Servo is starting to become usable
book.servo.orgI genuinely wonder what "starting to become usable" means there. Is it the first time that anyone can load a simple static page? Or is it good enough to browse most of the web, but some features are missing like... I don't know WebRTC or passkeys?
A simple static page is not equal to 'starting to become usable', that'd be 'passes the first experimental test'. By 'starting to become usable' I mean Nextcloud and all related apps I've tested work as intended and I'm posting this message from within servoshell. I've kept track of its progress just like I did back in the time of its spiritual predecessor - Mozilla, then Phoenix which became Firebird which became Firefox - and was until recently not able to get it to show much more than that simple static page you mentioned. While it may have passed a bunch of verification suites it - servoshell if not the Servo rendering engine itself - it was not usable for much more than showing that static page. Now, it is. Its usability resembles that of early, pre 1.0-versions of Firefox. I used those just like many others did, warts and all, and reported and where possible fixed bugs.
Why now post this for Servo and servoshell? Because it seems to be close to the same state as Firefox was when it became somewhat usable and with that it became possible for others to start using it and report any bugs and hiccups. The more people do this, the fast the progress will be. A usable alternative browser engine and with that usable alternative browsers is a big, BIG thing and worth reporting as well as supporting in its progress.
I was kindly asking more details about what "starting to become usable" means (i.e. should I try it or does it feel like a tech demo?), and I feel like I am being sermoned in return :-).
> Its usability resembles that of early, pre 1.0-versions of Firefox.
Not sure what it means to me. Pre 1.0 of Firefox was what... 20 years ago?
> The more people do this, the fast the progress will be.
As a developer myself, I kindly disagree. I mean this is not a general rule. When my code is not ready, having tons of people test it doesn't help me: I just ignore their reports because anyway my code was not ready for their testing.
Is the Servo team asking for that, or is it just your feeling?
> Not sure what it means to me. Pre 1.0 of Firefox was what... 20 years ago?
Why does that matter? For those who were around at that time the release of the Netscape code was a big thing. This led to the further development of the Mozilla browser which was close to the closed-source Netscape suite, including loads of tools which were used by few but increased the code size and complexity. Eventually a new browser shell was created based on the Mozilla Gecko engine and the XUL user interface markup language. The release of the first usable versions of Firefox promised to fulfil the expectations raised by the release of the Netscape code: a relatively lightweight, capable browser which could be the tool to escape the stranglehold which Microsoft had on the web through its (ab)use of their desktop platform dominance and the Internet Explorer browser.
Now we're close to being back to where this started with Blink/Chrome instead of Internet Explorer - and for web developers the annoying presence of Safari with its lackluster adoption of 'standards' which threaten the app store income stream - and there is a clear need for an alternative. Servo may end up being this alternative no matter whether the Servo developers deem this to be the intended purpose of their work or not. As to whether they consider their code to be ready for testing I can only refer you to the site I linked to.
I have the idea you're reticent in accepting the possibility of Servo becoming the needed alternative to the Blink/Webkit hegemony. Am I correct in thinking so and if so, why is this the case?
Also, as a developer myself I subscribe to the 'release early, fail fast' philosophy. Early failures tend to be small failures, easily corrected. Servo has been under development for more than 13 years as a research project, first inside Mozilla, then outside after Mitchell Baker considered her ideologically driven political activism more important than Mozilla's actual core mission of developing web technologies and fired all Servo developers in 2020. Development continued nevertheless, now by volunteers. Here's where to go if you want to contribute to the development:
> I have the idea you're reticent in accepting the possibility of Servo becoming the needed alternative to the Blink/Webkit hegemony. Am I correct in thinking so
No, you are not.
I guess I was just trying to kindly ask "can you describe in a few words how usable Servo is today?". I was genuinely interested. Not anymore, but I do strongly hope that there will be an alternative someday. Apparently not today.
> I was genuinely interested. Not anymore, but I do strongly hope that there will be an alternative someday. Apparently not today.
Why are you no longer interested? The more eyes, the shallower the bugs, the more heads and hands, the sooner they're squashed and with that the the closer that alternative will be viable. It doesn't happen by itself, not back in the Netscape->Mozilla->Firefox days and not now. Run it besides your current browser for less critical tasks - like commenting here - and see how far you can take it. When it tells you it crashed - which for me has happened a few times, not the shell but the engine - see if you can recreate the crash and report it or - better still - submit a patch.
> Why are you no longer interested?
Because of your sermons :-).
> The more eyes, the shallower the bugs
Honestly at this point they don't seem to need help finding bugs or missing features. As long as the devs can find them in 10 seconds themselves, they don't need reports from users who won't contribute code.
> Honestly at this point they don't seem to need help finding bugs or missing features. As long as the devs can find them in 10 seconds themselves, they don't need reports from users who won't contribute code.
Then why, oh sage, does the project ask people to Get Involved [1] and tell those who click that button - the first prominent one on their landing page - the way to the code [2], the way to contribute [3], the way to build the project [4] and where to report any issues they run into [5]?
Also, what is that nonsense about the developers being able to find them in 10 seconds themselves? It sounds to me like you've swallowed the LLM propaganda hook, line and sinker and consider field testing and bug reporting superfluous now that your personal Robby the Robot takes care of everything. Maybe he will, or maybe he won't, who knows? Talking about LLMs, the project does have some stipulations around using them to contribute [6] which spell out the reasons why Robby might not be as helpful as he seems to be.
There is a current thread going on this subject related to one of the LLM companies pushing their latest iteration of a 'bugbot' [7] which might end up showing some pros and cons of this technology. It works fine until it doesn't, the problem is it will be just as confident spewing nonsense as it is when it has found real problems. It is up to you to decide when you're dealing with the former and when with the latter.
[1] https://servo.org/contributing/
[2] https://github.com/servo/servo
[3] https://book.servo.org/contributing/getting-started.html
[4] https://book.servo.org/building/building.html
[5] https://github.com/servo/servo/issues
[6] https://book.servo.org/contributing/getting-started.html#ai-...
Posting this using servoshell, the rudimentary browser UI based on the Servo engine which has made great strides lately and now is starting to become usable for general-purpose browsing.