Just say NOOOOOOOOOOOOOOOOOO!

Why? What you’re asking for doesn’t actually make much sense, when one thinks carefully about the differences between applications and documents. And also keep in mind that this is not just a ‘Flash thing’, it is true for any other type of Web-hosted application (AJAX, Silverlight, JavaFX, etc.)

Why not?

1) Browser history support is a combination of: UNDO mechanism and NAVIGATION. Let’s discuss UNDO first. There are plenty of examples of traditional (i.e., non-Web, such as: client-server, mainframe, native Windows …) applications that do not implement an UNDO mechanism. Sure, UNDO is a superb idea–but it is not a requirement. So, how about the navigational aspect? What it effectively gives you is a linear (non-branching) context stack (with backwards and forwards traversal). You could find that same UI concept in 1970s vintage mainframe applications, so this is NOT an innovation in interaction design. More to the point, if an application developer wants to support it, it can be easily enough added to that application. (But I think that GUI widgets like tabs and breadcrumb trails are a better idea, in general.) The one thing that the browser history support gives you is a standardized way of traversing the context stack. That is, indeed, worth something; in particular it improves usability for new or infrequent users. But not worth enough to make it a requirement, IMO.

2)Bookmarking is a conflated term. If you’re talking about bookmarking the splash screen/login form/’home page’ of an application, then you can do that today and it works fine. So, the only thing I see as being in dispute is deep linking. What does deep-linking mean, semantically and technically? It means being able to tag and persist the current state of the application and then to restore that state on demand. Hey, that sounds really cool, let’s demand that . . . Just one problem: its wicked hard to do (at least for non-trivial applications). The only general examples I’ve seen of technology that does that are Operating Systems (i.e., Hibernate, or Suspend/Resume), and general purpose Virtual Machine technology (VM snapshotting and machine migration). I’m actually moderately optimistic that, with some progress in VM technology and maybe faster hardware, this type of technology may become more feasible _at the level of browser platforms and even applications_. But we’re not there yet (IMO), so I don’t think it is realistic (or fair) to have deep-linking as a requirement.

Thanks for reading my ramblings,
Banned in Boston

Look at the Java ecosystem as an example (even though it took long to get Java completely open). Struts, Spring, Eclipse, OSGi, Ant, etc, etc. These are the kinds of things that grow when you are embraced by the open community.

IMO – overnight Adobe’s market would increase an order of magnitude. They’ll be able to compete better with Microsoft. Because on a purely technical level – Silverlight is much more exciting than Flash.

Moonlight is already open source and had a 64 bits port available from the very beginning. Adobe only now announced the availability for 64 bits linux.

Interestingly enough, parts of the Silverlight 2.0 such as the CLR not only have good open source implementations but are under a standard (ECMA 335 FYI).

[1] http://www.go-mono.com/moonlight/

http://codinginparadise.org/weblog/2008/11/how-flash-can-integrate-with-open-web.html