News for hackers. Stuff that matters.

A comment on Browsers. Parent comment

Re:Browsers by austincheney on Sunday December 16, 2018 @10:27PM (#18696724)
IE, at its height, was available outside windows. https://en.m.wikipedia.org/wiki/Internet_Explorer_for_Mac

At that time, 2001-2002, IE6 was highly praised. You don’t get to 96% marketshare by being something everybody hates. The hate came later when CSS demand ramped up and it became clear IE6 used a non-standard box model. Worse still MS essentially abandoned the browser, the result of total marketshare dominance, and tied it directly to the operating system.

The reason people are complaining about this is the lack of diversity. This has proven very bad in the past. It essentially makes the platform a proprietary space even if the dominant browser is open source. The key here is who directs the technology and direction of the platform.

[ Parent | on HN ]
Re:Browsers by mdasen on Monday December 17, 2018 @12:52AM (#18697144)
It was sort-of available for Mac. IE for Windows was based on the Trident layout engine. The Mac version was based on the Tasman layout engine: https://en.wikipedia.org/wiki/Tasman_(layout_engine). While they were both branded as "Internet Explorer", they weren't the same browser.
[ Parent | on HN ]
Re:Browsers by kitsunesoba on Monday December 17, 2018 @01:16AM (#18697234)
Tasman was actually quite good, too. It was a great deal better than trident in many regards, ranging from the ability to properly handle transparent PNGs to more complete and correct CSS support. IE for Mac was a great browser for years after IE 5/6 had grown long on the tooth. I always wondered why it didn’t get ported to Windows to replace Trident.
[ Parent | on HN ]
Re:Browsers by wvenable on Sunday December 16, 2018 @10:51PM (#18696810)
> IE6 used a non-standard box model.

Which we all use now anyway because it was right way to do it...

    * {
      box-sizing: border-box;
    }
[ Parent | on HN ]
Re:Browsers by yuchi on Monday December 17, 2018 @01:48AM (#18697326)
No no no, that's a completely different issue — we are talking about hasLayout stuff here.
[ Parent | on HN ]
Re:Browsers by austincheney on Monday December 17, 2018 @12:33AM (#18697072)
That is a matter of opinion.

I don't do that and I avoid code that does. In the days of IE7 I learned to write CSS code that was conformant to both box models. Now I just write to the standard box model. CSS isn't something that will ever be particularly challenging once you have gone through the extremes of edge case use cases.

[ Parent | on HN ]
Re:Browsers by wvenable on Monday December 17, 2018 @11:29AM (#18699932)
It's not a matter of opinion, content-box is much more difficult to work with when using fluid layouts than border-box which is why all major CSS frameworks use border-box now.

At the time, the CSS specification didn't make it clear which way to calculate widths so Microsoft, probably due to their experience building graphical layout software, chose border-box while all other rendering engines at the time went content-box probably without considering the consequences. Realizing the error, the box-sizing property was added to CSS to fix this.

[ Parent | on HN ]
Re:Browsers by austincheney on Monday December 17, 2018 @09:22PM (#18704066)
You claim it’s not a matter of opinion and due to personal opinion. Weird. Then you invent a history from which to draw imaginary conclusions. It is perfectly ok to have a personal opinion on CSS. You don’t have to justify it and you certainly don’t have to conjure a dillisional alternate reality.

Imagination aside, IE3 was the first browser to offer any CSS support.

[ Parent | on HN ]
Re:Browsers by earthboundkid on Tuesday December 18, 2018 @12:39PM (#18708592)
You're just wrong. The CSS box model as specified by the standards was totally unusable for making a website where you want boxes to have certain sizes with certain margins and paddings, especially in the days before you could use calc(), so you couldn't mix pixels with percentages. It's true that it was the standard, and so MS should have implemented the standard, but it's also true that what MS shipped made way more sense and every frontend dev turns on their box model when it box-model: border-box became part of the standard.
[ Parent | on HN ]
Re:Browsers by robin_reala on Monday December 17, 2018 @12:49AM (#18697134)
Couple of corrections: IE6 actually fixed the box model (that was the difference between the new standards mode vs quirks mode that matched IE5), and IE on the Mac used a completely different rendering engine (Tasman instead of Trident).
[ Parent | on HN ]
Re:Browsers by austincheney on Monday December 17, 2018 @01:38AM (#18697292)
The box model issue was independent of quirks mode. Quirks mode was triggered by HTML not conformance to a modern doctype (HTML4, XHTML 1.1, or one of the three XHTML 1.0) which produced a different DOM tree and thus different layout/appearance when CSS was applied. This was particularly frustrating because even when MS fixed the corresponding CSS defects their DOM tree remained non-standard until IE9.

I remember having to deal with all these things when working with the company's aging CMS. It was a great tool, but the version the company was using was old (before standardized doctypes in 1999) and inserted a comment into the top line of code throwing all documents into quirks mode in IE.

The box model was corrected when IE released version 7, which still also featured quirks mode rendering for backwards compatibility.

[ Parent | on HN ]