Reading Marc Andreseen’s recent comments about browser innovation go me thinking about alternatives for web navigation… specifically, what the weaknesses currently are, and how to improve them.
- threading/path navigation – currently, browser histories don’t keep track of the path of you’re browsing: which links you’re following, in what windows (or tabs!), etc. If they did, not only could you very easily pick up where you left off (also for crash recovery), but definitely help you find things, especially when combined with
- historical search – yeah, everyone’s talked about this before, but it still doesn’t seem to exist. In the world of cheap 200GB hard drives, why can’t I keep a searchable cache of all the pages I’ve browsed? Well, it’s not for any technical reason
- statistical bookmarks – would be neat if various filters could be applied to your history corpus to autogenerate bookmarks out of visiting patterns
- intrapage bookmarking/annotations – it’d be nice if you’re location in the page would also be stored. This could be done in a combination of ways – character offsets, xpath, rolling checksums / character diffs. In fact, it’d probably have to use all of those if you’re storing against changing pages (of course, local cache would help alleviate some of this).
- browser lookahead – prefetching simple HEAD requests right now to show link status overlays w/ size/type/status info would be a nice addition to the current prefetching implementation
- tabbed browsing manager – coming soon?
Non-nav improvements:
- syncing – bidirectional profile auto-syncing
- better blocking – for images, other media types that let you block by URI fragment, or properties, or any combination thereof
- site-by-site plug-in control
- fine grained transfer control – the ability to do absolute/percentage based throttling by host, window, etc.