Re: [Fsfe-uk] ReplaceFlash, was: BBC digital curriculum service in England

[Alex Hudson]

On Thu, 2004-10-14 at 10:48 +0100, MJ Ray wrote:
> I think the NPD survey is the one at 
> http://www.macromedia.com/software/player_census/flashplayer/ but 
> where are the quoted European results and methodology?

http://www.macromedia.com/software/player_census/flashplayer/version_penetration.html

Methodology and sample is at the bottom.

> That's certainly far higher than I would expect from experiences 
> around here.

I think we're talking about different markets, though. I'm talking about
direct commercial experience developing online language learning
products. 

Interestingly, though I can't talk too much about it, I do have
experience of two products: one based in Flash, the other in
cross-platform HTML and Javascript (although using Real for multimedia
audio and video). Let me tell you; the Flash was quicker to develop and
much more robust in the marketplace. Yes, it does depend on who
developed it, and yes there are other factors at play - but trying to
replicate Flash-style interactive play in other technology is not a
simple game at all. The HTML version (which did work on Netscape 4 up,
pretty much) ended up being more difficult to support, and when browsers
changed (as they inevitably do - the big one being when MS moved from
Netscape plugins to ActiveX) the course had to be bug-fixed. And yes, it
was written to standards. The standards may not change; the browser
support sadly does (witness Moz dropping SVG, I guess).

> >> I'm aware SMIL is a few years behind in terms of installations and 
> >> applications, but they're not show-stoppers for a long-term 
> >> direction. How  isn't it as capable?
> > SMIL isn't even close to the same feature set.
> > [...] It's a completely different kettle of fish to Flash (if it was
> > competitive, Flash player certainly wouldn't support it). [...]
> 
> I'm puzzled by this bracketed comment. Does Flash player support SMIL? 
> If so, isn't its installed base at least as big as that of Flash 
> player?

Well, yes and no. I probably should have suffixed "Flash" with
"authoring tools"; I'm not sure you can actually get the player to deal
with them the same way you can get Real to deal with it. I would guess
it doesn't, the download is pretty small.

Even so, saying "I support SMIL" is still quite a weak statement. Real
supports SMIL, but only really as a framework to sync together Real
media. It's not actually buying you that much, since "SMIL for Real" is
not going to be playable in IE, for example. Plus, SMIL only provides a
small fraction of what Flash provides.

http://www.cwi.nl/projects/Ambulant/Why.html (and the AMBULANT project)
looks interesting for this discussion.

> Three approaches, I guess:
> 1. file format conversion - are there any already?
> 2. develop new authoring tools - the GIMP problem of user inertia for 
> existing users
> 3. persuade "Flash director" to support standard formats - seems 
> unlikely
> 
> As far as I can tell, we're best to encourage some combination of 1 
> and 2.

1 is probably the most straight-forward for now; developing authoring
tools would probably be insanely hard.

As regards the discussion of plugin activation via a browser, yes it is
possible, but you wouldn't want to do it for a number of reasons.
Firstly, browser APIs are different - IE uses ActiveX, Moz uses
XPConnection, etc. etc. I can't remember the practical differences this
entails, but I would suspect it would require at least different code
and I would also imagine feature coverage is not identical (I seem to
remember this used to be a problem; maybe not any more).

Also, it's not really a huge amount of use for timing purposes. If
you're using SMIL, you would want to do all the timing stuff in there,
because that's what it's good at - relying on Javascript calls would get
stuff out of sync very very easily. But, if SMIL is being provided by a
plugin, it would have to be the "top most" content provider (i.e., SVG
support provided by the browser wouldn't be accessible to the SMIL
plugin, and wouldn't be much use). I don't think SMIL operating in a
plugin would be able to time an SVG animation taking place in the
browser, and even if it could, the two drawing areas wouldn't be able to
overlap, so you would have to do all the graphics in the browser which
probably brings you back to faulty timing.

Plus, Flash does have other nice features like loading data in the
background. Given SMIL isn't scriptable, I don't think it can offer that
kind of feature the same way Flash does (it's not the same as
pre-loading things that might be coming up). A SMIL user is more likely
to see "loading" screens than a Flash user, unless the media is highly
linear (you might be right about SMIL - it seems to have some non-linear
support, but that really isn't what it's aimed at and I don't think it
will be very good at it).

If we do want to investigate SMIL more thoroughly, I would guess that
Helix is the place to start. It has SMIL 2.0 somehow, it also has a
range of audio and video codecs. The question is, how do we get SVG (or
a similar graphic) into the Helix player? 

Cheers,

Alex.