I'd like to thank Gerv for agreeing to an interview with LQ. He wanted to make it clear that while he is a long-time Mozillian, his views are not official positions of the Mozilla Corporation.

LinuxQuestions.org) First, let me say thank you for agreeing to an interview. Tell us a little about yourself and how you first got involved with Mozilla.

Gervase Markham) My name is Gervase Markham. I’m a Christian, a Liverpool fan and a lover of good cheese, and I live in Sheffield in the UK with my wife and our three young boys.

I’ve been with Mozilla, as a volunteer or employee, since 2000. I got involved when I read a Slashdot comment (!) from an existing Mozilla contributor called Matthew Thomas. It said that if Mozilla failed, then Microsoft would get control of the web. I thought that the web was too awesome, even then, to be controlled by a single company, so I decided to help Mozilla out. Sixteen years later, I’m still here. I’ve done many things in my time, but I currently work mainly on Public Policy, which I tend to summarise as "persuading governments not to make unhelpful laws about the Internet". My current focus is copyright reform in the EU; you can read our policy positions on the Mozilla Policy blog.

LQ) A throwback question that I asked in a 2002 interview of Asa from Mozilla: What do you consider Mozilla's place in the current browser market? If you couldn't use a gecko based browser, which one would you use?

Gerv) Exactly what Firefox should be (which is perhaps another way of asking what its place in the current market is) is something which has been debated recently. You can read my views on that question on my blog. Some metaphors I think work are “butler” and “exoskeleton”. I do think that what people expect from a browser, and what a browser can usefully do for them, has changed a great deal since 2002.

In terms of differentiating ourselves from other options, I think Mozilla’s non-profit status makes it easy - we should be unambiguously on the side of the user, and not on the side of someone else’s bottom line. And we need to both do that, and get better at explaining it.

As a Linux user, if I couldn’t use a Gecko-based browser, then I wouldn’t have much choice of rendering engine! And that is one reason why Mozilla’s continued success is so important. I’m really glad the Microsoft is writing a new, standards-compliant engine, but they haven’t said anything about Linux support. So I’d probably use Chromium (not Chrome) with some changes to make sure my privacy was preserved. At least, until Servo is ready :-)

LQ) I was disappointed when Mozilla declined to participate on Bad Voltage (and subsequently declined to even offer a statement of any kind) after a Mozilla-related segment. I was even more disappointed when Mozilla initially declined requests to be interviewed by LQ, something they have been happy to do before. Based on feedback I've heard, Mozilla now often declines interviews. This seems to point to a culture shift within Mozilla toward being less open and transparent with external organizations, even Open Source related ones. Do you think this is a fair assessment? While some amount of cultural change is inevitable in an organization that has grown as much as Mozilla has, do you think this change will have a negative impact on the long term growth and perception of Mozilla?

NOTE: The person previously interviewed by LQ was on leave and agreed to another interview after I sent these questions to Gerv. More on that soon.

Gerv) I personally declined a Bad Voltage request for an interview because it became clear that one member of the BV team was very keen to discuss gay marriage - a topic which I have no interest in discussing in a broadcast interview (or any interview). I would not be surprised if the other refusal(s) BV got were for similar reasons. I was rather disappointed that this possibility was not mentioned at all in the follow-up BV segment where the team speculated on why Mozilla might have refused.

However, I do agree there is probably some truth in the idea that Mozilla is now more cautious about talking to the media than we have been in the past. As you note, some degree of cultural change comes with organizational growth; it also comes when you have a higher profile. But the events surrounding Brendan Eich’s appointment as CEO last year probably also have something to do with it; that was a very painful period for everyone in the organization, and I don’t think that many people would say that the press attention was helpful as we tried to work through it. More generally, Mozilla regularly struggles with the tension between being open, and kicking off an unnecessary press storm which has to be dealt with. If I were planning our office, I’d put the PR team next to the bar...

We do need to make sure Mozilla both speaks and acts in a way which demonstrates that it’s clearly a part of the open source community, with everything that entails. I’m not sure we’ve got this right at the moment, and want to look at ways we could do better.

LQ) Do you think it's accurate to say there is a bit of internal struggle happening within Mozilla about the future direction of the organization and the best path to take to achieve the Mozilla mission?

Gerv) In an organization of 1000+ full-time people and 10,000+ part-time people, it would be very unusual if there were not differences of opinion as to our future direction! Mozilla has made a number of strategic realignments recently, trying to respond to what is a fast-changing market. Unanimity on every aspect of these changes would be… surprising. Robust debate about such things is important; if there’s one thing we could improve on, I’d say we need to make sure our discussions engage the entire community outside the walls of the Mozilla Corporation as well as inside.

