in the meantime, i suppose adobe/flash have enough room to improve and provide something that html5 might not be able to cope when it’s finally ready and ‘standardized’ on all browsers. i think all talks are valid, but simply saying that js/css can replace flash seems to me a lot of naivety or uber belief in browsers’ development.

In general i dont think flash will be replaced by HTML canvas in near future. Something like flex framework will help to grow the adoption of html5, for now nothing like that emerges.

“Features like secure streaming, dynamic buffering, and rich video controls will keep professional from switching technologies.”

Actionscript and Javascript are alike in more ways than one, rich video controls on Flash aren’t any better than the same recreated in Javascript (and arguably Javascript is easier to pick up than Flash despite sharing a similar syntax).

And things like secure streaming and dynamic buffering have already been figured out long before the HTML5 video spec was proposed (for plugins like Quicktime).

“…the notion that HTML5 can kill off Flash, Silverlight, Unity3d, O3D, or any other rich multimedia plugin ignores the facts that these tools were created for large scale Application, game development, and even 3D multimedia development.”

To mention, O3D is planned to be integrated into web browsers alongside HTML5, it shouldn’t be clustered with Flash, Silverlight or Unity3d plugins that fork web development or make it proprietary, especially being that O3D relies on Javascript for coding.

“You could argue that someone could write a 3D engine that uses the Canvas, but it wouldn’t have the speed as a plugin that has access to low level processing and video card integration.”

I’ll point you here then: https://wiki.mozilla.org/Canvas:3D

Oh also that’s why O3D is being integrated into browsers as a standard.

“What I am trying to say is, use the right tools for the job. Silverlight is a multimedia tool meant for .NET developers. Flash is a multimedia tool meant for non-programmers and Java-style developers. HTML is a great tool for web page markup and some presentation content. Trying to bend a tool beyond their intended use is only abusing the platform. Picking the right tools for the job is the most important project decision you can make.”

Ok, I can understand what you’re saying here (side note, Flash is for those familiar with Javascript not Java) and yes HTML is for markup and presentation but that’s why we have other standards like Javascript and CSS and plus aren’t all visuals apart of presentation.

Alongside this we have another working standard Native Client (http://code.google.com/p/nativeclient/) which will allow development of web objects in native C code, allowing developers to create wrappers for things like .NET, Java, Python, Actionscript (via Gnash), and so on. Killing the need for proprietary again.

Plus the web was designed for sharing information and communication openly, this shouldn’t be expressed by proprietary plugins and protocols.

from reality, I agree the author.

from the ideal, I disagree.

Let’s take them to the ideal:
HTML is browser based in a standard way;
RIA is a virtual layer beyond native platform in a non-standard way;
native program is platform-dependent.

so, from cross-platform criteria, HTML is standard and is/will be used most widely. RIA give the same ability, but with non-standard(and less) support.

from functionality criteria, HTML have less function than RIA, and RIA is less than native program.

from the Trend, HTML will catch up some funtion, and RIA do this from native program.

during the Trend process,HTML will always lack function than RIA, and RIA less than native program.

to the end of the process, HTML will be the base of WebOS,RIA is no reason to exist any more, and so do the native program – expect that HTML program with native module. — total ideally.

why?
that implies:
programs should be written under a total standard way, run totally platform-independent. HTML is the standard so far, RIA and native program is NOT.
native functionality should be embeded as plugin only if necessary.

How?
the answer is un-predicatable.
competition, collaboration anywhere, invention anytime,
any decision matters much.

what’s my point?
choose by the reality, think for the future. that’s all.

Most of the actual implementations, so far, use the underlying operating systems media plugin, some of which (like Quicktime) have similar capabilities to Flash – and with Flash video also moving towards a MPEG/H.264 based format, and opening up the protocols, I think we’re rapidly approaching a point where non-DRM video can be regarded as standardised – thanks to Flash for creating that standardisation.

(It could have been Real or Quicktime, both of which are not content to be plug-ins, but insist on taking over other parts of your computer, and conducting their war against Microsoft).

Equally, some implementations of HTML5/CSS3/etc do use GPU acceleration, and I believe there is talk of creating

As for GPU acceleration, etc – Canvas 3D is a (prototype) binding of OpenGL with JavaScript, and some of the browsers also use GPU acceleration to improve performance (and notably this is something that has only recently come to Flash).

While I think the Flash platform will continue to advance, it’s also a case that as basic browser technology catches up, those advances will be into increasingly niche areas.

This is exactly the same as the debate on web apps/RIA vs Desktop apps – an SDK like Cocoa on OS X is still way ahead of Flex, but for the vast majority of tasks web apps are capable enough.