2021-03-24, updated: 2021-07-22

Tested with Nyxt 2 Pre-release 6.

Tags: press.

Radio Študent Interview

Radio Študent Interview

By John Mercouris

We recently had the pleasure of speaking with Radio Študent from Slovenia. You can find their website here: https://radiostudent.si/.

A brief description of who they are (from their website):

RADIO STUDENT is an independent medium. At Radio Student there are only a small number of full-time employees, which are joined by over 200 regular contributors: journalists, art and music critics, D.J.s, announcers, technicians etc. The majority of them are students and young people, and a number of them are highly respected intellectuals, journalists, artists and university lecturers.

You can see the full quote and more information here: https://radiostudent.si/info.

In our interview, we covered Nyxt, what makes it different, and what our vision is for the future. We also covered a wide range of interesting topics about the modern Internet, and the upcoming challenges we will likely face.

Half of the interview is in Slovenian, and the other half in English. Please enjoy :-)

Radio Student Interview

Direct link to MP3 file:

Radio Student Interview

Transcript provided below by Pedro Delfino:

1 - What is the Nyxt browser?

John: The Nyxt browser is an experiment to see how we can re-imagine the browser interface and what kind of tools we can develop to enable people to effectively utilize the Internet for the purposes of, you know, research and any type of knowledge work that they would be doing.

So, we do a lot of things in the context of the chrome and the browser. The chrome is the User Interface. And we also provide a lot of tools for analyzing the content of documents and traversing documents in intelligent ways. That's kind of a short summary. It is a browser designed to enable you to process the information on the Internet. And navigating it as efficiently as possible.

2 - Do you use Nyxt on a daily basis?

John: So, I actually do use Nyxt on a daily basis with one exception being Youtube. So, Youtube is kind of a pinpoint in Nyxt since it is a very heavy website. And our main engine right now, which is WebKit, does not handle it quite as well.

There are alternative websites you can use, like Invidious, which is a different front-end for Youtube. But I have not set those up on my machines. I do not know if Pierre's usage is just like that, but for me is my daily.

Pierre: Yeah, I use it as a default browser. For me, it has been extremely useful. I would only resort to other browsers for the very rare use cases in which Nyxt does not work. Like John mentioned Youtube.

Well, John did not mention it, but in the future, we would like to fix this incompatibility issue by adding another web renderer/web engine which is based on blink used by [Google] Chrome. Then, ultimately, I would like to give the choice… I mean, I would like to fix this incompatibility issue permanently by allowing users to use the engine they want. So, they would not have this tight dependency on the web renderer.

3 - Is Nyxt designed for productivity? Or customization? What is Nyxt optimized for?

John: Yes, so the first one aligns a lot with what I said earlier. The browser was designed to enable you to be productive. We built a whole bunch of little tools for that purpose.

After working with it for some real-time, we realized that the real value proposition that enables productivity is the customizability of Nyxt. And this means that literally every parameter of Nyxt can be changed by the end-user. You can even do that while the program is running.

So, that is why the slogan went from "Be productive" to "The Internet on your terms". Because this emphasizes the control that you are given and the ability to engage with the Internet how you want in a way that suits your work-flows.

We recognized that we cannot predict at every eventuality or potential on how people are going to need to interact with the content on the Internet. So, instead, we give them a framework so that they can build whatever interaction method that they need.

4 - Conkeror seems to be similar to Nyxt in some ways. How would you compare Nyxt with Conkeror?

John: So, I used Conkeror as well, a long time ago until the deprecation of the (inaudible). Well, I do not know how to pronounce it, I think it is pronounced as XULRunner (by Mozilla).

And it was great. It was a good tool. But the biggest problem was that it was tightly coupled to a specific platform. And we have noticed this has happened many times before. For example, many extensions programmed for Firefox and many WebKit (inaudible) web browsers ceased to work when WebKit 2 came out.

So, one of the huge differences in Nyxt is that we have gone through painstaking efforts to make sure we are not tightened to any specific platform. In the past, we demonstrated Nyxt running with both web-engine and WebKit. Even though these are completely distinct engines.