LQ) There seems to be a perception in the community that Mozilla has become less responsive to feedback (to be fair, this isn't a new issue. See https://bugzilla.mozilla.org/show_bug.cgi?id=374002 for an example from 2007, which got comments for almost 5 years). Any comment on this?

Gerv) I’m not sure how one would measure “responsiveness to feedback”. One trouble we do have, and have always had, is that browsers excite a lot of passion in people, because they use them every day, and a small group can be very vocal in demanding what they want even if it’s not the right thing when you consider the big picture. MNG support would be one controversial example I could cite - 8 years on, people still bring that up. So ever since I joined Mozilla 15 years ago we’ve been dinged as being unresponsive to feedback, and it’s hard to tell how much of that is code for “they didn’t do what I wanted them to do”.

LQ) While browser statistics are notoriously difficult to track and hotly debated, all sources seem to point toward a downward trend for Firefox on the desktop. At LQ, Firefox isn't doing too badly. In 2010 Firefox had a roughly 57% market share and so far this year it's at 37%. LQ is a highly technical site, however, and the broader numbers don’t look quite so good. Over a similar period, for example, Wikipedia has Firefox dropping from over 30% to just over 15%. At the current rate NetMarketShare is tracking, Firefox will be in the single digits some time this year. What do you think is causing this decline and what does that mean for Mozilla?

Gerv) Firefox having 100% market share has always been a non-goal for us; the world we want to see is one where there are multiple standards-compliant browsers to choose from. So in that sense, the current situation is a success. Firefox market share is not an end in itself; it’s a means to an end of advancing our mission.

However, we also need sufficient market share to have impact on things like web standards; the stewards of other browsers sometimes support the right thing, and sometimes don’t, and we have to be creative in building alliances, but in the end our market share is directly related to our chances of getting our way. And for that reason, any reduction is bad news, because if we reach a point where we no longer have sufficient market share, that threatens our ability to advance our mission. We recently launched a new marketing campaign, taking into account the fact that many people now use multiple browsers, based around the message “When it’s personal, use Firefox”. Marketing isn’t all we need, of course, but telling people about what you are doing helps. Not enough people yet realise that Mozilla is a non-profit with a mission.

LQ) While Mozilla revenue numbers take a while to be released, the general consensus seems to be that moving from the deal with Google, to the deal with Yahoo and other local providers will result in less revenue. While the move may have some non-monetary advantages such as independence, flexibility and a better alignment with the Mozilla mission do you anticipate the revenue impact will be severe enough to impact the long term goals of Mozilla?

Gerv) There does seem to be a persistent thing going around that the Yahoo! deal (and associated deals to make other engines the default in other markets) were bad for Mozilla financially. I’m not authorized to say anything that hasn’t already been said publicly, but I will point to the positive statements that were issued by Mozilla in the aftermath of the deals being signed, some of which made very clear that we are in a stronger position now financially. Yes, that word is “stronger”, not “strong”.

The search market needs more competition, and I’m glad we are a part of making that happen. Whatever engine you use, this is going to be good for you.

LQ) With both revenue and market share declining, does Mozilla still have the clout it needs to direct the evolution of the web in a direction that is open and transparent?

Gerv) And it’s at this point that I have to dispute the premise of your question - see above :-) But I would say that if Firefox had had a higher market share at the relevant time, then the discussions about EME (the W3C standard for DRM in the browser) might have gone in a different direction. In the end, standing alone, we decided that implementation was better than accelerated irrelevance.

So in two ways, the EME situation underscores how important for our mission we think it is for Firefox to have substantial market share. If we’d had more of it, the discussion might have been different; and to avoid losing it, we had to do something painful. The more people use Firefox, the more likely it is that we can direct the web to be a force for good. We could certainly use more clout, and your vote counts.

LQ) Firefox was initially created because some people within the Mozilla Project felt that Netscape was too bloated, partially due to feature creep. Has Firefox succumb to Zawinski's law and become the thing it aimed to replace?

Gerv) Well, we haven’t yet succumbed to Zawinski’s law, unless you count being able to render “gmail.com” as being able to read mail. In fact, depending on how you look at it and how you assemble the history, Firefox is one of the only examples of the law working in reverse - Mozilla’s flagship browsing product used to be able to read mail (when we shipped the Mozilla Suite) and can no longer do so.

