Mozilla lays off 70

Mozilla uses AFL, which is a genetic algorithm that tests code paths. They are also transitioning to Rust, which will give them a much bigger safety guarantee over most of their code and a much smaller audit surface for the rest.

Rust has nothing to do with QA things since most bug are business logic / UI / UX related. QA don't test for those kind of errors.

They're transitioning quite slowly, a sibling comment mentions https://wiki.mozilla.org/Oxidation#Rust_Components this page. Even accounting for how they're obviously prioritizing their efforts to get the best bang for the buck, there will be a lot of critical C/C++ code in Mozilla products for a pretty long time.

In my experience, that’s the exact sort of errors that a QA team our position tests for.

American Fuzzy Lop? It's a fuzzer. I've never heard of a fuzzer described as a genetic algorithm that tests code paths, but that is technically correct.

A fuzzer is not going to replace unit tests or good SDLC, which often involves QA.

Their comment meant that QA don't test for the kinds of errors (safety) rust protects against.

Mozilla is good at everything except making browsers. rr is the best debugger I've used. I wonder if there's a way for them to monetize Rust. It seems impossible to monetize a programming language without owning a platform e.g. Microsoft (*.NET), Apple (Swift), Borland (Delphi), JetBrains (Kotlin). FirefoxOS for mobile with Rust as the first-class citizen would have been huge for them. Maybe it's not too late. Take note Mozilla.

If there's anything they're going to be in the history books for, it's going to say Rust, not Firefox.

Hi, I'm the architect of rr.

Mozilla is very good at making browsers. Making browsers is incredibly hard and it's amazing that Firefox is competitive with Chrome given a fraction of the development resources Chrome has.

Mozilla has made some big mistakes, but so have the other browser vendors. It's easier to brush over your mistakes when you have an ocean of resources and market power.

Mozilla is already in the history books for Firefox, any way you slice it. Firefox has been the only serious refuge from proprietary browsers and their vendors for nearly two decades, and has made history in many ways. It would only be omitted if the book was written by a competitor's PR department.

Quality and popularity are rarely correlated, but such metrics are even further confounded when you have monopolistic factions like Google and MS pumping huge money into user acquisition for their competing platforms. It's ridiculous to imply that Firefox is being rejected by users just because Google has been dumping millions into getting people switched to Chrome (and doing so not only with marketing dollars, but increasingly with bully tactics).

In terms of contributing to declining market share, my personal opinion is that Mozilla's big mistake with Firefox was killing embedding around 2010.

Thanks for rr - it's such an amazing tool, and a really impressive engineering accomplishment.

Indeed, and if you count the value of in-kind support for Chrome, such as constant "switch to Chrome" advertising on the most important Web real estate in the world, we're talking billions, not millions.

I would love someone to estimate an actual figure for total Chrome marketing spend including in-kind. I'm guessing it would be well over $10B.

GA's are a good generative path for fuzzers; at a low level that's exactly what they do.

Thank you for rr, once again. Great tool. I hope to contribute one day, as soon as I find an itch to scratch.

Could be, but it's awfully hard to say. What would have been cut to make room for embedding work, and what would the impact of cutting it have been? Did we know enough at that time to make a good decision about that?

Given that the embedding API would have had to change dramatically due to e10s pretty soon after 2010 (and maybe again due to Fission), I think 2010 was probably a bad time to promote a stable embedding API.

Embedding doesn't rank high on my list of Mozilla's mistakes.

If Mozilla used all its income derived from Firefox on Firefox would it really be that underfunded compared to Chrome?

Absolutely.

You're welcome!

One contribution we can always use is people writing and talking about rr. One of the biggest things holding rr back is so many people just don't know it exists (or they know it exists but they don't appreciate what it could do for them).

Absolutely it would or wouldn't be underfunded?

I assume he meant "absolutely it really would be that underfunded", because that is true.

Elsewhere in these comments people have estimated that Google pays at least 1000 people to work on Chrome. That's about the size of all of Mozilla, and a lot of those Mozilla staff are necessarily not working directly on Firefox --- you need HR, accountants, marketing, etc. Also, Google pays its developers significantly more than Mozilla does, on average; Mozilla developers tend to get big raises when they move to Google.

And that's just direct spending. Historically Google has done a lot of Chrome marketing on its Web sites, which is prime advertising real estate that would cost astronomical amounts of money if it was for sale. And historically Google has paid hardware and software vendors to preinstall Chrome, which is also expensive, though I'm not sure how much that happens these days.

I've not seen any estimates that Google employs 1000 devs to work on Chrome. Only that in the more than decade long history of Chromium 1000 Google devs have worked on it in total.

I don't want to dig through hundreds of comments again --- but as someone who worked on Firefox for 15 years and knows a bunch of Chrome people, I believe those estimates.