(And what about HTML 5?)
These are a few, hasty views of a hobby programmer, on the current animated debate as to whether ECMAScript is finished for the foreseeable future (but for minor tweaks and some security additions), or whether it should take a bold leap forwards (and take a few risks at the same time).
Personally, speaking as a hobbyist who uses HTML 4 and ECMAScript 3 to develop lightweight applications (I have developed a 2-page outliner, time recorder and other applications that run the browser), and as someone who is not quite time-rich enough to make the jump to .NET or Java, I am all for it. But then, I do not have to worry directly about political, commercial, security, technical, interoperability, accessibility, bloat, complexity, implementation, or other considerations.
For me the issue is this:-
(a) should the browser continue further down the path it is already on, and become an open standards internet-oriented "application" platform in its own right, as a complimentary partner for, or even an alternative in some instances to, Silverlight, Java[FX], and Flash[Flex]; or
(b) should the browser stick to its core role as navigator of HTML 4 documents on the World Wide Web, and only be used for a specific set of applications (e.g. e-commerce, blogs, map browsers, mail clients etc) which are adequately catered for already, and otherwise be limited to being a gateway to more advanced applications written for Silverlight, Java[FX], and Flash[Flex]?
Implicit in this, is whether there something about Silverlight and .NET, Java[FX], and Flash[Flex] which means that some people might still prefer to use the browser to deliver some kinds of applications. Does HTML and CSS provide more flexibility for user interface design, and does it provide an immediacy, being delivered through the browser, that other options don't give? We have got used to web pages on demand, so we want our applications on demand. But don't Flash and Silverlight at least already match the browser in this?
There also seems to be a broader issue about encroachment of standards into previously proprietary domains, and whether there is as obvious a case for this as there may have been for the original static World Wide [Information] Web. E.g. the open document format now competing with Microsoft's own document format, and HTML 5 trying to set a standard for "application" mark-up, which overlaps considerably with XUL, XAML, and numerous other schemes for application mark-up languages.
I find this to be a difficult question to answer, because the range of possible applications seems to be unlimited, and there may be more than one way to implement or deliver a given application (from full desktop installation, to static HTML pages delivered by a server side processor and database) each of which could be equally valid; so the role or evolution of the browser cannot easily be said to be closed.
The browser (almost from the start I would guess) already delivers applications through the classic series of server generated web pages, in the form of e-commerce, e-mail clients and other remote application services, and these represent a vast part of the internet. Stuffing the browser with the DOM, Scripting, XML and an HTTPRequest opens the door further, by freeing developers from having to deliver the application through a series of separate processed pages, and enabling an infinite variety of user interfaces, and so brings the browser to the threshold of actually (with some rough edges, and performance limitations) looking like a classic application; staking a claim to some of the territory which .NET, Java, and Flash were already in or going for. Indeed, the Adobe Air product brazenly says that HTML/Script can be an application development platform and run "stand-alone".
If the browser is to go further, then there is a case for giving it a more fully fledged application programming language and HTML 5, and calling this new browser platform the
World Wide Application Web.
Fundamentally, the question is, whilst there may be a compelling case for the a World Wide [Information] Web and an HTML and CSS standard, what is the need for a standards based World Wide Application Web, and why shouldn't the browser stop in its tracks and leave this to competing proprietary plug-in solutions in the form of Silverlight, Java[FX] and Flash[Flex], or even full Java and .NET assemblies? Don't .NET, Java, and Flash already do everything one might need of an [internet] application, and more smoothly than the browser, so why crowd the application development space yet another option? Or perhaps we should even abandon the whole thing and go back to basics; writing yourself a bespoke application as a native executable in Borland's Pascal (as a programmer friend of mine considered doing when unable to choose between the browser, .NET and Java).
Is it the case that just as the World Wide Web freed information, there is a social need for an accessible World Wide Application Web to free applications, or is it too soon to tell?
To close with a glib statement: I would prefer to have more choice and competition than less. Just because the browser and the web started out in life being one thing, does not mean it should not evolve, if the demand is there.
To answer my title: my vote is 4.