IE7 Beta Released

Internet Explorer 7: Now in beta testing for developers

[chadmg]

[microsoft.com...]

Improved design to make everyday tasks easier and faster, with better navigation through tabbed browsing; inline search right from the toolbar; shrink-to-fit Web page printing; and a streamlined, redesigned user interface (currently in its early stages in Beta 1).

And they REALLY like RSS feeds according to their next improvement.

New tools to take you directly to the information you want through support for Web feeds (RSS) that includes automatic discovery of web feeds (RSS) on Web pages, basic Web Feed (RSS) reading capabilities, and basic support for saving Web feeds (RSS) as a new kind of favorite.

Anybody want to share their MSDN subscription? ;)

[Hester]

While that's true in terms of IE being the majority browser, many top designers feel it is actually easier to code for Firefox first, then move on to IE. The reason is that you don't encounter IE's infamous layout bugs, which of course don't exist in Firefox. You can program to the standards and fix the page in IE afterwards when it's almost ready to go live. Otherwise you can waste a lot of time in IE avoiding bugs.

[Farix]

The graph "Web Browsers Used to Access Google" found at [google.com...] is clear enough, developers should code for IE and - if necessary - tweek a bit for Firefox.

Popularity means squat when it comes to the development end of things. The old adage, "send strict, accept loose" is what's at play here. By using Firefox, developers are more then likely to "send strict" because Firefox is much more likely to display markup problems then IE ever will. The tweaking only comes in when IE hasn't implemented something from the W3C specifications or has implemented it incorrectly.

Remember, the W3C specifications--warts and all--are the gold-standard for which browser implementation and rendering are compared to. Not what IE sets as a standard--or lack there of.

[zafile]

What does "standard" mean?

The page [google.com...] also helps. The graph "Operating Systems Used to Access Google" is key to find a more accurate answer.

To obtain 2005 stats, I adapted older Zeitgeist statistics and mixed some numbers at [news.com.com...] . Then, I made a projection.

The numbers obtained "describe something very precise". More than 90% of the world's computers employ the Microsoft's operating system. Obviously, this is the "standard".

Some organizations claim to follow the standard without taking account the 90% plus figure.

"You can program to the standards and fix the page in IE afterwards."

I rather program to the 90% plus figure.

[Robin_reala]

Zafile - you're building in forwards-incompatibility by doing that. Every current browser manufacturer (including Microsoft) are building their browsers to conform to standards. You're thinking too short term.

[Farix]

What does "standard" mean?

As far as HTML and CSS is concerned, it's whatever the W3C says it is. They are the group recognized by all browser makers, including Microsoft, as the ones who defines the standards.

[zafile]

"you're building in forwards-incompatibility by doing that".

Since I follow the 90% plus standard, I only use the latest authoring tools made by Microsoft.

[Farix]

Then you write inefficient markup.

[JAB Creations]

Hester....

One - AAA accessiblities and usability require the use of many things, espescially :hover. You can NOT use onmouse JS attributes to emutlate hover effect in MSIE. Unless you want to bloat your code with various hpyerlink (as you can not really slim down the a attribute in CSS when you clean it up) you're going to want to do menus a different way.

The idea is to make the page load fast (dialup), be accessible to the blind and deaf (should be normal standard but that becomes a strong selling point at least) and to be as eye candy as possible (via rollover effects in example.

Two - I clean both SP2 and non-SP2 systems and regardless of how they vary case to case, the way they fall from grace is always the same.

I'll reply to anything anyone asked later as I'm traveling tonight...

[drhowarddrfine]

zafile,
As of Friday, your "90% standard" is correcting itself to follow the W3C. At least in beta2 of IE7. So if the "90% standard" is the one to follow, why are they changing to the W3C standard? Why have they joined with WASP to learn where they must concentrate to "fix" IE7 toward the standards (to paraphrase their statement).

See the IEBlog and read about how much they realize IE is far off the course for errors and bugs and how they desperately want to fix this so they will follow the W3C.

