Almost a great idea

Here's an example of a good idea gone wrong.  I saw a link for a web site that exists solely to advance openness in government.  How cool is that?

U.S.
Image via Wikipedia

I clicked around for a bit, eventually reaching a place where I was supposed to be able to submit an idea for government, where it would (presumably) be viewed and discussed among my peers.  When I clicked the button to enter my idea, I was prompted to log in with OpenID (again, very cool).  I logged in and clicked the button again, and was rewarded with the following barfage:

500 Servlet Exception

[show] java.lang.NullPointerException

java.lang.NullPointerException
	at _jsp._jsp._includes._build_0header__jsp._jspService(jsp/includes/build_header.jsp:37)
	at com.caucho.jsp.JavaPage.service(JavaPage.java:61)
	at com.caucho.jsp.Page.pageservice(Page.java:578)
	at com.caucho.server.dispatch.PageFilterChain.doFilter(PageFilterChain.java:195)
	at com.caucho.server.webapp.DispatchFilterChain.doFilter(DispatchFilterChain.java:97)
	at com.caucho.server.dispatch.ServletInvocation.service(ServletInvocation.java:266)
	at com.caucho.server.webapp.RequestDispatcherImpl.include(RequestDispatcherImpl.java:494)
	at com.caucho.server.webapp.RequestDispatcherImpl.include(RequestDispatcherImpl.java:358)
	at com.caucho.jsp.PageContextImpl.include(PageContextImpl.java:1008)
	at _jsp._jsp._includes._autoselect_0header__jsp._jspService(jsp/includes/autoselect_header.jsp:23)
	at com.caucho.jsp.JavaPage.service(JavaPage.java:61)....

Close, guys. Very close!

Please don't do this to your customers, okay?

Reblog this post [with Zemanta]

More on “HTML is dead”

Microsoft's S. Somasegar ("Soma"), who heads the Developer Division, posted on his blog yesterday about "Key Software Development Trends".

I was pleased to see him include "Proliferation of Devices" among the top trends in development, but there was obviously an acute case of tunnel vision at work here, because Soma completely neglected all the non-Microsoft devices that people seem to insist on using.  I know, I know -- you can't have a Microsoft VP publicly acknowledge Apple on a corporate blog; that's just craziness, but the rest of us sure can.

Windows v0.0
Image by . SantiMB . via Flickr

As I wrote in my last post, the proliferation of devices -- especially across platforms -- has the potential to impact me in a pretty fundamental way.  It wasn't long ago that I could develop a web application and have a reasonable expectation that most clients could use it, regardless of their client architecture.  Differences existed among browsers, to be sure, but there were relatively few "can't get there from here" moments.

Today, I see a fractured web.  Technologies like SilverLight have the potential to radically improve both the user experience and development productivity, but these benefits are hugely marginalized when they only run on some of the platforms that our customers use.  Today, Microsoft's technology leadership position is challenged on every front:
  • The desktop.  Obviously, this is where Microsoft lives.  Between Windows and Office, this is not only where Microsoft asserts its most solid dominance, it's also where they make most of their money.  In terms of technology adoption, though, they're facing a 1-2 punch: they've struggled to get people to upgrade Windows, and they've struggled to get people to upgrade IE6.  The Windows upgrade problem also impacts SilverLight availability, since most people get SilverLight with an operating system upgrade.  According to Riastats.com, SilverLight's current penetration is just above 50%, which is a good start, but not yet where it needs to be.  Adding insult to injury, more people are running Mac OS every day, and Linux continues to grab scraps of market share, too.
  • Internet Explorer.  When people leave IE6, there's a good chance they're going to FireFox or Chrome instead of to IE8.  Microsoft's once near-monopoly in browsers has been eroding over time.  It's now imperative that any UI technology that runs on a browser must run on most, if not all of them.
  • Windows Mobile.  The delay in getting Windows Mobile 7 out the door has absolutely killed Microsoft here.  They got caught by the iPhone in much the same way they were caught by the internet, and we've yet to see a credible response.  At this point, even Mobile 7 is lights-out fantastic, it's got a pretty huge uphill battle to gain relevance, let alone dominance.
  • XBox.  The XBox is a real success story at Microsoft, but again, it's not dominant, with a market share somewhere around 25% of all gaming consoles.

There's no question that what Microsoft really wants is for us to develop software with Microsoft tools on a PC running Windows, and then to distribute this software to consumers who are also running on a Microsoft-powered device of some sort.  When Microsoft held near-monopoly positions on every platform where computing reasonably occurred, this was tenable.  Today, though, it's just not reasonable.  Customers are computing on a dizzying array of devices, and not all of them are powered by Microsoft.

