It may not have syncing, but Stainless[1] has a very clever notion of bookmarks that <i>are</i> sessions. Makes it ridiculously easy to e.g. have a bookmark for each account you use on site X, and open them all simultaneously. I absolutely loved it. Unfortunately it's a mostly dead project, though it's open source now.<p>[1]: <a href="http://www.stainlessapp.com/" rel="nofollow">http://www.stainlessapp.com/</a>
Isn't this similar to what Firefox does - where the browser UI is written using javascript and XML - which is why it's so easy to write add-ons for it?
While the project objectives are very nice, I hope it solves one other problem, the "contentEditable" problem<p>In the pre-historical days, web pages are static, served as a file blob, in one way, or both ways (FTP)<p>Then there came server side templates, we now have dynamic pages.<p>Then CGI and Perl came, we have <form> so user could put their contributions to the server.<p>Then there's AJAX opened the possibility of "Web apps"<p>However, all of existing efforts left behind a very basic concept of HYPERTEXT: It is always read-only or partially read-only. The browser is good at display HYPERTEXT or manipulating part of it, but not create those.<p>We have WYSIWYG editors, and the famouse Markdown (and alternatives), they lowered the level of entry of a "writable" Web.<p>I hope in one day, users can write hypertext freely, as easy as make an edit on contentEditable, the user input & interactions could become parsable data, and the developer's layout/css tweaking and js debugging could be made from web browser Firebox/DevTool directly back to LESS/Coffeescript.<p>Browsers, really, should consider shifting from a consumerism tool to an authoring platform. Branched, incremental and versioned.<p>My two cents.
Cool project, not what I expected from "Next Generation Web Browser".<p>I was hoping for a project to build an HTML/CSS/JS suite using OpenCL and OpenGL directly, in a functional style which can be parallelized and make use of immutable data structures in a sensible way. Backed by a built-in object store database.<p>I do a lot of OpenCL, so if anyone wants to write <i>this</i> browser, let me know. I will join your mailing list.
Actually a lot of their ideas are similar emacs. Have a few core pieces written in C and then everything else is written in elisp. Even the stackable navigation and WebViews are the equiv of buffers.<p>Conkeror already has stackable navigation and can be extended using javascript. In many ways the ExoBrowser is simply taking Conkeror to it's logical conclusion.<p>Very interesting, I'll be interested in seeing where this goes in the future.
With stacked navigation allowing lots of open tabs, I hope we can also have efficient maintenance of those tabs. I don't need them to keep running javascript in the background. My machine shouldn't sag under the weight of a few hundred open tabs. After they've been inactive for a while, archive their state to disk and replace the view with a screenshot. Pull everything back into ram when I go back to the tab. If I take a second or two to look at the page before I try to interact with it, I won't notice a difference.<p>Let me whitelist a few exceptions, like gmail.<p>Built-in buttons to save/bookmark all open tabs would be nice too. And when saving a page, if there's a page title use that as the filename.<p>Optional feature: insert a meta tag to the html that gives the url of the page, so when I mention it online I can dig up a link to it, without having to bookmark in addition to saving.<p>Also I'd like a complete implementation of Vim and a pony.
Very interesting, but I can't quite rap my head around the idea. Aren't the protocols and the execution engine the plattform on which a browser is built? I find it very interesting to think about it what the minimal executable code is and how we ship it over the wire. Chrome was originally an OS as well as a browser (that's how I understand the history).
Looks like a clone of a project I did in xul a while back. <a href="http://nochrome.tp23.org/" rel="nofollow">http://nochrome.tp23.org/</a>
Got stacked tabs. And comes as expanded hackable js. Good to see movement in the browser space.
Interesting project, but I think it would be useful to write a simple tutorial on how to customize the browser.<p>I've looked at the API but that doesn't really tell how to setup the browser, how to run custom code, etc.
I'd be really interested to see some innovation around bookmarking.<p>My bookmarks are a hopeless unsalvageable mess. I have around 900 bookmarks, but have little hope of finding the one I need when I want to.
How does this compare/contrast (if at all) to the recently announced new browser engine, NiDIUM: <a href="http://www.nidium.com/" rel="nofollow">http://www.nidium.com/</a>
I don't know what experimentation you plan for synchronizing browser state, but Chrome and Firefox already can sync practically everything, securely.
Why not go all the way and move the rendering engine out of the privileged base of the system? Why is all this HTML and Javascript garbage in "kernel land" anyway? Why not make your browser kernel concerned solely with providing asm.js or PNaCL plus WebGL and some other primitive services? Then everything else can be preloaded libraries. And what on Earth does Node.js have to do with any of this? Why not just create some Javascript bindings to WebKit and V8, then.. write the browser in Javascript? The browser needs less of this kind of ridiculous bloat, not more.
up to I looked at the draft for $5082, I did not believe ...that...my friend was like actualey earning money in their spare time on their laptop.. there moms best frend has been doing this for less than eight months and as of now cleared the loans on their villa and purchased a new GMC. look at more info big57.com