[encyclo]

We're getting rather off-topic here, I feel. No-one is denying that IE6's enormous market-share means that pages must be made to display well in that browser. From then on, many prefer to work in Firefox and edit for IE, others prefer the reverse - each must find their own preferred methods of working.

But the advent of IE7 will mean a much greater consistency accross browsers, as all the major browsers (IE7, Firefox, Safari/Konqueror and Opera) will closely follow the defined standards in both HTML 4.01 and CSS2. It will be much easier to develop in the browser of your choice, with only minor adjustments for the others, followed by corrections for IE6.

[Farix]

Back on topic, there is more news from Chris in the comments section:

As for the "* HTML" selector issue - actually, it's currently fixed (that is, it no longer works) in beta 2; however, I'm on the fence as to whether we should ship that (it does help our appearance on the acid2 test), since it is in use in the web today for browser switching. I'd welcome feedback on whether we should fix it in IE7 or not.

So go over there and let your opinions be known.

[MatthewHSE]

Farix, do you have a link to that article? I can't seem to find it myself and a search is yielding no results.

Personally, I say axe it. It's only being used as a hack, and if IE7 doesn't require hacks, it would be nice if it didn't read them, either. On the other hand, we won't know what hacks it does require until after it's been out awhile.

Incidentally, I've been toying lately with the idea of whether or not it'd be helpful if every browser had some sort of unique "simple hack" similar to

* html

so you could target it specifically. On the whole I'm not tickled with the idea, but it would open up some interesting possibilities. But it will never happen anyway.

[Wertigon]

The graph "Web Browsers Used to Access Google" found at [google.com...] is clear enough, developers should code for IE and - if necessary - tweek a bit for Firefox.

Ah, yes... The "latest" browser stats, now more than a year old. The stats listed are from before FireFox 1.0 was released - Since then FireFox has gotten ~70M more downloads, which certainly has put a dent in Microsoft's market share. To quote a very well known American - "There are three kinds of lies; lies, damned lies and statistics." 'Nuff said about that.

Coding for a specific browser is a bad idea - You code for the webstandards which all browser vendors (Yes, even Microsoft, even though IE6 has some really annoying bugs) do a fairly good job at supporting. I'm not saying Microsoft's PERFECT by any means, and they deserve all the flack they get for keeping back the web, but IE6 is still much more standard-compliant than, say, NS4.

Of course, IE compability is important, but IE is *one* browser among *many*.

Back to the topic at hand, I'm definitely looking forward to IE7, even though I've personally moved on to run Linux exclusively. And I'll be one of the first to applaud the IE7 team when the full release comes out - they've done a GREAT job on their browser. :)

[edited by: encyclo at 8:17 pm (utc) on Aug. 2, 2005]
[edit reason] appropriate language please, see TOS [/edit]

[drhowarddrfine]

"IE6 is still much more standard-compliant than, say, NS4."
You might as well have said IE4, they're both long dead.

"..they've done a GREAT job on their browser."
In what way?

[JAB Creations]

Saying something is better then Netscape 4 is like saying my computer is better then one in the garbage because I have a 1 megabyte of fast page memory.

[Robin_reala]

OT, but "Lies, damned lies, and statistics" is actually a quote of Disraeli, a Victorian-era British prime minister.

[Wertigon]

"IE6 is still much more standard-compliant than, say, NS4."
You might as well have said IE4, they're both long dead.

Yeah, but I dare say that NS4 when it was released was even less standard-compliant for it's time than IE6 is now.

IE6 has a few quite annoying bugs, but most of them can be overcome by a few clever hacks or IE-specific code (like conditional comments), and the rendering bugs atleast doesn't crash the browser. Compare that to NS4 where you actually could get the browser to crash by using 100% standard-compliant CSS code.

"..they've done a GREAT job on their browser."
In what way?

