Chakra: Interoperability Means More Than Just Standards

re: Chakra: Interoperability Means More Than Just Standards

grinder mill — Thu, 02 Sep 2010 03:30:34 GMT

haha

re: Chakra: Interoperability Means More Than Just Standards

Mitch074 — Wed, 01 Sep 2010 09:31:53 GMT

@André R.: +1

re: Chakra: Interoperability Means More Than Just Standards

André R. — Wed, 01 Sep 2010 08:29:57 GMT

> function declarations anywhere in a statement

Simple, allow with a notice or warning in console if your running IE9 mode and disallow totally in "use strict"; mode.

re: Chakra: Interoperability Means More Than Just Standards

Kevin — Tue, 31 Aug 2010 16:11:44 GMT

Your first example about the /]/ vs /\]/ RE inadequately demonstrates what you want to convey. First, you admit that the standard specifically says /]/ is illegal, you tell us why, then you imply this is a form of "consensus feature." You define "consensus feature" as an implemented "feature" that's not part of a standard. You saying the illegal /]/ (by the standard) is a feature not part of a standard conflicts with itself.

For this example it's best to note that this broken-ness is ultimately the website implementor's fault for not coding for standards, but also the browser maker's fault for also not implementing the standards. In this example, neither party followed the standards, and thus both have made this standard worthless. Implementing /]/ as if it were legal simply because there isn't a matching [ and thus it is parsable is a standards implementation issue and ultimately not relevant.

re: Chakra: Interoperability Means More Than Just Standards

Mitch074 — Sun, 29 Aug 2010 08:11:33 GMT

@Allen Wirfs-Brock: I didn't mean that IE copies the other browser (you'd have integrated Webkit's KHTML if that were the case), but that in cases where practices deviate from the standard in a not-too-dangerous way, you do as other browsers do. Which is not so bad, after all.

However, I'd really like an answer on my idea of warnings in case of standards-deviating code that is still parsed and run by IE - is it possible? Can it still be implemented? That would be darn useful if built-in.

re: Chakra: Interoperability Means More Than Just Standards

Brophy — Sun, 29 Aug 2010 01:08:34 GMT

"The whole codec war is over anyway. With MPEG-LA's announcement of royalty-free streaming for non-pay content forever, WebM is stillborn."

Okay. There's not much I can do for you here. It's immensely disappointing that you seem to profoundly misunderstand what the MPEG LA's "announcement" actually means. Whether you're paying for it or not, it is the MPEG LA's contention that no usage of H.264 can occur without the largess of their licencing grace. On the bright side, it is with some considerable amusement that I watch people actually defend the abject absurdity of licencing required for the mere *transmission* of a file over the Internet. It really is quite an astonishing case of Stockholm Syndrome.

re: Chakra: Interoperability Means More Than Just Standards

Clue — Sat, 28 Aug 2010 15:01:08 GMT

<<does this mean you'll be deciding to bundle WebM in IE9 by default>>

Of course not. Instead, they'll wait for one of their competitors to bundle it (Google, prolly), wait until they distribute a few million copies, and then have MPEG-LA sue for billions.

The whole codec war is over anyway. With MPEG-LA's announcement of royalty-free streaming for non-pay content forever, WebM is stillborn.

re: Chakra: Interoperability Means More Than Just Standards

Brophy — Sat, 28 Aug 2010 11:27:04 GMT

"How do we decide whether to implement a feature that isn’t included in a standards specification? Like all browser providers, we often have to make this decision."

Extending this line of reasoning, does this mean you'll be deciding to bundle WebM in IE9 by default?

re: Chakra: Interoperability Means More Than Just Standards

Allen Wirfs-Brock [MSFT] — Sat, 28 Aug 2010 00:23:47 GMT

@Irakli Gozalishvili @ced

Regarding getter/setter definitions in object literals and trailing commas. Those are features of ES5 and are supported in IE9 standards mode. See my previous post about ES5 support in IE9 for details and a link to the actual ES5 specification. See blogs.msdn.com/.../enhanced-scripting-in-ie9-ecmascript-5-support-and-more.aspx

@mitch074

>> IE 9 seems to stick to "what other browsers do"

Actually I was trying to say something quite different. We don’t just do what other browsers do. We prefer to have a standard to guide us. When we don’t we evaluate each issue in a principled manner as explained in the post. Sometimes the result is that we do “what other browsers do” and sometimes is isn’t. Two of the example I used are for issues where we decided not to do what other browsers currently do. I chose those specific issues to talk about because I wanted to explain exactly why IE9 is not providing some JavaScript features that exist in other browsers.

@orcmid

You’re exactly right. The big problem with consensus features is that they are not written down anywhere. Even browser implementers don’t necessarily know what they all are until the go back and reread 10 year old code. This is really a failing of the standards creation process. True consensus features should be incorporated into the appropriate standards. The consensus regexp extensions probably should have been added to ES5. However, as far as I can recall nobody suggested it. Things that are already implemented and work just aren’t thought about a lot. This is something I hope we can do better in the next edition of the ECMAScript specification.

@David Bruant

>> I think that the correct behavior is to throw an error since an interpreter should help detecting obvious semantic errors and this is clearly one

The correct semantics isn’t so obvious. If it was, you wouldn’t see different browsers implementing different semantics for const. Typically there are many possible and plausible semantics for a programming language feature. That’s why standards are important. They tell implementers which of the many possible alternatives is the one that is supposed to be interoperable with other browsers.

Allen Wirfs-Brock

re: Chakra: Interoperability Means More Than Just Standards

Sigh — Fri, 27 Aug 2010 22:42:54 GMT

Troll elsewhere, sloth. You clearly haven't bothered to read anything about how the web browser market actually works (Hint: users uninstall browsers that don't work on the sites they like.) and claims that IE doesn't support "the most basic CSS specifications" indicates that you don't have any clue what you're talking about. IE8 has better support for CSS2.1 (the latest recommendation from the W3C) than any other browser.

Publishing HTML doesn't make you a "professional web developer". Professionals have enough pride and integrity to post their rants under their own name.