How to design a perfectly usable browser



  • The overwhelming majority of browsers right now (including Vivaldi) is a Chromium skin. The rendering engine is the same, so the difference that makes a conscious user pick one over another is the interface design, accessibility, configurability, or some very special extra feature, e.g. to handle proxies or cookies or webpage font/color modifications or whatever in a convenient or superior manner.

    Unconscious users use whatever their computer comes with or whatever the web pushes on the computer (remember that Chrome 1.0 arrived as sneakware embedded in Java updates for Windows), but users with their eyes more open have already, prior to Vivaldi, found what they need. If Vivaldi wants them onboard, Vivaldi must already be roughly as good as those other options, by having the corresponding perspective to begin with.

    Some concrete alternative examples I have in mind are Qutebrowser and Otter. Qutebrowser is a toolbarless keyboard-driven browser where every action can be bound to a keyboard shortcut and rebound on the fly in a command-line-like interface that displays helpful hints wherever possible. The missing feature (for now) is ability to edit menus.

    Otter comes with configurable toolbars and has a dialogue for rebinding keyboard shotcuts and, again, every action can be bound and rebound as a matter of general principle, just like in old Opera. And in Otter you can rewrite menus too, even though it's via the complicated and error-prone JSON format.

    The developers of those browsers took Opera's orientation to accessibility and UI and have begun to implement it methodically. This makes them usable, not by virtue of taking hints from old Opera, but because old Opera embodied plenty of common sense.



  • What's your point really? Are you indicating Vivaldi isn't usable? I just tried out Otter and it's still a release candidate after all these years. The icons are incredibly ugly and have wrong shading. When I tried to move the tab bar, the webpage currently open completely scrambled and was unreadable. If we are talking usability, Otter browser is completely unusable. I'd rather run Firefox or Chrome or Safari, or whatever. This project is doomed to fail, and I'd imagine the same goes for Qutebrowser. More power to them if they have the right mindset and want to build a browser the right way, but that's not how you build a functioning browser. Oh well, maybe they can get there in the 20s, if the classic desktop and its operating system is still a thing by then...



  • @luetage Perhaps idiosyncratically, by "usability" I first and foremost mean configurability to the very core. Displaying websites is trivial, that's what the rendering engine does, and since pretty much every GUI browser has the same rendering engine, then the basis for choosing one browser over another is not the rendering engine, is it? Instead, I look for an easy way to set up my custom keybinds, custom menus, custom css, etc.

    Vivaldi right now has some select stuff under Preferences/Keyboard, but why only some select stuff? Is there a good reason to withhold anything, such as "copy page address" under Preferences/Keyboard? Then there are the "Page actions" to filter colours in webpages (an excellent thing to include as far as it goes), but why don't the settings work consistently on every page and why don't they stick across restarts? It used to work in old Opera, so this is not something impossible or difficult. It's something we have been deprived of.

    I agree that Otter is not progressing as one (myself, really) would hope. However, the developer's mindset and aims are instantly discernible when you open it. At least it has personality out of the box and it is a reasonable option for the main browser if you use KDE (because there it has no issues with themes). Qutebrowser has progressed much better and is documented much better too. Qutebrowser has been where I want it for over a year now, feature-complete without any serious reservations. Warmly recommended. They are not doomed projects. They already have their niche and user base. They are doomed only if you mean that they will last as long as the enthusiasm of the main developer lasts, but this applies to everything, really.

    With Vivaldi, I don't see much point with it when I open it. Yes, it's way more configurable than Chrome and provides good competition to the modern Opera, but those are (very) low bars to cross, seriously. I am okay with browsers whose default interface is passable, but isn't pretty much every browser somewhat passable? When I need something slightly different from the default interface and the application does not have the configurability, then the interface is more of an obstacle instead of being an opportunity.


  • Moderator

    @ersi Perfection is unattainable, as no two people have the same standards.

    I think you assume compatibility is automatic. It is not. It requires a lot of coding to make a "Chromium" browser that does not use the Chromium UI just work. And that work has to be revised weekly, as Chromium changes.

