« Apollo ignorance | Main | 10 Apollo faves »
May 07, 2007
Silverwatch
Silverwatch: I pay attention to anything called "flash killer", particularly when it comes from Microsoft. I don't have much to say myself yet -- they still haven't reached feature-complete, and their marketing story is still shifting today -- I'll avoid giving them help by discussing specific issues at this point. But last week I read all the public commentary I could find, and a few hidden themes emerged: the people who had the most positive reaction were from the .NET community; the shifting deliverables do not indicate a coherent internal process; Microsoft video may have been refactored to be more Flash-like, but so was the Microsoft logic engine, the "CLR". More after the jump...
The people who were the happiest about Silverlight were .NET developers. There were a few "competition for Adobe is good" and a few "I'd like to write Ruby for the desktop", but the preponderance of positive comments were from all-Microsoft shops, who now see a way to reproduce Flash-like goodness from the past five years of the Web.
This constituency carries some implications for what the technology will achieve in the future. I've hated most of the "only-in-IE" interfaces I've seen in the past. Giving this crowd video and motion won't make such interfaces much better. The people who are reluctant to use technology outside of the Microsoft stack are very different, psychologically, from those who have been doing good Flash and Ajax work the past few years. I see Silverlight as commoditizing recent design principles, rather than exploring the new. Different types of people are attracted to different approaches to technology and inclusiveness.
(One warning sign: The level of knowledge among the MS fanboy posts was very, very low... we'll likely see dumb branding-based arguments online over the next 6-12 months.)
Odd thing I expected to hear more about from the blogosphere last week was the erratic delivery... last March the 1.0 plugin was expected for the end of 2006, complete with logic engine ("CLR"). Last December we saw a pre-alpha preview version, and the internal logic engine was ditched.
But the big news at their April 2007 event was the shipping of the first beta plugin. That's later than we would have though 14 months ago, back at MIX06. More confusingly, internal logic abilities had returned, but as a 1.1 alpha. Even just the parallel alpha/beta release should have drawn a collective "huh? wha?" from the blogosphere... two distinct development tracks is a very unusual pattern, hinting at a confused and shifting internal process. That's what we on the outside can see.
(I would have expected a single beta plugin with internal logic. The fact that there are two distinct downloads indicates some backstory that we on the outside have not yet heard.)
For video, the move to a Flash-style browser plugin is a clear admission that the Windows Media Player style of "we'll show your content in our own branded window" approach failed. Microsoft has reacted faster than Real or Apple here. (Well, it hasn't "failed" so much as "failed as the sole web-video experience"... the Adobe Media Player will be taking a next-gen approach to dedicated video-viewing applications.)
But there was also a significant retrenchment on the "Common Language Runtime" as well. Five years ago Microsoft press releases touted this client software, but it never took off, with only a few consumers installing the large logic runtime engine on their machine. The Linux emulation in Mono was always in uncertain status too. Meanwhile Adobe Flash Player not only introduced Just-in-Time native-code compilation, efficient memory management, and ECMAScript 4 innovations, but got 80%+ consumer installation in well under a year. Microsoft's strategy here is still not clear (only a few picked up on Tim Anderson's mention of incremental installation, or Ed Burnette's "camel nose" analogy), but it's obvious that CLR, like WMP, failed in its original goals, and was refactored to be more Flash-like.
My strongest takeaway, though, remains the one I had after the March 2006 Microsoft conference -- interaction designers have won; the aesthetic that so many developed over the past ten years has taken over the face of modern computing. When Microsoft tries to bring their only-in-IE troops onboard, then something big has happened in the world of interaction design.
Posted by JohnDowdell at May 7, 2007 05:51 AM
Trackback Pings
TrackBack URL for this entry:
http://weblogs.macromedia.com/mtadmin/mt-tb.cgi/8694
Comments
The only thing that has caught my attention in terms of silverlight thus far is that they plan to go after mobile as well (the details are sketchy but some mobile silverlight action has been spotted by several bloggers). I enjoy flash lite very much, however, I find it unfortunate and time-consuming to have to worry about developing for flash lite as well as regular flash. The nice thing about flash for pocket pc is that you can develop apps for it in exactly the same manner and interface as you do with regular flash (so long as you keep all the features at the flash player 7 level). That to me has been a great experience, but dealing with flash lite is a bit of a pain. Adobe should focus on truly creating a write-once, run everywhere environment for the flash platform. Granted that not all devices have the same resolution, or same modalities for interface, or same graphical capabilities, etc. But there has to be a better way to resolve this than just making a new fork in the road and calling it "flash lite". If microsoft silverlight developers are able to write code once and run it on both desktop and mobile environments, then our beloved flash platform may face a critical challenge. I doubt that microsoft has solved this issue already but if I were them, that's what I'd go after.
