The Fox Hunt - Firefox and friends compared

Updated: November 20, 2017

Are you familiar with the lyrics of The Reflex, by Duran Duran? The reflex is a lonely child, who's waiting by the ... and so on. Well, Firefox is not a lonely child. In fact, the Firefox family has many members, and now that we're facing the crucial moment of truth, whether to use Firefox or not, at all, given the radical change a-coming' with Firefox 57, it is time to give the entire series some extra spotlight.

In other words, let's try to figure out which of the Firefox siblings is the most suitable for everyday use. We're talking look, compatibility, the ultra-important extensions, security, performance, and such. Today, we'll have a wrestling match between Firefox, Waterfox and Pale Moon. As requested by you, readers. Shall we?


Before we begin ...

This is not a browser benchmark. In fact, most benchmarks are largely irrelevant, because they are often too academic or too specific. Then, Javascript execution times are perhaps relevant if we're only talking about websites with pretty much images-only swipy type experience, the kind (mostly) plebs use, but if there's so much as one sentence of text to read, the exercise becomes completely different.

So if you have a browser and you want to time its speed, then there are the following considerations: default build versus custom compiled, 32-bit and 64-bit applications, operating system, running services and current system load, underlying hardware components, device network type and speed, network peripheral capabilities and performance, worldwide geographical location and ISP, time of day, and then finally, the wider repertoire of websites being accessed and used, which again have their own peculiarities, including server load, use of CDN, content type, and so much more.

Unless you can test over a pool of say 100,000 users worldwide against 1,000 websites, it is almost impossible to get an accurate picture of the proper technical browser speed. But then, it doesn't really matter. For most people, perception is what counts, so if they feel something is faster, then they will think it is. The circumstances are less important. And with all that in mind, let's blaze on.

Firefox 56-57 (give or take an Integer)

So, first we have the original article. As I've written in my review, Firefox 57 is better than the previous 53 versions, as it combines adequate looks with security and performance. The new underlying improvements help it gain speed and perceived speed against the likes of Chrome.

I don't have too much to add here, except that I have Firefox 56 - current at the time of writing - installed alongside Firefox 57 Beta, Firefox 58 Nightly and Firefox 57 Dev, all 64-bit versions in the Kubuntu 17.04 Zesty distro on a Lenovo G50 test laptop.

Firefox with WebExtensions does not have the plethora of extensions that earlier builds do, so different profiles and different versions above come with different add-ons. The standard edition comes with all the bits and pieces, and there's an extensions for everything. The higher releases are rather limited.

Firefox running


The browser is fairly fast - and elegant. My one speed benchmark was running a full HD video in Youtube, in a second browser tab, in addition to a static HTML page. I checked the overall memory consumption on idle, and CPU behavior while playing and decoding the video. Again, here, there's a lot of variety, as there are tons of formats. In general, if you want to check and compare, this will be my Megane at Spa racing clip.

The standard edition (Firefox 56), with all its sub-processes (two additional Web forks) uses about 350 MB of RAM with a single tab open, plus some 100 MB of shared memory. With the Youtube clip playing, CPU usage hovered around 20% on average, with occasional spikes to about 25-27%.

Firefox 56, memory

Firefox CPU usage

Now, if we look at the Firefox 57 Dev build, memory figures are largely identical to the standard edition. There isn't really any major difference. Likewise, processor utilization was very similar.

Firefox 7 dev, memory

Pale Moon 27.5.1

This is an interesting project. Pale Moon tries to be what Firefox was before Australis, building on this foundation and adding to it. In the past, it was only about keeping the stupid new interface away, but recently, it has become much more. Pale Moon now uses its own rendering engine called Goanna (as opposed to Gecko), and it also has its own subset of extensions, and there's been some divergence from Firefox proper.

I tested version 27.5.1. The looks aren't native to Plasma - more Debian/Gnome stock if you will. But then you do get the sane interface, with tabs on bottom as it should be, and everything else that constitutes sane, non-monkey behavior for a browser.

Pale Moon running


Pale Moon offers a hybrid extension model - both what you can find on Mozilla's site prior to WebExtensions, and its own subset. In particular, since I'm using Adblock Plus in Firefox, it was interesting to learn that this particular add-on is not supported in Pale Moon. In fact, it is blocked because of known stability issues, and you should use a custom alternative. I'm not interested in politics, only the technical bits.

Pale Moon addons

The first hit is not compatible. The REAL alternative is not listed here.

I was able to install ABP, which is annoying in that you get an extension but then it gets blocked, so why allow the installation in the first place? Also, I'm worried what this means for the future of Pale Moon if standard extensions disappear or stop being developed, how will the browser handle the change?

ABP blocked

In the end, I did have an adblocker, but it wasn't the most productive exercise. Also, UI wise, there's inconsistency in the size and quality of icons, and they do not all match up, and the non-native KDE integration in this particular case is also jarring. I know that Plasma users will be a tiny minority of Pale Moon users, most will be running Windows and that's where the focus is, but still.

Icons are inconsistent


A very interesting set of results. Day to day, Pale Moon is faster and more responsive than Firefox 56. Things like opening new tabs, for instance. Newer editions of Firefox narrow down the gap, with Firefox 57 Dev and Firefox 58 Nightly matched to it. But then I'm asking myself, if Pale Moon can be fast and responsive with an "older" engine, what did stop Mozilla from doing this in the first place?

Memory wise, same conditions like Firefox, it eats 220 MB RAM with additional 85 MB of shared memory. This is about 20-25% less than Firefox, so that's nice to see. Probably all the telemetry stuff stripped out. However, with the HD clip playing on Youtube, Pale Moon is a bit less energetic than Firefox. It ate more CPU, at about 30% on average, and it felt more sluggish. Minuscule, nerdy differences.

Firefox 56, memory

Pale Moon CPU

Waterfox 55

The third of the trio is Waterfox. This browser is very similar to Firefox, and while it does follow the same fast-build plus ESR model, it may switch to WebExtensions entirely some day in the future. Undecided yet, which is a bit alarming. The major differences are, like Pale Moon, it strips away all sorts of telemetry, data collection, pocket, and other really unnecessary stuff. At the time of my testing, version 55 was the official and current build.

For now, Waterfox 55 has the same extensions model like Firefox 56. So I had no issues whatsoever adding stuff and customizing the browser. In fact, I even made it look like my default profile, and they look identical. Speaking of profiles, Waterfox lets you import either Firefox or Chrome stuff, so that's quite nice.

Imported user settings


Waterfox running


With the same set of extensions like Firefox, Waterfox is a bit hungrier both memory and CPU wise, which can possibly be explained by the fact it's one version below the parent, and there have been additional code improvements since. Of all the tested browsers, Waterfox ate most RAM, about 420 MB, roughly 20% more than Firefox and almost double what Pale Moon does, with the shared memory again 50% and almost twice as much, respectively. Not sure why, but there.

Firefox 56, memory

CPU behavior was somewhere between Firefox and Pale Moon, about 25% on average, with the overall browser responsiveness a bit behind the rest. I guess browser engine improvements, this one. Nothing cardinal, though.

Waterfox, CPU

Overall performance summary

With all three browsers squared, what we have is the following: Newer versions of Firefox are ever so slightly more responsive, which is what matters to most people. Technically, they are also faster, but that's nerd cream.

Pale Moon is really good for simple stuff, but it struggles a little under heavier load, yielding its performance lead to Firefox 57 and above. We're talking tiny percentage, and for most people, it won't really matter, but then we're in this manic race of matching up to Chrome, so in sadness swim we must.

Waterfox takes the last place, matched and slightly below Firefox 56, but then, it's probably as it's running one version behind (55), with improvements added since, and also possibly because there are subtle differences in the builds and whatnot. Not bad, just the way it is.


So what should you use? Well, it depends. You want extensions, the entire repertoire as it's meant to be? Go with Pale Moon, but be aware of the inconsistencies and problems down the road. However, another piece of penalty is less than optimal looks. If you are more focused on speed and future development, then it's Firefox, as it offers the most complete compromise. The add-ons will make it or break it. Waterfox makes less sense, because the margins of benefit are too small.

My take is - Firefox. It's not ideal, but Pale Moon does not solve the problem fully, it combines nostalgia with technicals, and that's a rough patch, even though the project is quite admirable in what it's trying to do. Alas, I'm afraid the old extensions will die, and the new ones won't be compatible, so the browser will be left stranded somewhere in between. But hopefully, this little comparison test gives you a better overview and understanding how things work.

Finally, we go back to the question of speed. We've seen how one flavor of Fox stacks against another, but what about Chrome? I will answer that in a follow-up article, which will compare Chrome to Vivaldi, again based on popular demand, and then we will also check how all these different browsers compare using my small, limited and entirely personal corner of the Web. Stay tuned.