- From: Alan Stearns <stearns@adobe.com>
- Date: Mon, 28 Oct 2013 13:21:02 -0700
- To: "www-style@w3.org" <www-style@w3.org>
- CC: Sylvain Galineau <galineau@adobe.com>, "L. David Baron" <dbaron@dbaron.org>
On 10/28/13 7:40 AM, "Sylvain Galineau" <galineau@adobe.com> wrote: >So if I wanted to build a web version of the widely used >Tweetdeck app, one named flow per feed would be a great way to separate >the content - individual feeds of tweets - from the visual feed columns in >the app. Sylvain brought this up in the context of wanting to attach event handlers to regions (to reorder feeds, for instance). But I think it's a very interesting case for flow-into generally. It's a more complex and compelling version of the 'breaking news' use case where some content gets redirected by a named flow to a more prominent container. In the Tweetdeck case, you'd have your content which is a chronologically-ordered list of tweets. Then you have N containers for each tweet category/search you want to display. These containers might be elements, or in the near future they might be custom components. The issue is how you match the content to the container you want to display it in. Without named flows, you either have a server process that does some violence to the content to wrap it in the various containers, or you have some client-side script that uses DOM manipulation to scramble the ordered list into the intended containers. If the user changes one of the feed searches, then you have to run the process/script again to re-mangle your content into the new display. With named flows, you can leave the content as it is. Tweet elements get assigned to different named flows, and those named flows each get assigned to a single container. When the user changes a feed search, all you have to do is change some CSS and/or class assignments in order to get the new display. The content always stays intact in its original structure. I think that named flows provide a better content/presentation separation story for this case than currently available methods. Thanks, Alan
Received on Monday, 28 October 2013 20:21:34 UTC