM, eric.durbrowcomcast.net (Eric Durbrow Ph.D.)
> wrote:
>
>> On the Acid3 test it seems that the highest score has been reached
>> by Webkit rather than the latest Safari and Firefox beta. And Webkit
>> seems considerably faster and lighter than the other two. And after
>> 2 weeks, it has not crashed. And Webkit does seem to allow for some
>> Safari plug-ins like Inquisitor and AdBlock. So if we don't use many
>> extensions in Firefox is there any reason why users should consider
>> Safari or Firefox? What are webkit users missing?
>
> My understanding is Webkit is the framework used by Safari to
> render HTML. It isn't a browser at all. Consequently, saying
> Webkit is "faster and lignter" than Safari really doesn't make
> sense. Safari users are webkit users.

There is a webkit nightly build right now, it scores in the 90's on
acid3. Safari 3.1 score 75%.

http://nightly.webkit.com I beleive

At 06.24 -0700 2008-03-19, Bill Rowe wrote:
>On 3/18/08 at 2:01 PM, eric.durbrowcomcast.net (Eric Durbrow Ph.D.)
>wrote:
>
>My understanding is Webkit is the framework used by Safari to
>render HTML. It isn't a browser at all. Consequently, saying
>Webkit is "faster and lignter" than Safari really doesn't make
>sense. Safari users are webkit users.

Eric talked about "WebKit Nightly" (at least in the subject, but left out the "Nightly" part in the message) which is a special Safari version based on the very latest WebKit build, as opposed to the shipping regular Safari.

http://nightly.webkit.org/
http://webkit.org/blog/29/nightly-builds/

I think bitreader is right. I am presently using the latest build of Webkit which scores 93 on the Acid3 test. This is considerably higher than Firefox 3 or Safari 3.1. And when I go to the "about Safari" menu option it reports Safari 3.1. I don't quite understand what bitreader means by patches Safari. It certainly works in Safari 3.1/webkit nightly and appears as a preference panel in Safari. Perhaps from the user's perspective it doesn't matter whether it is input manager, patch, or plug-in.

And yes, Safari Ad Block (SafariBlock) does not work as well as Firefox but it seems to work well enough for me.

Perhaps a better question is whether it really matters if a browser scores very high on the Acid3 or has a second or two better speed. Given that Safari 3.1 and Firefox 3 have similar speed perhaps it will come down to extension usefulness... Thoughts?

On Mar 18, 2008, at 5:01 PM, Eric Durbrow Ph.D. wrote:
> On the Acid3 test it seems that the highest score has been reached
> by Webkit rather than the latest Safari and Firefox beta. And
> Webkit seems considerably faster and lighter than the other two.
> And after 2 weeks, it has not crashed. And Webkit does seem to
> allow for some Safari plug-ins like Inquisitor and AdBlock. So if
> we don't use many extensions in Firefox is there any reason why
> users should consider Safari or Firefox? What are webkit users
> missing?

WebKit users are missing the time they spent repeatedly downloading
WebKit instead of doing something else with their lives. I'm kidding.

Nightly builds are not going to have gone through the same quality
assurance testing that a Safari update has. Their site says "Nightly
builds of WebKit are provided for testing purposes only" for a
reason. The old definitions don't seem to apply anymore but in some
sense a nightly build is not even beta software.

This matters not only for application stability but also rendering. A
small (but growing) number of web designers might look at how Safari
renders a particular bit of CSS or supports some bit of JavaScript.
If WebKit does it differently, even if it's "better" or more correct,
it can produce unexpected results.

In terms of features, if you're running WebKit then you are in a
better position to compare WebKit and Safari. I just downloaded the
current build and glancing at the menus and preferences, it looks
just like Safari. It seems to share Safari's history, bookmarks, and
preferences, that's kind of a shame. It was fairly easy to make a
Firefox Beta launcher that points it at a different profile from non-
beta Firefox. I don't know how easy it would be do to the same with
Safari and WebKit.

On Mar 20, 2008, at 9:42 AM, Eric Durbrow Ph.D. wrote:

> Perhaps a better question is whether it really matters if a browser
> scores very high on the Acid3 or has a second or two better speed.
> Given that Safari 3.1 and Firefox 3 have similar speed perhaps it
> will come down to extension usefulness... Thoughts?

i tend to have groups of bookmarks that i open in tabs. if i open a
couple dozen tabs at once, i can start working with the first one in
Safari immediately. Firefox chokes terribly under that kind of load.
and at least up until v3 i couldn't get past how horribly fugly
Firefox looked out of the box. so for me Safari/Webkit Nightly is a no-
brainer.

Webkit Nightly users might want to check out NightShift, a nifty piece
of freeware that automates the nightly download process.

On Thu, Mar 20, 2008 at 2:42 PM, Eric Durbrow Ph.D.
<eric.durbrowcomcast.net> wrote:
> Perhaps a better question is whether it really matters if a browser scores very high on the Acid3 or has a second or two better speed. Given that Safari 3.1 and Firefox 3 have similar speed perhaps it will come down to extension usefulness... Thoughts?

Certainly. Google's Browser Sync is so useful to me that it makes
Firefox use a no brainer for me.
(<http://www.google.com/tools/firefox/browsersync/>)
Maybe one can obtain similar functionality with .Mac (though I don't
think so, since it seems limited to bookmarks syncing) and Safari, but
that's not a free service.

-Thomas

There's quite a bit more to browser compatibility than the Acid3 test scores.

Although WebKit now scores 93 out of 100 tests, I can't use it with
Jira or edit pages with Confluence based Wikis (both Atlassian
products). I also know that Safari has problems with Joomla too.
Whenever I use any of these products, I switch to FireFox.

Now, I know Confluence does specifically check for Safari browsers and
refuses to load their GUI editor if you're using Safari. I don't know
if there is something specifically in Jira that prevents Safari from
working.

It may come down to a problem with TinyMCE which is a standard GUI
editor for blogs, wikis and CMS systems. I know JCE on Joomla has
similar issues, but may be based upon TinyMCE code. That might explain
Joomla and Mambo issues, and it might explain why Confluence simply
refuses to load TinyMCE if you have Safari. However, it doesn't
explain the problems Jira has.

As I once told someone, there are two web standards: 1). What the web
standards community decrees (WaSP and WC3), and 2). Whatever Internet
Explorer does. For now, most webpages seem to follow the second
definition of "web standards".

--
David Weintraub
qazwartgmail.com

On 20-Mar-2008, at 07:42, Eric Durbrow Ph.D. wrote:
> Perhaps a better question is whether it really matters if a browser
> scores very high on the Acid3 or has a second or two better speed.
> Given that Safari 3.1 and Firefox 3 have similar speed perhaps it
> will come down to extension usefulness... Thoughts?

Now that the dunderheads in Redmond are actually taking a serious look
at standards, full compliance with acid3 will be more and more
important. The fact is, the web is still struggling in its infancy,
and has get to shake itself loose of all the crap that was created by
Netscape and Microsoft. Once we have real stable standards that
everyone follows, the web is going to get a whole lot more useful and
interesting.

On 3/20/08 at 6:42 AM, eric.durbrowcomcast.net (Eric Durbrow Ph.D.)
wrote:

>I think bitreader is right. I am presently using the latest build of
>Webkit which scores 93 on the Acid3 test. This is considerably
>higher than Firefox 3 or Safari 3.1. And when I go to the "about
>Safari" menu option it reports Safari 3.1. I don't quite understand
>what bitreader means by patches Safari. It certainly works in Safari
>3.1/webkit nightly and appears as a preference panel in Safari.
>Perhaps from the user's perspective it doesn't matter whether it is
>input manager, patch, or plug-in.

I suspect for most users, the difference between a plug-in and
input manager are moot. A plug-in makes use of documented
interfaces and doesn't patch Safari. An input manager on the
other hand is more akin to the INIT files that did neat things
in OS9. And as a result is somewhat more risky. But because
input managers do patch applications, it is possible to achieve
results that simply cannot be done with plug-ins.

>Perhaps a better question is whether it really matters if a browser
>scores very high on the Acid3 or has a second or two better speed.
>Given that Safari 3.1 and Firefox 3 have similar speed perhaps it
>will come down to extension usefulness... Thoughts?

If the browser renders pages for sites you visit acceptably,
then it should not matter to you how well it scores on Acid3.

As far as extensions, the more you use the more impact there
will be on the speed of rendering pages. There is a very
definite trade off here.

At 06:48 -0700 UTC, on 2008-03-21, David Weintraub wrote:

> [...] Safari has problems with Joomla [...]
>
> Confluence does specifically check for Safari browsers and
> refuses to load their GUI editor if you're using Safari. [...]
>
> It may come down to a problem with TinyMCE [...] JCE on Joomla has
> similar issues [...]
>
> As I once told someone, there are two web standards: 1). What the web
> standards community decrees (WaSP and WC3), and 2). Whatever Internet
> Explorer does. For now, most webpages seem to follow the second
> definition of "web standards".

You might be interested in The Web Repair Initiative, see
<http://webrepair.org/>, which sort of tries to fill that gap. The primary
goal is to improve the quality of the output of such systems, but of course
the usability/accessibility of the (content) management interface is of
interest as well.

Btw, while FCKEditor and TinyMCE are perhaps the best known, there are quite
a few more such embeddable editors tat can be used by Web publishing systems.
The WRI lists 27 such editors:
<http://webrepair.org/strategy/known%20systems?SwitchFunction=on&FunctionIs=inline%20editor>.
If you are lucky enough to be the admin of a problematic CMS, it might be
worth giving another editor a shot. (Not that every editor can that easily be
embedded in every publishing system...)


--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

As I stated before, there are two browser standards: One set by the
various standards bodies and the other is whatever IE does.

The biggest problem Safari has right now is Midas/designMode support.
This is something cooked up by Microsoft, but followed by Mozilla
Foundation for FireFox and now Opera. Although not an official web
standard, it is used by Wordpress, TinyMCE, Joomla, Mambo, and many
other web editors. I also suspect the problems Safari has with Jira
and Confluence are in a similar vein.

I understand that there is now a new (and special) version of TinyMCE
that works with Safari. Websites need to download this version and
configure their sites, so that Safari browsers use the new Safari
compatible TinyMCE. I don't know how many have done this. Confluence
is merely testing for Safari, and if you're using Safari, just doesn't
give you the Rich Text editor.

I still have problems with Safari with the latest version of Joomla
although if I replace the TinyMCE editor with JCE, it seems to work
better with Safari.

I like Safari, and I even use it on Windows now because I like the way
the font renders. However, I know there are many sites where I can't
use it. When I need to surf one of those sites, I have to use FireFox.

--
David Weintraub
qazwartgmail.com

At 4:22 PM -0700 3/21/2008, David Weintraub wrote:

>As I stated before, there are two browser standards: One set by the
>various standards bodies and the other is whatever IE does.
>

David, et. all:

The following link nicely explains IE8, W3C standards and all that.
The headphone discussion is just an intro to the main topic.

http://www.joelonsoftware.com/items/2008/03/17.html

--Jim

> The following link nicely explains IE8, W3C standards and all that.
> The headphone discussion is just an intro to the main topic.
>
> http://www.joelonsoftware.com/items/2008/03/17.html

The blog entry referred to of Joel Spoelsky has ignited quite a stir.
It should not be read without at least knowledge of Mark Pilgrims
interesting and funny comments (hat tip: John Gruber,
daringfireball.net):

http://diveintomark.org/archives/2008/03/18/translation-from-ms-speak-to-english-of-selected-portions-of-joel-spolskys-martin-headsets

Best,

Kim

On Sat, Mar 22, 2008 at 6:04 AM, Jim Carr <jimcarrmac.com> wrote:
> At 4:22 PM -0700 3/21/2008, David Weintraub wrote:
> David, et. all:
>
> The following link nicely explains IE8, W3C standards and all that.
> The headphone discussion is just an intro to the main topic.
>
> http://www.joelonsoftware.com/items/2008/03/17.html

I understand the issue, and I know people who are working on IE 8 and
really are attempting to at least get the sucker to pass the Acid2
test.

