WHATWG Weekly: Web Notifications
Much sleep has not been had, so if you read something silly, it might just be me. There’s good news too, Shelley Powers volunteered to start writing the WHATWG Weekly starting March 21. Her main interest is HTML, so you might need to pester her on twitter (@shelleypowers) or email ([email protected]) to get other things covered.
Robert O'Callahan once wrote a great post on The Essence Of Web Applications. Nonetheless, there are some features desktop applications have that would be quite useful in the web application space. The Web Notifications work is one part of that puzzle, bringing an API to the web to show simple application status messages to the user. This way e.g. Yahoo! Mail can notify you of incoming email even when your browser is running in the background or Yahoo! Mail is not the active tab.
Purging Link Relationships
Per a decision of the W3C HTML WG the
first, and their synonym link relationships (values for the
rel attribute) have been dropped. Since
archives was similar to
index that has been removed too for consistency. These relationships were never that useful to begin with so I suppose it is nice that authors no longer have to worry about them. I.e. either worry whether to add them at all or whether they would be appropriate in a certain situation.
- Philip Jägenstedt made several small enhancements to the HTML5 Tracker, including better icons for the browsers. Lets us know what you think!
- Tab Atkins would like to reuse
<canvas>for images with structured fallback.
- Dave Kok put forward a detailed proposal for Session Management.
- David Bruant brought a discussion on the interaction of multiple global objects when using ECMAScript to the attention of the WHATWG, which happens to be standardizing some of that.
- Christoph Päper asked whether we should have a new native control for (value, unit) tuples.
- Edward Gerhold joined the quest (pretty sure there is one) to improve the application cache feature.
- Lawrence Rosen presented a license proposal to the HTML WG. Unlike the WHATWG the W3C has a restrictive license for its specifications, but it is considering changing that.
- Since the work on a new cookies specification is practically finished (waiting for the IETF to assign it a RFC number) Adam Barth put forward Origin Cookies. A proposal to solve some of the problems with cookies as they are today.
- Henri Sivonen found joy where most would find despair.
Give me my index, up, last and first links back. It’s not that they cost you anything.
I haven’t had time to read [email protected] regularly in awhile, so I wanted to say thanks for doing this — I’d have missed the multiple-globals discussion migration (metastasization?) to it without hearing of it here. (I’m initially skeptical of his proposal that WebIDL is the right place to specify multiple globals work in the language, but since I’ve done no more than read the initial email you link I certainly haven’t given the idea fair consideration yet.)
This is sad.
Before HTML5, there was a pretty good consensus and a number of largely deployed implementations of these useful link relationships (think e.g. accessibility), with long history since at least HTML 3.2:
The notions have been apparently mixed-up by the editor(s), leading to a mess (for instance erroneously attempting to put “index” and “contents” as synonyms of “top”), and throwing the baby with the bath was maybe the simple way out (and maybe the intended result, as I suspect after having seen the animosity of some people on the matter).
I think it is a pity to reduce this heritage, and editors should remember the importance of relationships between documents in the original HTTP/HTML vision.
Personally, I use these features constantly, and even on this very blog (blog.whatwg.org), I make use of them (thanks to their nice implementation in my Opera browser).
I hope this feature will be able to make it through somehow, maybe during later reviews…
After having received some good explanations from Karl, the situation for link-rel may not be that bad after all (from the viewpoint that I defended in my previous comment), if it turns out as follows: External repositories with larger collections of values (including the removed ones) may instead be relied upon in the future, such as http://www.iana.org/assignments/link-relations/link-relations.xml
Here are the minutes of the decisions: