Wednesday, March 09, 2011

At Work with Linux: Chrome 10 on RHEL 6

The latest Chrome updates came rattling down the Internets and landed on my RHEL 6 systems today. And, as usual, I decided to throw a test sample or two its way to see how Chrome 10 handled them. For this I went to Google's very own Chrome Experiments page.

I wasn't trying to be particularly sneaky or mean spirited. I picked the "Sintel Goes Boom" WebGL experiment because it was the first on the page (well, top upper left corner for those cultures that read left-to-right, top-to-bottom). And sure enough, Sintel did indeed go boom. Along with the sandboxed page. I even got a nice crash warning on the Gnome desktop.

Chrome 10 still has issues running WebGL on RHEL 6

Which, in the Universe's Grand Scheme, is not that big a deal. If nothing else I verified, once again, that crashing in Chrome's sandbox does not crash Chrome. For those of you curious as to the details of what happened, here's what RHEL 6 reported.

And the WebGL-generated crash report

Same cute VW ad, using HTML 5
I did manage to stream video again, after going back to YouTube's HTML5 opt-in page and opting in again. I did this for Chrome 9, but it apparently forgot that opt-in when automatically stepping up to Chrome 10. No biggie, re-visit the page, click the link, and we're good to go again. Nothing out of the ordinary, at least on RHEL6. Still good performance and decent audio.

Tankworld works. Unfortunately, I don't.
I did go back and try out some more Chrome experiments, finally hitting one that didn't crash and burn: Tankworld. I played with it a bit and was sufficiently impressed with its playability and reasonable performance.

Although I'm something of an grumpy old dog when it comes to technology these days, I have to admire and appreciate what's been happening across the board with Chrome, Firefox, and Internet Explorer, especially with regards to HTML 5. For the first time I can actually see someone writing non-trivial web-delivered applications that run across multiple platforms on the same browser (Chrome, in this case) without the need for separate specialized native-platform plugins.

What I'm most afraid of is that I'm going to run into the same issue with HTML 5 that I face with Java - write once, debug everywhere. But then again, maybe not. This is 15 years after Java was released. Perhaps we've collectively suffered enough pain with the Java write-once-run-everywhere torture that this time around we'll honestly follow through on that promise, especially with regards to graphics and multi-media. I can only hope.

No comments:

Post a Comment

All comments are checked. Comment SPAM will be blocked and deleted.

Note: Only a member of this blog may post a comment.