But, when you're writing webpages, you write webpages for everyone
else, and webpages for IE. If IE didn't control so much of the market,
many webpage authors would simply put IE users over to a link that
says "This page doesn't work with IE. Get a real web browser" and be
done with it. Unfortunately, IE controls 80% of the market, so you
design your webpage, makes sure that FireFox and Safari render it
well, then you put in the special tweaks for IE. As Joel explained,
it's one of the reasons why IE8 is having so many problems. Sites see
it as a Microsoft IE browser, and then do the tweaks for IE instead of
the standard pages like it desires.

Then there are things Microsoft created like Midas/documentMode.
Because so many sites depend upon it, other browsers like FireFox do
their best to make their browsers work with them even though they are
not official specs. Heck, I've seen the specs for Midas/documentMode
from Microsoft, and they're pretty poor. I think Mozilla rewrote them,
so they can be "compatible" with them, and Opera is taking the Mozilla
specs.

The problem is that WC3 hasn't adopted Midas/documentMode and Acid3
won't reflect them. You can be a purist on this issue (and from what
I've seen in the WebKit forum, WebKit took the purist position for way
too long). But, the problem is that so many sites use modules like
TinyMCE that depends upon Midas/documentMode that people will simply
think Safari is the problem, and not lose web standards.

I am glad that test like the Acid3 test will help browsers conform. I
am thrilled that the WebKit team is leading the race towards
compatibility, but Midas/documentMode isn't an official standard, so
the Acid3 test score won't reflect it. Unfortunately, even though
Safari hues closely to "the standard", many people won't use it
because it simply doesn't work with their websites.

By the way, the latest build of WebKit fails to do a drop cap on my
site that all other browsers, even Safari 3.1 render. I've reported it
to WebKit as a bug.

--
David Weintraub
qazwartgmail.com

> The following link nicely explains IE8, W3C standards and all that.
> The headphone discussion is just an intro to the main topic.
>
> http://www.joelonsoftware.com/items/2008/03/17.html

Do keep in mind that, dispite his overall credibility, Spolsky has
occasionally been accused of being a Microsoft fanboy (just like
Gruber, dispite _his_ overall credibility, gets his fair share of Mac
fanboy accusations). To put Spolsky's view on web standards (by the
way: no quotes necessary, Joel!) in perspective, then, you might also
want to read this "slightly" sarcastic blog post:

<http://diveintomark.org/archives/2008/03/18/translation-from-ms-speak-to-english-of-selected-portions-of-joel-spolskys-martin-headsets
 >


And don't miss the comment linked to in the update to that post:

<http://diveintomark.org/archives/2008/03/18/translation-from-ms-speak-to-english-of-selected-portions-of-joel-spolskys-martin-headsets#comment-11565
 >

(Caution: both come with a mild dose of strong language...)


Regards,

Jochen.

On Mar 23, 2008, at 4:13 AM, David Weintraub wrote:
> As Joel explained,
> it's one of the reasons why IE8 is having so many problems. Sites see
> it as a Microsoft IE browser, and then do the tweaks for IE instead of
> the standard pages like it desires.

And even sites that tried to prepare for IE8 were blindsided by the
lateness of Microsoft's correct decision that IE 8 would default to
standards mode. Until just before release, it was going to default to
IE7 mode.

This should be mostly moot right now, as most users have no good
reason to have the IE 8 beta. (I gulped and installed it, since my
Vista machine is even more a sandbox than my Mac laptop is. And IE7
mode seems to achieve being IE7, at least on the sites that don't like
IE8 mode that I've tried. (twitter.com being one of those--the text
contents align oddly, with most but not all being well to the right of
the images--or that was the case the one time I tried it.)

