I am also running an insider build of the OS as I want to be on the bleeding edge. My own browsers is a developer preview. This all taught me quite a few things.
- There is a huge difference between a browser and a webview
- It is possible to build all kind of fancy interfaces in web technologies. But often there are hard to work around security, performance and accessibility issues.
- Day-to-day browsing sucks when your browser is a moving target
The main thing it taught me that as developers, we have a massively skewed view of what a browser is. Or, more accurately, what matters to users.
A browser is a web display tool
Photo by Kevin Ku
In my meetings with clients, I found a few things that mattered most to them:
- Performance – a browser should start up fast, not bog down the operating system eating resources and stay responsive. This is not only about first use, but often a big need for long-term use – for example in a kiosk scenario. Browser restarts are uncommon and should not be necessary
- A reliable content display – a browser should also render content instantly. Content could be HTML documents with all the cruft they come with these days. But it is increasingly important that media of all kind plays smoothly. And often there is no chance to rely on a plugin or third-party player. This means a browser should be a video/audio player. It also should be an ePub and PDF reader. And it needs to render huge images and WebGL/WebVR content without any hiccups.
Neither of those should come as a surprise – this is also what we as developers expect. As developers we also want more. We put a lot of stake into the developer tools experience. And we want browsers to support the newest and hottest experimental technology. And we complain to other browser makers if not all are on par with the others on both these counts.
Most clients that wanted a bespoke browser, however, had other priorities.
A browser is an enabler
Photo by rebeck96
I was happy to hear that every customer I worked with put a high emphasis on accessibility. I was less happy to see how tough it still can be to make a zoomable interface in a WebView. And how to get keyboard access right. And how to enable screenreader access and text to speech interfaces. But that’s my job as a developer. And yours, too.
A browser needs to enable people of all different abilities to reach what they came for. And ability isn’t a fixed state but fluctuates with environment and external influences. I found that out lately when I had to zoom my browser to 110% to read the average web site. A reminder that I am the age I look. I fixed it later by switching to varifocals, but I am happy that browsers allow me to fix it on the spot. Often a WebView embedded in an App doesn’t allow that.
A browser is a guardian
Photo by ShonEjai
A very common demand is to have blocking of unwanted content in the browser. This ranges from must haves like malicious site filters to popup blocking up to providing a pre-approved whitelist and block all else. Ad-blocking is also often a demand, but the big worry is that an opened document could give an attacker access to the computer.
Browser creators have a shared list of known attack sites to filter out. This is a great service that protected quite a few people over the last years.
Displaying insecure content in a secure site is another problem that I needed to work with. Whilst browsers have UI interfaces and blocking built-in some WebViews allow mixed content. That’s not good.
People also wanted to ensure that there is no dependency on plugins for video display or to outright ban Flash content. PDF was a big concern, too. And they were right to do so. Browsers do a lot to avoid malformed PDFs to spy on us.
Speaking of spying, a private browsing mode or a general browsing experience that blocks trackers was also high on the list. This is not surprising and a sensible feature to offer.
The opposite of a private mode is, however, also something that we keep underestimating.
A browser is a memory keeper
Photo by Rachel Demsick
Browsers are there for us to display the web, but they are much more importantly there to make this easier for us the more we use them. I felt that in my own use of an unreliable (preview) browser in a constantly changing configuration. I realised just how much it sucks having to constantly re-enter passwords. How annoying it is not to have sites I keep visiting autocomplete when I type a few letters of the URL. How “share with X” buttons become useless when you are not logged in. In essence, how much I rely on the browser to remember what I’ve done and intelligently help me along the way.
We don’t give browser makers enough credit for this amazing experience, as it is – like all good UX – invisible. A browser that automatically and safely stores and enters form data for me is a great help. It even protects me from entering wrong information and the frustrating re-filling experience. Yes, of course this is also an attack vector, but that doesn’t mean we shouldn’t celebrate its usefulness – we should make it more secure instead.
Having Firefox’s “awesomebar” remember my surfing habits over the last 10 years spoiled me. I don’t even remember the last time I bookmarked anything. My important finds I tweeted out anyway and let Pinboard remember them for me.
Others may as well have a lot of bookmarks. Together with synced autofill content and history it is great to be able to take this experience with you. Independent of computer, phone or device.
When it comes to answering what a browser does for us, I learned a lot in the last few months. I am more humble when demanding features of my colleagues and friends in other companies working on browsers. They are doing an incredible job building a piece of software that displays the web in an accessible and secure way. And once you learn about attack vectors of the web and webviews you are in for some rough nights.
I also learned that when it comes to who wins the most users in the browser space, it comes down to how they handle user data. Technical ability of a platform and “pushing the envelope of what the web can do” is exciting for us as developers. End users, however want the web to be easy, fast and secure to use. They also want their browser to know where they’ve been, and choose when it shouldn’t. They want a few things automated and get easy access to their content to send out to the web. From a privacy point of view, I also feel uneasy about this, but those who make this the easiest, win.
When it comes to what browsers can do, they’re much more on par than they’ve ever been – thanks to standards. The big differentiator isn’t what tech to support, it is how much of a travel companion on the web the browser is. It is a game of data, insights and good UX of the browser interface itself.
I am much more excited about what we can do in the future to help people keep their online identity in a browser without leaking it. And I would love to see more interfaces that teach people the value of their own data whilst using a browser.
A browser is not merely a display mechanism for the web. It is a record of your identity and history on it. This can be a great thing if done well, or a scary one if not. I’m worried that too many people on the web leave breadcrumbs without knowing it. They see the convenience of that without knowing the consequences or value to others.
I invite any other web developer to look closer at the interface elements and user-centric features of browsers. There is a lot of beauty and very clever, researched-based interfaces to be found. And I’d love to see this get more coverage in our circles, and not only which new API or tech feature is the next big thing.
View full post on Christian Heilmann