Well, for one, they seem to have fixed most if not ALL the known rendering bugs that exists in IE6, *and* added support for quite a few more things, like :hover on all elements and position : fixed;. Read SuzyUK's and Tedsters posts on that matter. Or Google for "chris ie7" to read the original blog post.

I'm not saying I'm trusting Microsoft, but I'm willing to believe that they might actually come around on this issue. I don't think Microsoft will have a next-to-monopoly in the future, I seriously don't, and it might just be that the people at Microsoft are starting to realise this.

However, I'm still on my guard and it wouldn't surprise me if it turns out all of this is a smokescreen. But if IE7 actually fixes everything Chris Wilson says it does... Yea, I'll be right there in the front and applaud them. And that's not a bad praise considering I'm a 100% Linux guy nowadays. :)

[CritterNYC]

The graph "Web Browsers Used to Access Google" found at [google.com...] is clear enough, developers should code for IE and - if necessary - tweek a bit for Firefox.

Those statistics are over a year old. Firefox 1.0 wasn't even released until September 2004. I compile a stats breakdown based on OneStat, WebsideStory and TheCounter on a monthly basis. While the individual stats themselves can be problematic, the trends they show when combined can be very handy.

+-----------------------------------------+
¦.Rendering Engine.¦.July.2005.¦.Mar.2005.¦
¦------------------+-----------+----------¦
¦.MSHTML-Modern....¦...82.81%..¦..79.99%..¦
¦.MSHTML-Legacy....¦....4.88%..¦...9.23%..¦
¦.Gecko............¦....9.05%..¦...7.78%..¦
¦.KHTML............¦....1.19%..¦...0.95%..¦
¦.Opera............¦....0.73%..¦...0.66%..¦
¦.Netscape.........¦....0.08%..¦...0.09%..¦
¦.Other/Unknown....¦....1.26%..¦...1.31%..¦
+-----------------------------------------+

From this, we can see that Internet Explorer has been slowly but steadily losing users to Gecko (primarily Firefox), losing another 1.5% of all internet users in just the last 4 months. Most of those users seem to be coming from defections from MSHTML-Legacy (Internet Explorer 4 and 5), but some are also coming from IE6 based on a great deal of anecdotal evidence. We can also see that Safari has been picking up users. Opera has picked up a couple as well.

Of course, that doesn't even get into the fact that it is best to code to standards first. If you code to solid standards with the correct doctype, you'll usually be fine in Firefox, Opera, Safari, etc without any tweaking. Getting it to work in IE6 will involve just a couple minor changes. And then you'll need to apply the box model hacks, etc to get IE5/5.5 working as expected.

When a web designer says they are building a site for Internet Explorer... they're usually building a site for Internet Explorer 6.0 on Windows. Which means it will work for a bit under 83% of online users. If you want to add in IE5, you gotta start applying some hacks... or start with an older IE5-friendly design. Even then, you'll still be ignoring over 11% of visitors. Most IE designers never get around to testing on anything else.

On the other hand, if you start with standards and testing in Firefox (or Opera or Safari for that matter). You'll have that full set in the bag to start with. And can then apply the few hacks needed to get both IE5 and IE6 working just as well.

In the end, it saves time to start from standards AND your site will be usable by more people.

NOTE: Rendering Engine Explanations
The rendering engines are combined, partially due to the necessity of the statistics, in the following manner:
MSHTML-MODERN: Microsoft's current engine used in Internet Explorer 6.0
MSHTML-LEGACY: The engine from pre-IE 6. (a rough combination based on necessity of the stats)
GECKO: The engine used by Mozilla Suite, Firefox, Kmeleon, Galeon, Camino, Netscape 6/7, etc
KHTML: The engine created within Konqueror and used in Safari
OPERA: The Opera browser's engine
OTHER: All the rest that don't fit into the above along with misidentifications, no useragent string, some bots, etc

Where a given stat doesn't break down into finer detail, the other stats are used to break the meta result down. For instance, if one only gives an Internet Explorer number... one that uses IE4/5/6 is used to break it down into relative numbers.