     --John

At 4:13 AM -0700 3/23/2008, Kim Gammelgård wrote:

>>The following link nicely explains IE8, W3C standards and all that.
>>The headphone discussion is just an intro to the main topic.
>>
>>http://www.joelonsoftware.com/items/2008/03/17.html
>
>The blog entry referred to of Joel Spoelsky has ignited quite a stir.
>It should not be read without at least knowledge of Mark Pilgrims
>interesting and funny comments (hat tip: John Gruber,
>daringfireball.net):
>
>http://diveintomark.org/archives/2008/03/18/translation-from-ms-speak-to-english-of-selected-portions-of-joel-spolskys-martin-headsets
>
Kim:

I'm wasn't saying Joel is the final word, just that it illustrated
the confusing situation. As does the discussion that followed.

In the real world, I see very clear differences between how browsers
work. And that isn't even counting sites that are IE or they don't
work. Or do something useless.

The management company for the HOA I live in has a login section
where you can check your account. Unless you are using IE, it will
simply redirect you to <http://www.microsoft.com/en/us/default.aspx>
where you will not find a Mac version of IE.

I don't do any browser detection on websites I am responsible for.
And I spend quite a bit of time trying to explain to Windows folks
that I can't add words in some places where IE shows plenty of room
and other browsers show type breaking and over-printing next line.
Not to mention settling for less than ideal spacing around images in
some cases because I want type to clear the image on all browsers.

--Jim

On Mar 24, 2008, at 4:16 AM, johnbaxterlistsmac.com wrote:
> (twitter.com being one of those--the text
> contents align oddly, with most but not all being well to the right of
> the images--or that was the case the one time I tried it.)

Checked later Sunday--(I did not write the above at 04:16 Monday)--
Twitter text alignment in the tweets is sensible in IE8 now, but IE 8
cuts off the bottom of the remaining character count above the update
text box.

   --John (who does not run nightly builds, even though our Lt Gov
will be knighted by the Spanish in a few days)

I really don't understand all the stuff MS is doing in order to
operate in "quirk" mode and the versioning garbage.

It's really very simple: Call IE8 another user agent name. Right now,
IE's user agent looks like this:

    Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0)

All they need to do is have a mode switch on IE8 which sets the
browser to follow the standards, and sets the following user agent:

    Explorer/8.0 (MS-Exp 8.0; Windows NT 6.0)

No muss or fuss. No major change in web standards. Websites that watch
for IE as a browser will simply treat this as a non-IE browser when
IE8 is in standards mode. If the user switches IE8 into broken... I
mean IE7 compatibility mode, just set the user agent back to the old
mode.

Websites are happy, standards committee is happy, users are happy,
Microsoft is happy.

Wait, scratch the last one. Microsoft isn't happy unless it can flex
its muscles and make everyone else sweat.

--
David Weintraub
qazwartgmail.com

At 04:59 -0700 UTC, on 2008-03-25, David Weintraub wrote:

[... IE mess]

> All they need to do is have a mode switch on IE8 which sets the
> browser to follow the standards, and sets the following user agent:
>
> Explorer/8.0 (MS-Exp 8.0; Windows NT 6.0)

Something like that would probably help, yes. But only partly, as it would
only affect browser sniffing, which is just one method w3bd335ign3r5 use to
break the Web. It would do nothing for all those sites that (needlessly) rely
on HTML/CSS parsing bugs, and/or use conditional comments in stupid ways.

The only right way to author for the Web while taking IE into account:

<!--[if !(lt IE 8)]><!-->
stuff for any user agent that is not IE pre-8 here
<!--<![endif]-->

<!--[if lt IE 8]>
stuff for IE pre-8 here
<![endif]-->

There is no need for browser sniffing, nor for messing about with 'CSS
hacks'. Just use conditional comments, the one positive thing M$ has done for
the Web so far.


--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>