Add-On topic of what I think is the most important component of a web browser: A timeline of JavaScript Just-In-Time (JIT) engines.
https://egbert.net/blog/tags/jit.html
sorry, last time I checked on March 2022, Google Chrome cannot negotiate for my ChaCha-only TLS website; instead try using a Safari, Brave, Firefox, Edge, Aloha, OnionBrowser, Orion, Links, or Lynx web broswer, to name a few).
Meanwhile it is an ongoing crazy ride just mapping the evolution of WASM (in my next planned blog).
It's a shame Opera didn't go to the open source route. Opera with Presto was a great browser.
One cool option missing from this list is OffByOne. Very fast and lightweight, Peak HTML 3.2.
https://wiki.c2.com/?OffByOneWebBrowser
I develop a framework for building hyper-compatible Any Browser websites (applications include retro-computing audience) and I include it in my test suite.
Some of you may enjoy watching this too: Desktop Browsers Market Share over time (pie chart race) https://plotapi.com/explore/view/563abc04-f196-4992-aefe-973...
> This promising engine [Servo] was developed by Mozilla[...] Mozilla fired a quarter of their developers, which apparently included the whole Servo team. There have still been some commits to the code since then (presumably by hobbyists) but it is questionable if Servo will have a future.
Why do people keep framing the story this way? Servo's future is in Gecko.
always_has_been.jpg
The Servo repo was a testbed that allowed people to work on new, Rust-based browser components without anyone having to pass the type of code reviews that are necessary for a production Web browser that is by the way already continually shipping to millions of existing users.
This (far too common) meme of Servo as a somehow failed separate browser engine that was supposed to, I dunno, be swapped out at some indefinite point and retire the lizard or something is very weird.
I wrote a browser for Macs (System 7) in 1995-7 through a contract with James Gleick's Pipeline ISP [0], commercially available in the US. Don't see it listed, will reach out.
This only briefly references Tasman which was the rendering engine from IE For Mac. It was in many respects better than trident - the big bit was it had a much lighter footprint which meant it became the engine used for mobile IE.
Curious what are the advantages of using expensive custom embedded browser engine like sciter / ultralight over system webview, or using system webview wrapper like https://github.com/webview/webview?
First graphical browser I used was Slipknot [0], which doesn’t appear to be listed.
[0] https://en.wikipedia.org/wiki/SlipKnot_(web_browser)
(Edit: I see it now, must have skipped over it before.)
Cool graph, but I think in cases where. Lot of your data ends up indistinguishable from zeros it's better to use something like a log scale. We all know Chrome is the biggest by far, but I have now well to tell how the usage of, say, Links has changed over time.
There is a large amount of HTML found on the web that only serves to wrap textual information and is therefore human readable without rendering. Basic text processing is enough. IMHO, HTML tables are where "rendering" becomes a must-have.
Others I know of: MMM (CAML Light with applets, '90s), Emacs/W3 (which Bill Perry claimed had the first CSS implementation), Abaco (Plan9)
Has anyone tried to make a browser using the MS Word engine? It would be horrible but also kinda neat.
What does it mean "WebKit peak 49% market share" in 2012. That doesn't pass the sniff test. AFAIK, Safari never had a big install based on Windows and MacOS+iOS certainly didn't account for 49% of web traffic in 2012
What comes after Blink?
pours one out for Opera 12 (Presto)
"Gecko (Firefox). Down to 4% market share, mismanaged by Mozilla which prioritizes pushing its toxic politics over improving the browser."
Unlike the author, who clearly has no agenda or axe to grind.
Obligatory reference to the inverse relationship between Firefox market share and the Mozilla Foundation's chair's compensation [1].
It is sad to see Firefox become so irrelevant. As much as people blame Google for this (and it is true Google relentlessly pushed Chrome) but people forget just how innovative Chrome was and how Firefox didn't respond to these issues.
I remember when Chrome launched and it was revolutionary how it was one-process-per-tab (technically, it's site isolation not tab isolation but let's not get lost in the sauce). No longer could an errant website take down your entire browser (mostly). I kept wondering why Firefox didn't copy this. It took them years. What were they doing?
Now I appreciate Mozilla bringing Rust into existence (not without problems and early design mistakes [2]) but the initial goal seemed to be rewriting the browser in a memory-safe language and that never seem to eventuate..
[1]: https://calpaterson.com/mozilla.html
[2]: https://www.pingcap.com/blog/rust-compilation-model-calamity...
Interesting synopsis and confirms something I've always felt: There was never really a plurality of web browsers, there was always one that held an outsized majority vs the rest and drove web development practice. The closest, perhaps, was a brief period before Chrome became dominate and IE was waning fast, where I believe Safari, Firefox, and Chrome held approximately the same market share vs IE, which would be in the 2010-2013 era (peak Webkit was 2012), which I personally regard as one of the most interest times to be both on the web and be a web developer, it was also before Chrome forked Webkit fully IIRC.
FWIW, I know having Chrome / Chromium as the overwhelming majority browser is not great, if for the sheer fact competition keeps everyone "honest" in a way, but they are by far the most "benevolent" from a developer perspective. IE was truly both stagnant and terrible.
EDIT: that's not to say I approve the Chromium dominance, as a daily Firefox user especially, but I would be lying if I said, from a developer perspective, that Chromium hasn't been pretty good so far on balance. They do innovate. They do push new features. They do usually support the latest specs. Though again, I don't approve of it being so dominate, I'd prefer a plurality. Its a shame that Microsoft didn't use Firefox as its base for new Edge