Mozilla lays out multi-process Firefox engineering goals
Mozilla's Chris Blizzard has detailed the non-profit organisation's plans for the engineering behind Firefox, as it looks to make the browser spread even more of its workload between multiple processes. Blizzard notes that, although the multiple process model for Firefox is not a panacea, "it does gives us a leg up on some of the more systemic problems".
Going multi-process for Firefox goes beyond the previous efforts which have seen some plug-ins and extensions relocated into their own process space. The eventual plan for the Electrolysis project is that the Firefox user interface and each web page will be handled in their own process. These "content processes" should lead to a more responsive browser where garbage collection is inexpensive and won't block the user interface from responding. It should also prevent image decoding and other expensive operations on the DOM from blocking responses. Multiple process Firefox will also make better use of multi-core hardware, allow for security oriented sandboxing, improve crash resilience and give better, more predictable, memory use.
If the advantages of a multi-process model sound familiar, it is because Google's Chrome and Apple's WebKit2 have been developed with the model in mind. The task is more complex for the Mozilla developers who have been working on "out-of-process" technology on the relatively green field of Mobile Firefox.
Meanwhile, Mozilla's Aza Dotzler, Product manager for the Firefox on the desktop, is trying to make 64-bit builds of Firefox for Win64 a "supportable reality". Beyond improving security and performance, Dotzler sees "download confusion", "plug-in issues" and "memory footprint" as solvable issues. Dotzler is looking to the community to help with the planning of work that needs to be done to make a 64-bit Windows Firefox easy to install and use.