Break from the future: Hold the new stuff and fix the web first

Hiatus is not a dirty word. Crevice is a dirty word, but hiatus isn't

These interwebs over here don't work at all

While Archibald’s argument is a good one, he glosses over the fact that often the web's version of features on native platforms are half-baked and have been for ages. Consider the Geolocation API, possibly the most useful HTML API out there. A mobile device in your hand can let a website know where you are and tailor information you might need – maps, nearby friends, the closest restroom and so on – to where you are. And of course it comes with the broader reach of the web, no building multiple apps for different platforms. Write once, run on the web. It sounds perfect.

You might think a company like Uber would jump on the Geolocation API. But Uber's service is currently only available through an app.

Opera's Bruce Lawson, writing in response to Koch says that's understandable: "Because sometimes the Geolocation API is not very accurate on the web, and really accurate location info is critical to a taxi-hailing service."

Lawson says that's no reason to put browser features on hold, writing: "That’s an argument for making the Geolocation API better, rather than stopping development."

Indeed it is not so much that the web is moving too fast, but that there has been seemingly no effort whatsoever to fix things that have long been broken. The Geolocation API has been available in most browsers for many years now and it hasn't become any less buggy or unreliable.

It's not the pace of feature development that is the problem, it's the quality. But development resources are finite and if browsers were to stop creating new features for a while and focus those finite resources on documenting, debugging and fixing those half-baked features like Geolocation support instead, wouldn't that in fact be pushing the web forward?

Forget the more complicated APIs even, what if developers could do something as seemingly basic as reliably style form elements across browsers? Wouldn't that be pushing the web forward?

This is the core of Koch's argument, not that the web doesn't need new features for developers and users, but that new features at the expense of effort spent improving what we have quickly becomes pointless. Broken or half-complete emulations of native features can be worse than no features at all.

It seems reasonable to argue that what the web needs is not a year-long moratorium on new features – which, let's face it, will never happen – but something akin to what Ubuntu Linux developers call "paper cut bug" fixes.

A paper cut bug is defined as "a trivially fixable usability bug that the average user would encounter on his/her first day of using a brand new installation". In other words, a bug that isn't difficult to fix, but causes real problems during day-to-day use.

Perhaps, since the web is currently semi-broken, it would be wise to throw out the "trivial" criteria and take some time to start fixing things.

If that means new features are delayed by a few months, will that mean the end of the web as we know it? Probably not. In fact, it might mean the beginning of the web as we want it. ®

Similar topics


Send us news

Other stories you might like