So, what is the difference between Conkeror and Nyxt? Well, resilience is a breakage. And number two, everything is written in Common Lisp instead of Javascript. So, you get all the features from the Common Lisp languages such as writing macros, redefining things on the fly, and dynamically adding attributes to classes. Javascript can do some of these things, but it cannot do all of these things and not with the same expressiveness. At least that is what we believe.

5 - What about extensions such as Vimperator, Vimium, Vimari, and Chromemacs that simulate Vim and Emacs in the popular browsers such as Firefox, Safari, and Chrome?

John: So, these plug-ins like vimperator, vimium, vimari… they are only part of the equation. They actually only usually cover link-hinting and navigation. And they do it in a rather simplified way. Since they are an extension, they are limited to being in Javascript more or less. And they run within a really specialized context. So, when you use Vimari or Vimium and you bring up the link hints, it is making a data structure that exposes in javascript and then it has a key listener, again in javascript, and you press a key and it navigates you to whatever link.

In Nyxt, what we do is that we grab the contents of the link hints. So we query the page, it gives us all the clickable elements, returning a list that is then handled in Common Lisp. So, we can do all sorts of algorithms from fuzzy matching on the URLs outside of javascript. And we actually multithread those. This is stuff that is simply not possible strictly in Javascript.

Now, it matters that we pull all this information to Lisp because we can transform it and integrate it to other components of Lisp. So, for example, with Vimari or Vimium or whatever, you do not have a javascript context. So you cannot say: "navigate to any link hinting in any tab".

But, in Nyxt you can do that. You can say: "show me all the links that exist in all of my tabs". It will produce a composable list of them and you could fuzzy search all of them at the same time. So there is a huge advantage to bringing everything to a lisp image instead of keeping it in Javascript for that ground.

6 - What is the target audience for Nyxt?

Pierre: So, the target audience is, well, a rather geek audience that we have. Mostly, tech-oriented users, power users, developers, and anyone who is really interested in collecting web content, analyzing web content. So, yeah, mostly tech-oriented user.

But then, in the future, we would like to make Nyxt a bit more accessible. Actually, ultimately, we would like to make Nyxt very accessible to the point that it could possibly be used by anyone.

Right now, for instance, we have a graphical customization interface. In the beginning, we could only customize it from the Lisp config file, which is really cool, right? Because you can program your browser. But is not really user-friendly. Now, we have a couple more of buttons and friendlier forward/backward buttons… You can also customize your default URL and start page with buttons. There is still a lot of work to do and ultimately we could have much more graphical customizations, much more menus, more mouse interactions, and so on. This is what we are aiming for.

7 - What about the web browser's competitive landscape? There are a lot of options, right?

John: Yeah, there are a lot of options as long as you want a browser that just has tabs on a screen, a URL bar at the top, and bookmarks as a list. If you want everything else, there is no option. All browsers available look the same.

It remembers me of a phrase from Henry Ford, who said: "You can have a model T in any color you want, as long as it is black". I really did not see any differentiation in the browser space. I also did not see any of them that were really focused on enabling the users to be more productive. There are attempts but none of them really hit the mark in my opinion.

Pierre: (inaudible) I would say that some browsers maybe they can satisfy the needs of tech-oriented users, especially when it comes, for instance, to Qutebrowser (inaudible some part) very good in that area.

I think they are focusing on the niche and maybe they would not try to go as far as empowering users with bigger data analysis features (inaudible).

8 - What is your view about the market/user share of web browsers? Is there an oligopoly or a monopoly? Is that a problem? Why?

Pierre: Maybe the first thing that is quite our view at the moment is the hegemony of Chrome, Google Chrome, the navigator. Well, not such a long time ago, I think Firefox was almost dominating in some countries, such as Germany. It was quite a tough battleground for web browsers.

But, today it seems that Google Chrome has won the battle. And I think we should really work against the monopoly here. We all remember the beginning of years the 2000s when there was a single browser, namely the internet explorer. Together with adobe flash, it was not exactly a pleasant experience for many users, especially the ones from non-windows platforms. They were having a hard time just browsing the internet which is crazy, right?

The internet should be fully accessible for everyone. So, I think we should (inaudible). I think it is important for us to not fall in there again.

Maybe something that John mentioned is that user interfaces in all browsers look the same. So, it feels that today the web experience is a very dumb down experience. It is a click experience and it feels like most users can only be passive on web browsing. This is sad, I would like to feel active when I am browsing the internet. That would be my opinion. What about you, John?

John: I think you have touched on a lot of great points. And one that I would like to head to is that many listeners could think to themselves: "yes, but the Google Chrome source code is available, it is an open-source project".

I would argue that Chrome is a "source available" project, not an open-source project. If you try to commit a change against the will of Google in the repository, you are going to find that this is impossible to do. And this represents a huge problem.

Google Chrome by having so much market share, so much power, can unilaterally create new web standards. If Google Chrome says: "Ok, we are supporting this feature, now", then people are going to start putting that on their websites. And if this is accepted by any committee-specification whatsoever, now, anyone who is not using the Google Chrome engine cannot use these websites anymore. And they can bake any sorts of non-sense they want. So this is a huge risk for the greater free internet. It is actually a really big problem but it gives you the illusion, the comfort of open. But is not really open source.

About the second thing that Pierre brought up, the [other] browsers treat their users as if they were incapable of doing anything but pointing and clicking. They do not give any control to improve things or to automate your work or analyze data on the internet.

They are really just… I do not know how to put it but is… What I would consider as just a dumb terminal, it just displays information. It does not understand any information.

Imagine if you are trying to consume books through photoshops or through (inaudible).. where you cannot search for text. So, you would need to open every page through the file open dialog. That is how it feels for me to browse the internet in another browser today.

9 - Question unknown

John: They are all related though, in the end. They are all problems that manifest themselves in this ecosystem due to the structure and design of modern web browsers. An (inaudible - guardian?) makes sense when your tool is so primitive

The UI of websites becomes more complex to handle the needs of the users because the browser interfaces are so infantile. What you have in your menu bar on your browser: quit, open, bookmark. You cannot even tell me what links exist on this page? You cannot even tell how my characters this page has?

It is nothing. You cannot do anything. Of course, yeah, that's why this (wall guardians [?]), these platforms, they get build up. It is a way for a tool to get built. There is a powerful opportunity there, so the tool gets built.

10 - Question unkown

John: One thing that I wanted to touch upon that Juri said, I think Google invests so much money on Chrome because there is an operational risk. Their core risk is the advertisement business. And once you got someone else calling the shots, you know, Mozilla committee foundation, something you do not control, they can make your business prohibitively difficult.

In my opinion, from what I can tell, google chrome is just a vehicle for Google to track you and learn about you without having to use cookies. You always sign in to your chrome account everywhere you have been, everything you have done, everything you clicked. So they make a profile for you and they do not need to use any third-party trackers.

And in fact, I think the evidence for this becomes more clear as Google proposes to do something with third-party cooking blocking. I cannot remember the exact details, but, basically, this would make it impossible for any advertisement entity besides Google to do this targeting advertisement.

So it is just all part of the bigger pieces… Somebody there saw some calculations and decided: "You know what? We are going to make a web browser!". And they went all-in. You know, they went all-in on it. They have this kind of twisted vision of the internet where people just use the internet via a Google Chrome book. Everything is done on the browser. I mean, that is absurd. It really makes no sense for me, whatsoever, honestly. I do not even how it got any traction. But here we are haha.

11 - Question unknown

John: Yeah, and you cannot even compete with them. They have got literally infinite money, they print money. For them is like… how can I say… it is not a direct contributor to their bottom line. They do not have to make money from chrome. They can develop chrome at a huge loss for as long as their want. And they just get the secondary benefit of the power. And there is no one way to compete with that. Like as I said, the death of that started with Internet Explorer who did exactly the same thing.

Microsoft said: " you know what, we (inaudible), we want our platform to support our protocols, our whatever". They want to embed active-X controls for whatever reason. They want you to have to use their platform. They can do it at a loss because they are making their money elsewhere. You cannot compete with them. That is kind of how it goes.

12 - About the death of the web browser

I think, actually, the death of the web browser came a little bit earlier. I think it came out when the first version of Internet Explorer came out. I am not trying to say that internet explorer was a terrible browser and that was the death of the internet. Rather, that it was free and it came with an Operating System. Suddenly, it was impossible for anyone to have a (business) model to compete with the web browser. Why would you pay to install a web browser if your computer already comes with a free browser?

That is huge friction in the way of adoption. And so that completely destroyed the business plan from Mozilla. In the interim, they kind of chug along with being supported by Google, having this default engine. That is all their make they money. Nothing has changed.

In the past couple of years, you have seen them doing a whole bunch of experiments to redeem themselves, a VPN service, a Mail service… A whole bunch of things that are completely orthogonal to web browsing. And without much success. That is what I think is the reason for their layoffs. They invest a bunch of money and they do not get anything from it. So, now they are trying to cut their loss. Fortunately, that means a death of a company usually. But, (inaudible).

13 - About the web engine

Pierre: The current engine used by Nyxt is WebKit. More specifically, WebKit gi-gtk. (inaudible). And, well, we have a second engine which we partially support, but the implementations are still in their way. It is a Qt Web Engine which is a fork of blink used by Google Chrome.

We started for WebKit for a couple of reasons. One is that we try to do something different and not fall for the monopoly of web renderers. And the second reason is because WebKit gi-gtk is actually quite pleasant to work with in terms of API. It is highly customizable, they export a lot of (inaudible)… A huge interface that gives a lot of control over what you can do with the rendering. So this is perfect for us because we want to give as much customizability as possible for our users.

Now, to the second renderer Qt web engine. Well, as we mentioned before, we also want to provide good compatibility for our users in the sense that they can, if one website does not work with web gi-gtk, they can always use the other web browser to surf the web. So it is a fallback anyway.

John: Yeah, you can have multiples renderers simultaneously. It is not the easiest thing to do but it is possible, yeah.

Pierre: There could be some limitations, though. We will see it. We still have to work on it. We have not tested everything.

14 - A browser engine programmed in Lisp?

John: There is. It is called Closure and we put it in our repositories (with the "s", "j" is for the programming language). Why did we choose Common Lisp?

Well, actually, Pierre and I debated quite a bit at the very beginning of the project. We thought about Racket and Guile.

But, ultimately, Common Lisp standout, because is the most sort of "batteries-included lisp". It has got a lot of great libraries. It is very stable. The spec does not change. There is almost zero (inaudible).

Anything that you write today is going to be useful for 30/40 years ahead. If you contrast that with javascript, you are going to have a lot of trouble running something written 5 years ago. Which is… I do not know… Is not a good way of running software to build an ecosystem in my opinion.

15 - About other protocols

Pierre: We have been asking about the (inaudible) for protocol before. Like, there are other protocols IPFS, and this one I am a bit more familiar with. So, it is definitely on our roadmap to improve support for these different protocols. And, for instance, IPFS is still supported by maybe 1 or 2 browsers at the moment. It is a bit niche. it would be cool to add support there. It would also make Nyxt a bit more attractive to people looking for more exotic protocol.

John: And there are other protocols too. Like Tor protocol, FTP protocol. I would like to add support for things like that. So that you do not have to pull another kind of program to interact with this kind of thing.

I remember that in the past, 10 or 15 years ago, browsers used to have facilities for this kind of operation. I remember the first Konqueror, with the K, where you could manipulate directories with SSH, SSHFS, or FTP whatever… It was amazing how it loved to create that exact same experience that Nyxt can interact with. It can mount FTP, or whatever, and it is exactly as using (inaudible). That would be really cool. But we are not that yet.

16 - How is the project currently funded?

John: So there is this European Union program called "Next Generation Internet Initiative" and what they do is that they disperse the funding with organizations they interact with. We have been interacting with NLNet which is an organization founded in the Netherlands to support, well, this kind of things, such as open-source projects - you can visit their websites. And then we are also in the NGI0 program supported by the funding box. And they are located in Poland.