[jerrylim]

anyone know where IE7 is downloadable? I'm searching for it.. :/

[Robin_reala]

Youll need an MSDN subscription. Unless you want to go via slightly more illegal channels.

[2by4]

IE 7 has flown the coop: search: ie 7 download

and you'll have no trouble finding a copy. It was out in the wilds literally days after the official msdn release.

[encyclo]

Like I said earlier in the thread, I would strongly discourage using any "unofficial" (ie. pirated) version of IE7 beta 1 - or anything that claims to be IE7 beta 1. You just don't know what you are really downloading, the beta 1 is not a good guide to the features in the future IE7, and there will be an open public beta available soon anyway via official channels.

[2by4]

The tabbed browsing is decent, but not all the bugs they claim as fixed are fixed. Still lots of work, but gives a decent idea of how it's going in general. It's raw, which is to be expected.

[lisa_w]

Mozilla has a wide variety of plugins to suit you browsing needs, does IE plan to follow suit? Or does Microsoft prefer to keep things simple for the basic user?

[JAB Creations]

The more share they loose the more they'll do. No -- unless you're talking about spyware, then you'll still have an immense (non)choice to pick from. They occasionally release a tool here or there (powertools) but nothing too grand or useful. IE users have to depend on people they don't know nor can trust for additional IE features.

[collymellon]

Fact - AAA accessible web-sites do not work in MSIE...

JAB I read everyone of your posts in detail and find very good info out of them - but the above fact you have stated is simply not true.

I'm interested to hear why you think AAA sites don't work in IE.

[MatthewHSE]

It's in Message 68:

One - AAA accessiblities and usability require the use of many things, espescially :hover. You can NOT use onmouse JS attributes to emutlate hover effect in MSIE. Unless you want to bloat your code with various hpyerlink (as you can not really slim down the a attribute in CSS when you clean it up) you're going to want to do menus a different way.

I don't know anything about it myself, but if the :hover capabilities on elements other than <a>'s are important for AAA sites, then, indeed, IE6 and lower just don't cut it.

<edit> I'm not JAB under a different username, but I thought he wouldn't mind if I quoted his previous remarks! ;) </edit>

[Robin_reala]

I'm pretty certain JAB is wrong on this one Matthew although if he'd like to back up his assertion then I'll be more than happy to look into it further. I've done a fair bit of accessibility work and I don't recall anything in WCAG1.0 that specifies that a 'hover' effect is necessary.

[collymellon]

That dosen't justify stating IE can't run AAA sites - I know they can (I work for the largest disabled access company in UK and had to build the site).

If :hover is needed anywhere else other than the <a> element - I simply accept I cannot use it. If certain features of a site have to be removed to conform to the AAA standards then so be it - Yes I agree IE's CSS support is simply not good enough in a world of accessibility standards.

More to the point if I get a site that needs to be replicated/re-built to AAA standard it will be rebuilt to them standards and pass all required HTML/CSS/AAA validation and run perfect in IE/FF/NS/screen readers etc. I just can't accept the statement that 'IE cannot run AAA sites' after spending so long running AAA sites in IE.

[JAB Creations]

Ok let me clarify as the devil is in the details! A good interactive website can not be AAA accessible in IE. That is of course IE6 and below as we have still yet to see the final release of IE7 with some big promises looming over our heads. One of example is the need to use the :hover attribute. Hacks do ~NOT~ count even as they may pass validation. The browser must be judged on it's W3C compliancy. At least in my opinion. Gotta have a selling point. ;-)

Thanks MatthewHSE...yeah I didn't read your post btw. I'm happy to clarify previos remarks, even if I'm wrong as to correct myself so I'm quite happy to have someone watching over me.

The new version of my own site will have extensive serverside scripting in order to suppport legacy browsers such as IE6 (including CSS hacks) as I'll be removing the current JS and doing everything as W3C and 508/AAA as much as possible (if not completely passing). I love a challange!