Yeah, this kind of detail was really missing from the public statements :)  I 
don't expect consumer-facing PR posts to go into nitty-gritty technical 
details, but it wasn't apparent that there was really anything more going on 
besides "nope, we're just going to rewrite it and move all the UI around in the 
process".  Appreciate the info.

On Tuesday, September 24, 2019 at 1:41:44 PM UTC-4, Kris Maglione wrote:
> On Tue, Sep 24, 2019 at 10:07:28AM -0700, mark.erik...@gmail.com wrote:
> >- I am very happy with the current Fennec app and its UI, and 
> >don't understand why Mozilla feels a need to drop that product 
> >and create a new one from scratch.
> 
> We're not creating a new one from scratch. Many of the component 
> parts of Fenix already existed, and the core is basically the 
> same as Fennec. That said, there are many reasons:
> 
> 1) GeckoView, which is a more or less drop-in replacement for 
> Android's WebView, is designed to be used by any application 
> which needs to embed web content on Android, including 
> potentially other web browsers. It's an important part of our 
> core mission to make sure the web remains a viable living 
> standard, with multiple competing implementations, to avoid the 
> sort of stagnation and vendor lock-in we saw when IE6 
> essentially ruled the world.
> 
> 2) We also already needed to maintain WebView-based browser 
> front-ends for configurations where shipping our entire 
> rendering back-end was not viable or practical. Given that we 
> already need to maintain these separate, already-compatible 
> front-end and back-end implementations, having to maintain an 
> entirely separate browser with a completely different front-end, 
> and a lot of different back-end glue, is just not a good use of 
> resources.
> 
> 3) GeckoView and the Fenix front-end are much more modern 
> frameworks than Fennec. They were architected from the ground up 
> using all of the knowledge and experience that we've gained 
> developing Fennec and a number of other experimental browsers 
> over the years. The result is that they are not only much easier 
> to develop and maintain, but also much faster and more resource 
> efficient.
> 
> 4) The current Fennec browser, unlike Fenix and our desktop 
> browsers, is very much single-process by design. Web content 
> runs in the same process as the browser UI. The native Java UI 
> allows us to use threading to work around some of the 
> performance problems inherent in this sort of design, but it 
> doesn't give us any of the security properties of process 
> isolation, which are becoming increasingly important in the age 
> of Spectre attacks. If we wanted to continue maintaining Fennec 
> apart from Fenix, we would need to drastically rearchitect it to 
> support process isolation for web content. And, given that 
> Fenix and GeckoView were designed to handle this from the start, 
> again, that would just not be a good use of resources.
> 
> -- 
> Kris Maglione
> 
> UNIX is simple.  It just takes a genius to understand its simplicity.
>       --Dennis Ritchie

_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

Reply via email to