So, basically, we set-up some milestones of things we want to achieve and then we go through, and we kind of justify how this supports the EU initiative, how it makes sense, and how it is actually an appropriate grant. And after we go to the whole approval process, we complete the milestones, verify that the milestones are completed then we get basically the dispersed funding. And this covers basically our food.

17 - How did the project start?

John: So, when Pierre contacted me, we talked about, OK, we really love writing this software, but in order to really do that, without doing something else on the side, you have to become a business, right?

Otherwise, you do have the day job and then you have got the side project. I know that I personally did not want to split my time that way. I have been doing side projects my whole life and I know it is exhausting. You have two things in-between in your head, you are working within two different environments and you just wanna to do one thing, and you want to do it well.

So we founded a company and we thought about how we are going to financially support ourselves in this process? So, initially, we did some contracting thing. I went out and I found some contracts and we did those while working on the side on Nyxt.

But it ended being a lot of the same thing. Pierre and I had normal jobs, and then we worked on Nyxt on the side because the contracting was taking a huge portion of the time. And then we started to think about how can we make Nyxt directly profitable. How can we make it such that our work directly aligns with becoming sustainable so that we can keep working on it and improving it to provide something for the community?

And so just by great luck Pierre. Well, it is not great luck. It is obviously a bunch of work. Pierre is in a bunch of mail-lists, so somehow and some way he got contacted and heard about the NLNet grant. So we applied, thinking "OK, maybe we will get the grant". You know, this was kind of what we need to survive otherwise we are going to work this on the side. Luckily, we got it. We were very lucky in that respect, you know. We are very grateful for that.

And then we kept working, working, working through that grant. We gotta figure something out, we gotta become (inaudible). So by the end of the grant, the NLNet people contacted saying: "hey, as we are winding down, we have some further people who can fund you". And we thought that was amazing, but what are the chances that we would be accepted again?

As you can imagine, you need a lot of runaways to turn a browser into a sustainable product. Because to start a browser with a minimal viable browser, requires years of effort. So, we worked, worked, worked, and, now, we are really close to our 2.0 release.

After our 2.0 release, we are going to have a pretty clearly settled API that we are going to slowly build of (inaudible). And this brings us to ours modernization plan and how Nyxt fits in with Atlas (the company).

Atlas is just the legal entity that represents Pierre, me, and whoever is working for us. What Atlas will do is it will take Nyxt hopefully, pending acceptance from the community with the users' votings, and Atlas is going to do applications on top of Nyxt.

Emacs, for example, has a bunch of applications that you might think of. Rest clients is an application for emacs. It is like a separate program, almost.

We are not gonna cripple or limit or inhibit any way nyxt artificially. But we are going to do is build value on top of Nyxt in these applications and hopefully, the people like them, and this can turn into a sustainable business that is not relying just on research grants to get by. Because our ultimate goal is to be able to produce something for everybody for the benefit of everybody in a sustainable way such that we can do the best work that we wanna do. That is ultimately the only way.

If you are working on something on the side, you really do not have enough hours on the day, honestly. Imagine if buying (inaudible) only worked on weekends. Where would it be today?

It would be nothing compared to what it is now. So we see it as a necessary step in terms of making an important project that we think will have a good impact on the world.

18 - Anything else you would like to say?

John: There is something that I said at the beginning but that I would like to emphasize again. Nyxt is specifically a tool to empower Internet users and that is why we say to "use the Internet on your terms".

The goal is to make something that is customizable for everyone because the Internet is a huge, absolutely huge library, the biggest library of human knowledge we have ever seen. And we think that current browsers inhibit access to this knowledge base. It is like as if everyone has to read a book from a microscope and we want to get rid of the microscope and just send the book down from the person.

Here you go…Here is… Do you want to turn a page with a fan? Go for it. Do you want to read with green lights? Fine! Here is the book. Here is the information. Just take what you need. We hope that this will carry on research and all sorts of knowledge workers to just do better work.