[jd sez: Reconciling laptop and pocket capabilities has been announced by the group here as the objective, but I haven't seen public estimates yet of delivery schedules. What you describe is the dream we're already pursuing. :) ]
Posted by: Pedro Carabeo-Nieva at May 7, 2007 08:23 AM
Hmm, one other element very much in play the past week... Grant Skinner mentions it in his analysis:
"In the short-term Silverlight offers a lot of opportunity. Microsoft has *really* deep pockets, and in typical MS fashion they are throwing a lot of money around to convince developers that Silverlight is worth learning. I know a number of Flash developers that are currently making decent money working on Silverlight projects that are directly or indirectly funded by Redmond. Adobe simply cannot compete in this respect. They have capital, but not the kind of "buy a small country" cash-on-hand that Microsoft has."
Posted by: John Dowdell at May 7, 2007 08:39 AM
You know me JD, never one to compliment MSFT on anything. I also can't help but think Silverlight is completely overhyped, BUT... I downloaded the beta plug-in to my 12 inch Powerbook a few days ago and they've already done one thing you guys haven't been able to do in 10 years of Flash: make it play smoothly on a Mac.
The embedded video was smooth as TV, as opposed to the jerky playback Flash has forever been guilty of.
I didn't play with the thing much after that because frankly, the "sample" content they provided was all pretty blah, and I'm still not willing to say the technology is a winner, but smooth playback on a Mac shows two important things to me:
a) They are serious about being cross-platform.
b) Apple now seems to be right in the whole blame game about which company was responsible for Flash's slowness on Macs.
[jd sez: Doesn't sound like a response to what I wrote; more like a new topic. If you saw a single video it would be good to analyze that case before generalizing to the whole: in-browser timing is a large subject.]
Posted by: Mike D. at May 7, 2007 09:00 AM
JD- i'd agree with you on .net houses. In the post "10 reasons why silverlight will succeed" the guy says that MS's main focus is on already entrenched developers who are looking for more tools from MS to extend their capabilities. MS from this angle seems to be getting the traction where they want it - With their current favorite dev houses. And i think that also explains the shifting focus over the dev cycle. I'm not sure microsoft has a vision beyond whatever the sum of the dev house requests were/are. Feels more like a .net update mentality rather than defining a new eco-system for app dev. Serious lack of vision for what the goals are. The apollo team was "on message" from the beginning with what apollo would be and what it would solve at version 1.
And i don't expect to see much better apps from them since i run safari-but i've survived without their apps so i am not worried
Posted by: Ethan Estes at May 7, 2007 01:20 PM
As a Flash developer I'm way more interested in WPF than Silverlight. It seems MS and Adobe are going towards the same destination but MS is coming from desktop tools and adding the Silverlight player as an extra... where Adobe is coming from the web player and adding desktop features with Apollo.
The contention that they're messed up because they're taking forever isn't the way I see it. I see it more like they're in it for the long term. Look at Central and how much that improved Apollo. The point is that some things take time. The things that are quick to develop are usually just as quick to go away.
Posted by: Phillip Kerman at May 7, 2007 02:32 PM
Video playback using SilverLight in FireFox under my Vista machine is very very smooth, uses very little CPU and memory resources, and at first glance, looks very 'high definition' and never skips or pops.
I wanted to try the demos to see what all the 'fuss' is about... and it sure seems to work great so far. I'm glad there will be some competition to Flash - it's always good to have.
From what I've read on other blogs, etc., SilverLight will use more available and common codeds, and the video or streaming licenses will be cheaper?
Anyways, I think MS may have a hit on there hands. Silverlight is sure generating a lot of press.
Posted by: Drake at May 7, 2007 08:54 PM
ok let's clarify things here
SilverLight 1.0 (WPF:E beta)
support only unmanaged code
SilverLight 1.1
support unmanaged and managed code
on the managed code supported you can use
languages based on the CLR (C# and VB.net for now)
or language based on the DLR
(Dynamic Language Runtime)
and there you can use IronPython and Managed JScript.
(note: Managed JScript has nothing to do with JScript.NET, it's the new version of the MS JScript engine).
more infos:
http://blogs.msdn.com/deepak/archive/2007/05/02/managed-jscript-is-availaible.aspx
http://tirania.org/blog/archive/2007/May-03-1.html
so basically, the intesting bit in SilverLight 1.1 is that you can either use C#, VB.net, IronPython, Managed JScript to program over it,
and the 1.1 plugin itself now total 4MB (it was less than that when only unmanaged code was supported but still 4MB to run a "mini" .NET VM is quite good)
now the question would be why ?
bah simply because if you don't have a faster engine, their solution does not really bring anything new, or let's say that it can not really compete with the speed of the Flash 9 VM :p
just my 0.2€ about the CLR bit of confusing infos in your post
Posted by: zwetan at May 8, 2007 07:46 AM