What's needed today from Microsoft is real development leadership.  Give us a runtime that works everywhere, and Visual Studio becomes an absolute no-brainer choice for development.  While this might seem like a daunting task, a good part of this work is already being done.  Miguel de Icaza's efforts with Mono have paid incredible dividends to-date.  Today, because of Mono, you can:

Microsoft, please jump on this bandwagon!  Visual Studio is so clearly superior to other development environments that its only current threat is developers hemorrhaging to develop for other platforms.  Making Visual Studio a true cross-platform tool could make that argument a non-starter.

Similarly, SilverLight could very easily overtake Flash as the most widely-available rich UI runtime, but support for more mobile platforms will surely help this cause.  In a recent blog post, Miguel talks about a library for the iPhone that helps define UI's more declaratively -- a trait SilverLight already handles well.

To envision the kind of difference this could make, imagine launching Windows Mobile 7 with a few thousand iPhone apps ready to run on it.  This makes WinMo7 a much easier switch for consumers, and it could be a reality if we were already developing iPhone apps in C#.  Aren't the folks in Redmond sick of seeing every company from Dominos to Nationwide Insurance telling us to download their iPhone app?  Why not level the playing field and let these companies publish an app that works on any phone?

It's within reach.

Reblog this post [with Zemanta]

If HTML is dead, what’s next?

The introduction of Apple's iPad got a lot of people talking about "apps" again. There's no denying the oppressive popularity of apps today; everybody's got an app store and everybody's playing catch-up with Apple. Apps are the new hotness.

Yesterday, Stephen Forte (Is the iPhone (and Android) the harbinger of death for web pages?) observed that apps kick the crap out of web pages when you're on a mobile device, which is why we're seeing an app revolution similar to the one that launched HTML (and the web) to prominence a decade ago.

Assorted smartphones. From left to right, top ...
Image via Wikipedia

The part he missed, though, is the negative impact of a fractured client landscape.

When you see a Fortune-500 company announce a new iPhone app, do you ever wonder what it expects its Blackberry customers to use?  How about Android?  Is the cost of the new app hotness a need to build four copies of every app?

On iPad day, I caught an interview on NPR's "Marketplace".  Josh Bernoff from forrester.com was talking about how the web is effectively shattering due to the different experiences on each of these platforms.  To me, this demonstrates that we're in the midst of a fundamental transformation.

The web (specifically, HTML) was the great equalizer. Any server could serve any client. This simple concept "made" the web.  We're now experiencing a shakeup to this universal access. The web is now accessible to more devices than ever, but the cutting edge is rich client development (apps), and this is hugely fractured. On the web, we have technologies like Flash and SilverLight, and on mobile devices, you can develop for iPhone, iPad, Android, BlackBerry, Palm Pre, Windows Mobile, and others.

Today's development tools give us no practical way to target all (or even most) of these client platforms "natively". This is not due to technical impossibility; it's a function of the power struggle that's occurring among these warring platforms. If Microsoft and Apple both wanted to see SilverLight run on an iPhone, I'm confident that it would have happened by now.

Instead, all the major players in mobile platforms want to own that whole space, and their proprietary UI's are required for this. If Apple, Windows Mobile, and Android all ran flash, for example, Apple's dominance in mobile devices would be severely compromised (after all, I can get the same "apps" on any device at that point, right?).  Apple doesn't want to see this, obviously -- it takes money directly out of their pockets.

The impact of the splintered web on developers is twofold.  First, and most obviously, every app must be coded from scratch to run on each platform a developer wishes to reach natively.  This is going to force a pretty uncomfortable reckoning with Product Managers, and it's probably going to mean that in many cases, only the top one or two  mobile platforms is served, leaving the rest of your customers to eat HTML table scraps.

The second impact on developers is a splintering of skill sets and tools.  If I want to port my .Net application to iPhone / iPad, I'm looking at a sizable intellectual and financial investment.  At a minimum, I need to buy an Apple computer, because you can't do Apple development on a Windows box (no monopoly there, right?).  Only then can I even begin to try to port or rewrite the app.  Tools like MonoTouch can help preserve my business libraries, but the UI transition won't be seamless.

In practical terms, the specialization needed to be good at developing for any of these platforms also means that any one developer can't be great at all of them, which implies that I need multiple developers to target multiple platforms.  This is starting to get expensive, now, isn't it?

In time, it's inevitable that the market will work this out.  One of the major platforms will win, relegating the others to the "Island of Misfit Technologies", or a number of them will agree to interoperate (via Flash, SilverLight, HTML 5, etc.).  In the mean time, though, businesses need to expect more expensive development if they want to reach all their users with native apps, and developers had better be prepared for more UI platform changes.

Do you miss the good old days of HTML already?

Reblog this post [with Zemanta]