WebGL is fancy,but performances are just not there yet. How long most people will stay on that site given how slow it is?and i have a pretty good computer(Macbook pro). it's not a geek website, webdevs should test performances seriously and ask themself if these kind of stunts are worth the effort. So yeah, there is a wow effect, but if i was the client,i'd be afraid of a pretty high bounce rate.
The message is clearly lost here,due to performance problems.
i was sitting at a google dev day once, and they were showing off all the cool new stuff you can do with canvas.
the crowd was cheering. google was the rockstar in the house.
my friend sitting next to me turns to me and says:
"great, now we can do what flash could 10 years ago".
that was the same year when the google munich chrome lead was bashing the nacl presenter in a private conversation for being backwards. he "couldn't comprehend it at all" "the cloud should be doing all the work".
yeah no shit, with all the junk layers we're putting on top of one another there's no way the clients can actually compute things themselves.
Interesting, I've never heard that criticism of NaCl before. I guess it makes some sense if you're engineering toward a "thin client" model with cheap, minimalist hardware.
The main concern about NaCl though, is that it's not portable, it's just Google's version of ActiveX.
So I guess he could have just said "great, now we can do what ActiveX could 10 years ago."
The problem isn't so much with WebGL performance. WebGL is easily fast enough for simple particles like this (although 2D canvas could probably also manage it). The problem is that a lot of web devs don't have a full understanding of the OpenGL pipeline, and how using tools and effects in one way rather than another can lead to crippling performance issues. That entire scene could probably be batched into a single render call.
WebGL has lots of potential to be lightweight and non-intrusive. I've started writing a series on the topic to try and dispel some myths of WebGL being slow and bloated.
Comments
ultra slow on my laptop, unfortunatly.
WebGL is fancy,but performances are just not there yet. How long most people will stay on that site given how slow it is?and i have a pretty good computer(Macbook pro). it's not a geek website, webdevs should test performances seriously and ask themself if these kind of stunts are worth the effort. So yeah, there is a wow effect, but if i was the client,i'd be afraid of a pretty high bounce rate.
The message is clearly lost here,due to performance problems.
same here, haswell laptop.
i was sitting at a google dev day once, and they were showing off all the cool new stuff you can do with canvas.
the crowd was cheering. google was the rockstar in the house.
my friend sitting next to me turns to me and says:
"great, now we can do what flash could 10 years ago".
that was the same year when the google munich chrome lead was bashing the nacl presenter in a private conversation for being backwards. he "couldn't comprehend it at all" "the cloud should be doing all the work".
yeah no shit, with all the junk layers we're putting on top of one another there's no way the clients can actually compute things themselves.
Interesting, I've never heard that criticism of NaCl before. I guess it makes some sense if you're engineering toward a "thin client" model with cheap, minimalist hardware.
The main concern about NaCl though, is that it's not portable, it's just Google's version of ActiveX.
So I guess he could have just said "great, now we can do what ActiveX could 10 years ago."
there are noticeably less dependencies to the underlying OS and a stricter security model. pnacl even goes further with being architecture neutral.
sometimes the inner qualities of a system are those that matter.
The problem isn't so much with WebGL performance. WebGL is easily fast enough for simple particles like this (although 2D canvas could probably also manage it). The problem is that a lot of web devs don't have a full understanding of the OpenGL pipeline, and how using tools and effects in one way rather than another can lead to crippling performance issues. That entire scene could probably be batched into a single render call.
WebGL has lots of potential to be lightweight and non-intrusive. I've started writing a series on the topic to try and dispel some myths of WebGL being slow and bloated.
https://medium.com/p/abf669857722
Part 1 of the technical aspects: https://medium.com/p/cebd176c281d
I like to play around with WebGL for prototyping ideas, but it be the second coming of VRML if it doesn't improve.
Every time I see a WebGL site, my NVIDIA card jumps to full usage, followed by increase of CPU usage across all cores.
For real OpenGL stuff, native is the way to go.
Unfortunately, it's so laggy as to be unnavigable. :/ Nice idea but as you say the message is lost.