norman.walsh.name: Comments on /2008/07/13/fluidhttp://norman.walsh.name/2008/07/13/fluid/comments.atom2012-05-23T13:12:04.353967ZComment 1 on /2008/07/13/fluidhttp://norman.walsh.name/2010/09/25/oauth#comment00012008-07-14T06:55:43Z2008-07-14T06:55:43ZDave Pawsonda39a3ee5e6b4b0d3255bfef95601890afd80709

.... And it's Mac only.

Comment 2 on /2008/07/13/fluidhttp://norman.walsh.name/2010/09/25/oauth#comment00022008-07-14T11:50:40Z2008-07-14T11:50:40ZNorman Walshda39a3ee5e6b4b0d3255bfef95601890afd80709

True. But the "site-specific browser" page on Wikipedia does have links to other things. Mark Birkbeck's "Sidewinder" demos at XML 2007 were particularly impressive.

Comment 3 on /2008/07/13/fluidhttp://norman.walsh.name/2010/09/25/oauth#comment00032008-07-14T12:59:55Z2008-07-14T12:59:55ZChris Messinada39a3ee5e6b4b0d3255bfef95601890afd80709

I'm a huge fan of Todd's work on Fluid as well. There are so many possibilities -- it boggles the mind!

For the PC, you might take a look at Mozilla's cross-platform Prism platform -- a project that has influenced Fluid's development:

http://labs.mozilla.com/projects/prism/

Comment 4 on /2008/07/13/fluidhttp://norman.walsh.name/2010/09/25/oauth#comment00042008-07-14T20:56:30Z2008-07-14T20:56:30ZLou Springerda39a3ee5e6b4b0d3255bfef95601890afd80709

Depending on how it is implemented, it may address an issue that vexes me with my overloaded browser: ill-mannered Javascript crapping up the entire browser process. (Can you spell SBOD?)

Process isolation is something that I *thought* was solved a long time ago, but seems to have been forgotten in the headlong rush to put gobs of logic and in the single running browser processes.

I'm betting we will eventually have some hybrid of browser+local web service to provide a unified architecture to address this as well as some other issues. I think running a browser instance for a single page as a first-class process is a step in the right direction. I'll trade some memory and context switches for some control and isolation, thanks.