Ajax Runtime Issue List Working Draft
From RuntimeWiki
OpenAjax Runtime Task Force Wish List
In the order of priority (more or less based on the number of references from the input documents):
- "The two HTTP connection limit" issue.
- References:
- Need for mediated cross frame messaging support.
- References:
- Discussion:
- JON: Although my wiki page used the topic name "Mediated Cross-Frame Messaging", I don't think we need the browsers to do any mediating per se, just provide a publish/subscribe mechanism for cross-frame messaging from a child IFRAME to/from its parent so that mashups can work in a performant manner. Right now, mashups have to resort to slow communications hacks such as the IFRAME proxy approach which requires chunking the data and polling for changes. We want to have a replacement approach. We don't need a W3C standard for this, we just need particular browsers to implement something that can be used by Hub 1.1 under the hood. Note that Google Gears probably can address this need, so most likely if Gears is bundled, we don't need anything else.
- Offline support (maybe via Google Gears).
- References:
- Discussion:
- JON: Ubiquitous deployment of Gears would be great, but if browser vendors are uncomfortable with Gears, then they should at least implement something that has the same basic functionality such that someone can write an Ajax wrapper library. (Note: Dojo Offline Toolkit is an example of one such library.)
- Event opacity.
- References:
- [Alex Russell]
- Echoed widely at Ajaxian, Alex' blog comments, etc.
- References:
- Mutation Events.
- References:
- Consistent Box Model(eg. APIs that gives reliable location of any element)
- References:
- Cross domain communication/HttpOnly cooki/JS
- References:
- Discussion:
- JON: Lately (and I am writing this 2008-01-09), there has been hot discussion at the W3C about their Access Control spec and some corresponding discussion on the OpenAjax Security TF mailing list. Lots of pushback on W3C Access Control from the community and some folks (such as me) saying that JSONRequest is a better option. This is a complex issue with many factors to consider and many differing opinions about all of the various factors. But the key thing to note is that W3C is in very active discussion about how to enable cross-frame communications.
- OnDomReady/OnLayoutComplete/OnLoad - event notification
- Reference:
- Discussion:
- JON: I'm not a big fan of this one. Yes, I see the need, but the problem is that this is the sort of feature that needs a standards organization to define it first before the browsers can implement. Therefore, we probably should send this request to W3C/HTML5 or WhatWG first (e.g., http://www.whatwg.org/specs/web-apps/current-work/#event2), rather than send it to the browser vendors.
- XPath support.
- References:
- 2D Drawing/Pixel level presention/vector graphics.
- References:
- Discussion:
- JON: Microsoft, I'm looking at you: implement SVG like the other browsers have done! It has been a W3C standard since 2001. Don't say it is too hard. In the meantime, you did Vista and the whole Silverlight thing.
- JIT for Javascript.
- References:
- [Ajaxian]
- Discussion:
- JON: After talking with various leading people in the Ajax community, what I heard was that JavaScript isn't the performance problem, it's "DOM" that needs to be sped up. Also, JIT is just one approach to speeding up the performance of raw JavaScript. Therefore, we might want to reword this one to be "Make scripting faster", and suggest that they speed up both their JavaScript engines (perhaps via JIT technology) and also speed up common DOM operations.
- References:
- Reliable caching of Javascript toolkit.
- References:
