State of Embedding in Gecko | Hacker News Reader

State of Embedding in Gecko

chrislord.net

17 pointsprotomyth9 comments

Wow, maybe he didn't mean to do this, but this is quite an indictment against the current Mozilla management team!

"Current"? Embedding Gecko has always been an exercise in futility, which is why everyone and their dog adopted WebKit with gusto. Simply speaking, Mozilla always expected people to build on Gecko, rather than with it. You were expected to go XUL and treat Mozilla as your main runtime, developing the whole program with JS on top of it. Embeddability was always an afterthought. They took this approach to extreme levels with FFOS.

Khtml/Webkit used the opposite model and won big.

Funny enough though, Mozilla's insistence on JS as general-purpose was likely a factor in the long-term survival of what was, back in the days, a wobbly and improvised browser-only scripting language. In many ways, you could not have had Node without the XUL years; but the JS explosion happened on Webkit components, (fatally ?) wounding the Mozilla platform. History is funny like that.

This really isn't an accurate representation of what happened. WebKit's success didn't happen overnight, and people weren't jumping on it from the get-go. In fact, there was a long period of people using Gecko and engineering their own embedding/browser solutions because it was widely, and correctly realised that Gecko was a more mature and featureful engine. This was my job for several years before I was employed by Mozilla (and indeed, I did do some WebKit work at the time too - but there was far less demand for it).

It was only through a long period of neglecting these users that WebKit was allowed to mature and become successful. There was a good reason for doing this (see yoric's blog post, linked in my other comment above), but I think we've past the time now where we need to reconsider this direction somewhat.

It didn't happen overnight because any embedding project takes time; but the webkit community expanded relatively quickly as Mozilla simply did not care about their (overall small and shrinking) own share of the market. You said it yourself, they just did not care enough. Look at Qt, which had struggled for a while with producing reliable Gecko bindings and then really turned a corner with Webkit.

The death of support for embedding Gecko is at least part of what eventually killed Camino, too. (I was a contributor.)

Camino was my favorite browser back in the days of PPC macs, I have really missed having a Gecko-based browser with a native interface on any platform since then. Ditto when Epiphany made the switch from Gecko to WebKit for the same reasons (and webkit-gtk is really awful compared to Gecko, every time I try to use GNOME Web these days I switch back to Firefox almost immediately because it doesn't load half of the web apps I use).

He's not just talking about embedding. He's also talking about their pivot into the IoT space.

(disclaimer; it's my blog post)

I don't think it is - this post by yoric goes into some detail as to why killing embedding in its previous form was a good idea: https://dutherenverseauborddelatable.wordpress.com/2016/03/0...

I do think it was a poor decision not to support EmbedLite more fully, but looking at bugs in Bugzilla, that appears to have been down not only to Mozilla's general direction, but also the decisions of individual engineers, with what looks like sane reasoning (that I partly disagree with, but then I'm likely not privvy to all the facts).

Hindsight is 20/20, it's easy to point at mistakes that have already been made and there's not much we can do except learn from them. What I'm really worried about is that we've committed to this 'no embedding' path purely out of apathy to change it now. The sort of "Oh, it's a really big job to do anything now, we don't have the resources so we're not going to bother".

yoric makes the excellent point that we're significantly cutting down our potential platform engineer employment pool. Much like him, I was also hired because I was working on embedding Gecko, and thusly, Gecko platform. I fear that we're going to miss out on whatever the next big thing is because we're not enabling people to use our (really rather good, and liberally licensed) technology, and that we're not going to be in a position to keep up with our competitors because we can't offer anything compelling to the people with the necessary skills to help us. That would be bad for the web ecosystem as a whole, which is something it appears I've come to give a damn about.

As for IoT, most new advances in technology are met with skepticism. While it's going to take a lot to make me say I believe that there's a viable market for many consumer IoT devices, I'm sure many people said this about smartphones, about laptops, about the TV, etc. and the Nest has already proven that it's possible to make a compelling product. I think it's sensible that we're investigating this area at the very least.

> That we haven’t supported this effort seems insane to me, especially as it’s been shipping for a while as the basis for the browser in the (now defunct?) Jolla smartphone.

Exactly. It's a shame Mozilla never accepted EmebdLite officially. What was the problem really?

Sailfish is not defunct by the way.

> The quality of the work done on this seems quite high to me, after a brief examination.

@romaxa knows what he is doing. He worked on embedding Gecko for a long time already (including in Nokia).

Reply on news.ycombinator.com