I think that people expect different things from a browser in 2015 than they do in 2005, and the web has significantly more capabilities - real-time audio/video communication, for example. The question is not about whether you add features or not, but about whether you are adding the right features. Of course, there’s room for debate on this too. But a browser has to render the web that exists, and today, that means a lot. Facebook’s current HTML UI, for example, was unimaginable ten years ago, and takes a pretty seriously capable piece of code to render it.

LQ) I noticed something interesting about feedback from my initial post. Many people commented that they have recently stopped using Firefox for technical reasons. Memory leaks and random crashes were the two most cited technical reasons, but there were others. Do you think the quality of the Firefox codebase has decreased? I don't seem to get the random crashes, but do notice the memory leaks. Do either of these happen to you personally?

Gerv) I don’t see many memory leaks or random crashes - I run our Aurora (alpha) version and my about:crashes log says I’ve had an average of 2 per month in the last 3 months. There is, however, a 100% CPU spin I hit quite often and need to track down. I do know we’ve been working really hard over the past couple of years on reducing memory usage, there are memory-usage tests on our continuous integration platform and any regression gets jumped on, and we have some of the best crash-reporting infrastructure of any open source project anywhere. Constant vigilance.

If Firefox is slow or crashy for you, suspect an extension first. Unlike other browsers, our extension model lets addons do anything; and as you can imagine, that has major upsides in available power, but a big downside in that a bad addon can cause instability. Try “Restart with Addons Disabled” on the Help menu and browse around for a while.

LQ) On the topic of my post, do you have any general comments or reaction based on it or the ensuing discussions?

Gerv) Your post surprised me, in that it seems like an odd reason to stop using Firefox. If you really think “the web would be a worse place without Mozilla”, you need to vote with your feet, rather than expecting someone else to support us so you don’t have to.

LQ) There seems to be a perception among some that Firefox has been chasing Chrome recently, especially from a UI perspective. Do you think this is fair/accurate?

Gerv) I noted to someone the other day that for some people, if we add a feature Chrome has we are “chasing Chrome”, and if we add one they don’t have, it’s “bloat”... I think that parity with Chrome on performance (for example) is table stakes for competitiveness in the browser market at the moment, and of course if there are new web platform capabilities, we both add those too. So I suspect people offering this criticism are mostly thinking of the UI. In terms of that, I don’t really know, but it wouldn’t be totally surprising to me if two different sets of UX researchers came to similar conclusions about a topic. And of course, when Firefox started, being like IE so it was easy to switch was (if I remember correctly) an explicit UX goal. Similarity is not necessarily due to lack of creative ideas.

Personally, UI changes don’t bother me too much. I’m fairly adaptable. I had no problem when Ubuntu shipped Unity, either :-)

LQ) The adoption of Firefox OS has been slow and the initial release contained similar vendor lock-in as Android and iOS (I understand this is going to be addressed in a future release, which is great news). What do you think are the chances of Firefox OS achieving success and how does Mozilla define success in the context of Firefox OS. How important do you consider Firefox OS to the future of Mozilla?

Gerv) Success for Firefox OS, from my perspective, would be analogous to success in the browser market - enough market share to provide an open alternative, and to drive the market in a positive direction. This could be measured in raw user numbers, but doesn’t have to be - Asa (who you interviewed a while back) is fond of saying that IE 7 was the best browser Mozilla ever produced, because the rise of Firefox caused Microsoft to have to get back in the game again after they shut down browser development post-IE 6.

I wrote a blog post a while back called “Success Is Not Inevitable”. The mobile phone OS market is amazingly tough. There is a real possibility we won’t succeed; but if we don’t, then the citizens of the world will have fewer options available, none of which would be anything like as aligned with Mozilla’s mission and the values of web openness and freedom as Firefox OS. We’ve made a really good start - 18 mobile carriers in 31 countries, if my figures are current - but we have a long way to go.

LQ) Where do you see Rust going in the next 1-3 years?

Gerv) Rust occupies a unique place in the language pantheon - it’s memory-safe, type-safe, parallelism-safe (is that a word?), integrates well with existing C, C++ or other compiled code (to allow for incremental replacement in an existing application) and it’s fast - as fast as C or C++ if the implementation did all the checks it should, and which Rust does. I’m no language expert, but I’m fairly sure nothing else ticks all those boxes. It’s just what we need for Servo, our next-generation rendering engine, but I’m pretty sure it’ll have a load of other applications too. Watch for it on the Internet of Things.

I’ve been saying for a year now that I’d love to see an SSL stack written in Rust. Servo needs one anyway, and given all the problems SSL stacks have been having recently, it would be a much-needed addition to the options. Of course, writing crypto code is not for the enthusiastic amateur; there are probably not many people in the world who could pull this off. If you are one of them, think about it.

LQ) A question from an LQ member: What is done to keep Firefox up to date from a technical point of view (IMHO, it can not compete with Chrome and other Webkit/Blink-Browsers anymore in terms of speed), how far is progress with the new Servo engine and will the extended capabilities for plugins be kept when a change happens. The Pentadactyl plugin I use is, AFAIK, not possible on Chrome and similar browsers due to a lack of capabilities for plugins, so that Chrome users have to use rather featureless copies of Pentadactyl, like Vrome or Vimium.

Gerv) Well, I think our speed is pretty competitive :-) Servo is coming along nicely, and we are in the middle of integrating the first bit of Rust code into Firefox itself, as a test. Once that’s in, more bits of Firefox will move over to Rust implementations.

As to what will happen with addons for a browser based on Servo, that’s much less about Servo itself and more about the browser built around the engine. That doesn’t really exist yet - at the moment, Servo only has fairly simple UIs built around it for testing purposes. But however Servo ends up being used by people, I think Mozilla is unlikely to forget that powerful addons, while being a right pain in the backside in some ways (see above), are also a big reason why people use our products.

LQ) Anything else you would like to add or think people should know about Mozilla?

Gerv) I hope that every LinuxQuestions reader already knows that Mozilla is a non-profit-making organization, with a mission to preserve and promote the good of the web. But I would add that the browser you choose is not just a statement about which order you like your UI buttons in, but a vote for how you want to see the future of the web.

So if you think Mozilla’s vision of an open web, available for all to innovate on a level playing field without toll booths or speed bumps, is a good one, then use Firefox, and encourage all your friends and family to do so as well. And more: test your websites with Firefox, use and give feedback on our developer tools, write your browser addons for Firefox, make sure your distro ships Firefox by default. Every little helps, and you can make a difference.

To address:
I found I was using Firefox for philosophical reasons (along with inertia). Me questioning whether those reasons still existed made me step back and ask if I'd use Firefox, today, purely for technical reasons. For me, based on performance and stability, that answer is probably No.

--jeremy

I've also sent this follow-up question:

LQ) In the interview, you said "Unlike other browsers, our extension model lets addons do _anything_" (emphasis yours) and "I think Mozilla is unlikely to forget that powerful addons, while being a right pain in the backside in some ways (see above), are also a big reason why people use our products". Based on comments from some addon authors such as Classic Theme Restorer (https://addons.mozilla.org/en-US/fir...orer/versions/) it appears what addons can do has recently been reduced in a way that is impacting existing addons. I'm not referring to the digital signing here, but to new validator checks that come along with the signing. Is it still fair to say addons can do _anything_? Do you think this decision will result in more power users switching away from Firefox?

--jeremy

I am not certain about the addons situation you mention, so I wouldn't want to comment definitively, but I _think_ there has been a bit of a misunderstanding around the new signing requirements. There is a validation check which comes along with signing, and some people interpreted messages, which were intended to be warnings, as requirements for signing. I hope that this will all work itself out.

Gerv

Thanks again Gerv.

--jeremy

Crom secure, flexible, free of charge. Google or Linux version of drivers weigh in closed or open. Mozilla is lagging behind a little.

Thanks for the interview, Gervase and Jeremy!

It's always nice to hear from the Mozilla/Firefox people.

Just to second Myk267, thank you both.

I still use Firefox because its goals for the web align with mine. But as Gervase said, "There is, however, a 100% CPU spin I hit quite often and need to track down" -- I hit that too. It usually takes several days of having Firefox open with multiple tabs before I hit it, but then it hits and I go nuts. My only add-on is Ghostery, but as far as I can tell - and I could be wrong - Ghostery reduces the frequency of my encounter with that bug.

I rarely have the mental energy for hacking on code in my spare time, the day job sucks the mental energy out of me. But I do think Firefox is one of the most worthwhile projects on the web to aid.

Michael I have the same problem. Whenever Firefox runs out of physical RAM, it starts trying to access the swap file. It does this very badly, which is why you end up with the 100% CPU hit (on one core; to people with multiple cores, it looks like less, but the browser still becomes unresponsive). I tried installing an add-on which unloads tabs after so many minutes of inactivity, ostensibly freeing up the RAM they used, but Firefox still seems to be caching everything. I think I tried turning off the cache, but it's still sucking up RAM for some reason.

The reason there is no serious Firefox fork in existence today, is because its code is figuratively similar to the Labyrinth of Crete. The whole thing is built from a single massive project that takes two hours (or longer) to download and compile. Snaking across all its code is this construct called the "security principal" which defines what can and can't be done. Half the code is in Javascript, and the other half is C++ and the afore-mentioned Rust. Mozilla relies on a system called XP-COM, an invention of IBM, both to interface with the operating system and to enable Javascript integration. XP-COM is the real root of Mozilla's problems: it is bug-ridden and by its very nature prone to security gaps. But it's everywhere in the code and it obfusticates the whole system to an enormous degree. To rip out XP-COM is tantamount to starting over from scratch.

I'd like to see a whole new project. I'd like to see people think critically and roll up their sleeves, and start re-inventing the wheel frankly. In my view, open source isn't truly open unless it's readable. That means shirking languages like C++ and moving to languages like BASIC, which by now has very competitive performance and capability. People want things to be easy, to make a quick buck. But if you want to make something that endures, you can't skimp on the foundations, because they are most important part of the structure.

It isn't a cache problem, or it's more than a cache problem. My personal desktop has 12GB of RAM and an (older) AMD 6 core processor, and Firefox and the desktop environment will be the only programs running. The "about:memory" tab in Firefox will indicate that I'm using 2.5 GB of RAM with maybe 50 tabs open, so there should be literally 9GB of RAM available for Firefox to use before hitting the swap file - and I still got 100% CPU usage for one core and hangs.

This hasn't happened to me in the past month, but I installed Windows on a spare partition for my kids to play some games and now Firefox never runs for more than a day before the computer is restarted.

To be fair to the Firefox developers, the project is old. Chrome and now Microsoft Edge have an advantage because they started fresh much more recently and learned from past mistakes of other browsers.

Regular Firefox is 13.5 million lines of code with just 7 million lines of C++ and just 1000 lines of Rust, https://www.openhub.net/p/firefox/an...guages_summary

Rust is in Mozilla's new project to rewrite Firefox called Servo: https://github.com/servo/servo (115,000 lines of Rust, 65 lines of C++, but the project is far from finished and production-ready) https://www.openhub.net/p/mozilla-se...guages_summary


Open source is open as long as the license is open. The people who write the code owe the people who download it nothing. It's beneficial for the code to be clean and clear, but not required.

C++11 and C++14 are excellent languages, and C++17 will be even better. A clean rewrite of Firefox using brand new C++14 would be fine. But again, Servo is the clean rewrite that Mozilla is working on and that uses Rust. Rust, at least in theory, makes it simpler to write secure code, and to write parallel code, and to write concurrent code versus any version of C++ without losing much to C++ in terms of performance.

I have a C2D with 2 gigs of RAM and 500 meg swap file, and it behave exactly the same way! Strange.

https://github.com/servo/servo/wiki/Design

^ My take on this is that Firefox will be less open source friendly than ever. I mean when you write a whole new language to write your browser... you don't wall your garden off more than that in all but name. You're right that open source is technically anything which is made publicly available, but there is a difference too between open source in name, and open source in substance. Firefox is not open source in substance by any definition because the barrier to contribution is practically insurmountable without collaborating with Mozilla themselves. The documentation is poor and spotty, and the Mozilla people frankly won't explain their often arcane technique except to project insiders.

It'll take at least 7 years to implement a RUST Firefox. We won't see most of what's in that project plan until 2025 at the earliest, assuming Mozilla endures that long (and it's questionable they will). Me, I'm only around because they are the only browser to still allow read/write access from the hdd without forcing the user to invoke a gimmick of some kind. I grant that they are the sole remaining resistance to Microgoogle/Googlesoft, but it's not very inspiring.

Unlikely or not, it seems that Mozilla indeed has forgotten and phases out their old extension APIs in favor of the limited APIs of the Chrome browser: https://blog.mozilla.org/addons/2015...refox-add-ons/
The proposal that the Mozilla developers will work with developers of popular extensions to add the functions they need is very short sighted, because obviously this can only be done for already existing extensions. You can't create a new extension to get it popular enough to be considered worthy of extending the API if you can't write your extension because of the limited API in the first place.

For me this decision is just another nail in the coffin of Firefox and I have increase time spend with evaluating other browsers that can offer me what I expect from a browser. It seems Firefox doesn't want to be on that list anymore.