By John Gruber
Tara AI — Build better software, faster
When you publish your opinions on a regular basis, it’s hard to resist the urge to gloat after you’ve been proven correct. (For example, I’ll be I-told-you-so-ing with regard to the iPod mini for the next couple of years.)
It’s also rather easy to ignore the times when you’ve been proven wrong.
It’s a good thing I wasn’t publishing Daring Fireball back in the mid-to-late ’90s, because if I had been, I’d currently be eating an awful lot of crow with regard to what I would have written about the web’s potential as an application platform.
At that time, at the peak of the Netscape-Microsoft browser war, the conventional wisdom was that the web was the future of application development. The technology certainly didn’t yet exist, but the idea was that Netscape’s web browser posed a serious threat to Microsoft’s Windows monopoly — that at some point in the future, user applications would be written to run within the browser.
Thus, Microsoft’s incredible change of course, going from more or less ignoring the Internet to completely dominating it within the course of a few years. The idea was that Microsoft killed Netscape because Microsoft saw them as a threat to Windows.
Me, however, I just didn’t buy it. I completely saw the potential of the Web as a publishing medium, but I just didn’t see how the Web was ever going to serve as a high-quality application development environment. The way I saw it, Microsoft killed Netscape not because it was a threat to Windows, but simply because they (Microsoft) wanted control over this new publishing medium.
I simply couldn’t have been more wrong. The conventional wisdom was in fact correct — the web has turned into a popular application development environment. Where I’d gone wrong was in getting hung up on the idea of it needing to be high-quality before it could become popular.
I was thinking in terms of the apps that I used every day, circa 1996: BBEdit, QuarkXPress, Photoshop, Eudora. There was simply no way that a “web app” could ever provide the same quality experience as the “real” apps I was already using. And I was right about that — the user experience of any app running in a web browser is crippled.
What I’d overlooked is that most people don’t use advanced text editors or desktop publishing software; and more importantly, most people simply don’t care about the quality of an app’s user experience. Not at all. They just want it to work, and to be “easy”.
My saying that web apps would never become popular was like a theater critic in the early 1950s dismissing television.
The user experience limitations of a web app are glaringly obvious. They simply don’t look or act like normal desktop apps. The browser in which they’re running — that’s a normal app. But the web apps running within the browser aren’t. They don’t have menu bars or keyboard shortcuts. (The browser itself does.) This isn’t about being “Mac-like” — it applies equally to Windows and open source desktop platforms. Instead of looking and feeling like real Mac/Windows/Linux desktop apps, web apps look and feel like web pages.
The persnickety little UI details I obsess over — these are nothing compared to the massive deficiencies of even the best web app. But most people don’t care, because web apps are just so damned easy to use. What’s interesting is that web apps are “easy” despite their glaring user experience limitations.
What they’ve got going for them in the ease-of-use department is that they don’t need to be installed, and they free you from worrying about where and how your data is stored. Exhibit A: web-based email apps. In terms of features, especially comfort features such as a polished UI, drag-and-drop, and a rich set of keyboard shortcuts, web-based email clients just can’t compare to desktop email clients.
With web-based email, you can get your email from any browser on any computer on the Internet. “Installation” consists of typing a URL into the browser’s location field. The location field is the new command line.
[appear] to have been designed by vi users (
kmoves up, and we are expected to memorize multi-key sequences for navigation).
Gmail’s threading and searching are indeed nice, but its overall look-and-feel is far inferior to that of a real desktop mail client. What it has going for it is what all webmail apps have — zero installation, zero maintenance, access from any computer, anywhere (including from work, a major factor for personal email). Gmail is simply better than the other major web-based mail apps; but Yahoo and Hotmail and the others are still ragingly popular.
What I missed when I dismissed them a decade ago is that web apps don’t need to beat desktop apps on the same terms. What’s happened is that they’re beating them on an entirely different set of terms. It’s all about the fact that you just type the URL and there’s your email.
What got me thinking about this was Joel Spolsky’s “How Microsoft Lost the API War”, a terrific essay published last week. The gist of Spolsky’s argument is that Microsoft’s crown jewel is the Win32 API — the set of programming interfaces that developers use to write desktop Windows software — and that web app development is gaining momentum, at the direct expense of Win32 development.
The reason the Win32 API is so important to Microsoft’s Windows monopoly is dependence: if your company relies on Win32 software, then it also relies on Windows. And conversely, as a developer, writing against the Win32 APIs allows your software to run on over 90 percent of the computers in the world. That’s the cycle that built a $50 billion pile of cash — customers use Windows because that’s where the software is, and developers write Windows software because that’s where the customers are.
Switching to, say, Mac OS X is an expensive proposition for a large corporation. Not only do you need all-new hardware, but you also need all-new software. And we’re not just talking about buying new licenses — for large corporations, we’re also talking about custom apps written in-house (what do you think all those Visual Basic developers have been writing all these years?).
Switching to open source desktops — KDE or Gnome or what have you — is also expensive. No, you don’t need new hardware, but you still run into the same situation with regard to software. (Yes, I know — you can run Win32 apps on Linux using the WINE Win32 emulator, or with Virtual PC for Macs, but these are second-class Win32 environments. I’m not saying it can’t be done, just that it’s unappealing.)
Switching to web applications, however — well, that’s different. It can be done gradually, because you can switch one app at a time while still running Windows, and thus, still running all your other Win32 software.
It’s not so much that switching to web apps is cheap, as that it’s easy. In fact, in many ways, switching your employees to web apps is even easier than upgrading the Win32 apps they’re already using. I.e. it’s easier for corporations to migrate to web apps than it is for them to stay Windows-only.
Web apps are easier to deploy. No need to install software on each client machine; there’s just one instance of the app, on a web server. Every user gets the latest version of the software, automatically.
Custom web apps are easier to develop than custom desktop apps. That’s not to say it’s easy to make a web app that looks and feels like a desktop app — that’s not really even possible. But it’s easy to write a web app that looks and feels like a web page, which is apparently good enough for most purposes, especially data-entry and data-retrieval apps that tie into server-hosted SQL databases.
And if you think the 90-percent market share of computers that can run Win32 software is huge — how many computers do you think run a typical web app?
Most email web apps (e.g. Gmail and Yahoo Mail) run on any computer with IE, Safari, or any Mozilla-derived browser. Most weblog web apps (e.g. Blogger, Movable Type, WordPress, and Textpattern) run in every browser I’ve ever tried. These apps are effectively usable from any Internet-connected computer in the world.
I’ve been thinking about the rise of the web as an application platform for a while. But what hadn’t occurred to me until I read Spolsky’s essay last week is this, which I think is quite remarkable: Microsoft totally fucked up when they took aim at Netscape. It wasn’t Netscape that was a threat to Windows as an application platform, it was the web itself.
They spent all that time, money, and development effort on IE, building a browser monopoly and crushing Netscape — but to what avail? Here we are, and the web is still gaining developer mindshare at the expense of Win32.
There are certainly exceptions — banking sites come to mind — but for the most part, web apps are being built to run in any modern browser, not just IE.
I think Spolsky is very much correct that Microsoft is losing the API war. But what’s ironic is that they’re losing this war despite the fact that they won the browser war. Winning the browser war — destroying Netscape — was supposed to prevent there ever even being an API war.