    I think you assume configurability is easy to achieve. It is not. Job 1: Make the browser work (requiring extensive weekly work to keep up with changes in Chromium) Job 2: Make the browser flexible and configurable, providing as many options as humanly possible. This is far and away the more difficult job, given that the underlying engine is in a state of perpetual change. First, options are hard to code, not easy. Second, the shifting platform underneath breaks things that were working yesterday, ALL THE TIME. Hence, to code an option is to commit to patching its failures on a continuous basis. So you make an option, it works on today's Chromium, tomorrow's Chromium breaks it, and you're right back to creating it again.

    All that said, the ultimate goal of Vivaldi is maximum compatibility, flexibility and configurability. No other browser has such a goal, so Vivaldi will always be ahead in that race - but it will always be impeded by the fact that no one will test in or code to a minority browser, and therefor it will always be forced to re-code, revise, and patch to account for the continually shifting ground underfoot, of an engine that predominantly caters to Chrome and literally does not care at all, that Vivaldi exists.



  • @ayespy said in How to design a perfectly usable browser:

    @ersi Perfection is unattainable, as no two people have the same standards.

    Which is why I don't ask for perfection. I ask for a methodical implementation of a few openly declared principles. It makes sense for Vivaldi to aim at something like old Opera was, no?

    I think you assume compatibility is automatic. It is not. It requires a lot of coding to make a "Chromium" browser that does not use the Chromium UI just work. And that work has to be revised weekly, as Chromium changes.

    I think you assume configurability is easy to achieve. It is not. Job 1: Make the browser work (requiring extensive weekly work to keep up with changes in Chromium) Job 2: Make the browser flexible and configurable, providing as many options as humanly possible. This is far and away the more difficult job, given that the underlying engine is in a state of perpetual change. First, options are hard to code, not easy. Second, the shifting platform underneath breaks things that were working yesterday, ALL THE TIME. Hence, to code an option is to commit to patching its failures on a continuous basis. So you make an option, it works on today's Chromium, tomorrow's Chromium breaks it, and you're right back to creating it again.

    Sure, it's hard. It's tough. But, I wonder, when some one-man projects can pull it off for several years as a side-job, then why not a commercial entreprise as their main job?

    All that said, the ultimate goal of Vivaldi is maximum compatibility, flexibility and configurability.

    Can you link me to a written mission statement? Even better - a roadmap? I'd like to see if those words are there and if they appear in a context that gives them specificity.

    No other browser has such a goal, so Vivaldi will always be ahead in that race

    Maybe, if you talk about commercial browsers, even though Vivaldi is, for me, disappointingly marginally ahead the current Opera. Among open&free projects I know some that are clearly far ahead in terms of at least configurability, and I'd say also in other areas, once we get down to defining and quantifying other areas.


  • Moderator

    @ersi Does this help?

    https://www.sitepoint.com/interview-vivaldi-ceo-jon-s-von-tetzchner/

    or this?

    https://vivaldi.com/story/?lang=en_US

    And of course we are only marginally ahead of Opera. They had a huge head start, tons of money, and ten times the size of development staff.



  • After perfection there can only be decay.
    I have installed and used practically any browser, some with interesting functions and good speed, others with several cores, such as the trimotor Avant, of course Opera and even this new thing that has been released, Opera Neon. very nice but I do not see much sense, apart from aesthetics, seeing it more as a concept than as a useful navigator.
    A lot of times I've also used Chrome, although lately it has not convinced me so much, other chromiums, like Maxthon, which is not bad and Yandex, also UR that announces itself for its privacy, which is not as much as they say, Firefox naturally, very valid although somewhat slow and in the end Vivaldi that in my opinion more is about this functionality I was looking for.
    Without a doubt, it also has its cons, like everyone else, but it sure is on the right track to become a reference. All together, aesthetics, functionality and speed, apart from being customizable like no other.


Log in to reply
 

Looks like your connection to Vivaldi Forum was lost, please wait while we try to reconnect.