Is the Web Dead? Just Ask Our Engineers

A long time ago, before the Second Coming of Steve Jobs, I found myself on a drive over the San Francisco Bay Bridge with Michael Young, now Redfin’s Chief Technology Officer. I asked him: “Isn’t some small part of you rooting for Apple?”

Having spent years porting server software for Windows, UNIX and Linux, Mike unhesitatingly said no. “Who wants to spend all day re-writing software so it runs on a Mac?”

With web applications overtaking both desktop programs for writing letters and big back-end systems for storing customer data, we thought we never would: most web software would just talk to other web software, not directly to computers themselves.

But almost every innovative company of the past two years, beginning with Foursquare, has started with a proprietary application, not a website. If a competitor to Redfin started today, it would give us a run for our money by building the best mobile applications for real estate.

All of this comes as a surprise to those of us who said goodbye to apps in 1999. Back then, if you had told me software companies would again be building device-specific applications  in lieu of websites, I wouldn’t have believed you. What’s really surprising: that we love doing it.

Apps wired directly to the hardware instead of running in a browser can make the device vibrate or take pictures, they can use the accelerometer or support gesture-based navigation. It’s fun to write software like that.

And this is the reason that I think we’ll continue to see proprietary applications. iPhone and Android apps may be, as Sasha Aickin observed in his excellent comparison of HTML 5 and native applications, an historical aberration, a temporary, Steve Jobs-induced distortion in the space-time continuum.

Or the web may be dead, as Chris Anderson just declared in his Wired cover story about the increase in apps that use the Internet but don’t run in a browser.

I agree with TechCrunch’s Erick Schonfeld that Chris under-estimates how wearying it is for software companies to support different devices, and how slowly online services will evolve — Redfin still doesn’t show school ratings in its iPhone app, let alone on Android or iPads – if they have to be re-written three or four times for every device.

But when I see how excited our engineers get when we talk about building mobile apps, it makes me think that the Apple era of proprietary apps won’t end any time soon.

Discussion

  • Michael Brandt

    Glenn, thank you for the post. This topic is really interesting to me.

    For every software engineer that loves programming for iPhone, there are 10 that hate having to support IE6. And the two platforms are quite similar – having proprietary standards and huge market share for their time. IE6 was once as viable a platform as iPhone is now (more so, actually: IE6 had 70% market share compared to iPhone OS's 30%).

    The idea of Android is different. Its competitive advantage for developers is that it's not proprietary – if you develop for Android you can reach many different phones on different carriers. Mobile websites made with HTML 5 are even less proprietary. Of course, as long as iPhone has Android beat on market share, and HTML 5 beat on tools and user interface, it makes sense for companies to develop for iPhone.

    But I bet I know what minority-share platform with unique standards that engineers in 5 years will be grumbling about having to support. It’s not IE6, but it’s close.

  • http://blog.redfin.com GlennKelman

    So you're saying that an iPhone will one day be like IE6? That's a better headline Michael than the one we have now…
    I agree that Android is an open operating system in a way that iPhone's operating system isn't, but it's still another port that Redfin would have to support, albeit for more devices.

  • Adam Wiener

    This topic is near and dear to my heart having worked in the bowels of Windows, Office, and SQL Server at Microsoft for 5 years before I joined the team here at Redfin where everything is native on multiple processor architectures and in a bazillion languages.

    However I totally agree with Sasha that if you really get much richer functionality by going native that might be just the ticket, multiple platforms and all! However as Glenn points out the “orthogonality problem” starts to bog down your organization. M features x N platforms where M and N are non-trivial means the % of time you spend on new features vs. porting and maintaining starts to shrink rapidly. And there are customers out there using those platforms who might be disappointed if you no longer supported their system….

    So what's an engineering organization that wants to stay nimble supposed to do? Well… you need to either limit your features or limit your platforms. Or both.

    The one advantage Redfin and phones have over IE6 (or almost anything from Microsoft) is that they are consumer devices. The expected lifetime of a handset is probably 2-3 years. IE6 has been in service for 10. So when you say iPhone 3 is no longer supported a year or two from now, Steve Jobs and AT&T have got your back. When you try to deprecate IE6 the IT shops around the world that form Microsoft's big iron customer base are railing against you (it's not Microsoft itself, they'd *love* to be rid of IE6).

    So while the proliferation of new devices means sometimes porting some of the same features over again, there's a compelling reason to upgrade. You get an accelerometer, a new form factor, a video camera, higher-res displays, or whatever cool things people come up with.

    One day all accelerometers will have a standard API and HTML version XYZ supported by every major browser will be able to access it directly. But by then it won't be cool anymore. Standardization is hard, boring, and time-consuming.

    The state of the art will always be native.

    • PeterisP

      I can't really agree to your opinion on support lifetime, like “The expected lifetime of a handset is probably 2-3 years. IE6 has been in service for 10.”

      Customers that are reluctant to perform a two-click software upgrade would be even more reluctant to shell out money for a hardware upgrade if the previous device still functions – and they last far longer than 2-3 years. If the devices are given for free by the phone operator, then it might work, but worldwide generally people are paying for iPhones.

      • http://blog.redfin.com GlennKelman

        Fortunately or unfortunately, you can't drop IE6 in the toilet, or leave it on an airplane the way you can an iPhone, so it tends to stick around…

  • http://blog.redfin.com GlennKelman

    Best line: “The state of the art will always be native.”

  • AndySousa

    First off, I am huge a Redfin fan and honestly believe that this is the future of real estate as we know it and the traditional realtor model will fall deep into the history books as that generation stops buying homes. I also am an iPhone user and have 5 on my account as the rest of my family uses them as well. With that being said I am writing to this blog on my iPad but unfortunately the rest of Redfin is barely usable. In your post you mention that a competitor that provided the best mobile apps would give Redfin a run for its money, well zip (not a great example) already has a mobile presence with iPhone, android, and iPad. I don't think these guys will bring you down because of it, but it is shocking to me that you are so reluctant to get an ipad and android branch out there. Who knows what the future will be for any of these platforms and even the Redfin model for that point, but bottom line is you need to program for today regardless of what we are told by the “experts” or what the standard is. Meeting customers needs to me should take precedence over ideologies of what the standards should be.

  • http://dannyman.toldme.com/ Daniel Howard

    Well, if the Printed magazine Wired thinks the Web is dead, then it must be true.