Bulk loading/saving big binary data from/to IndexedDB is basically as fast as accessing the file system. Using TypedArray or (upcoming) BinaryData, JavaScript can also manipulate binary data pretty quickly.
I haven't attempted doing any serious graph processing in JavaScript, but these days, JavaScript is extremely fast at tree manipulation (much faster than OCaml, which has long been the fastest high-level language for this kind of thing), so I would imagine that it's quite suited for the task.
If you are serious about this, you should come and chat on irc.mozilla.org, channel #openwebapps for instance. We'll be happy to help you get started, plus that kind of discussion is often how new features and optimizations get started.
Bulk loading/saving big binary data from/to IndexedDB is basically as fast as accessing the file system. Using TypedArray or (upcoming) BinaryData, JavaScript can also manipulate binary data pretty quickly.
I haven't attempted doing any serious graph processing in JavaScript, but these days, JavaScript is extremely fast at tree manipulation (much faster than OCaml, which has long been the fastest high-level language for this kind of thing), so I would imagine that it's quite suited for the task.
If you are serious about this, you should come and chat on irc.mozilla.org, channel #openwebapps for instance. We'll be happy to help you get started, plus that kind of discussion is often how new features and optimizations get started.