Monday, October 1, 2012

HTML 5 takes a beating from Facebook, boost from salesforce.com


And you thought it had been a rough few months for Facebook.  The past few weeks have been HTML 5’s turn.
As you probably heard, Facebook CEO Mark Zuckerberg took a few minutes out from his newly public company’s on-going Wall Street woes to take a few potshots at HTML 5 at the TechCrunch Disrupt conference.
‘Biggest mistake’
"I think the biggest mistake we made as a company was betting too much on HTML 5 as opposed to native," said Zuckerberg about their approach to developing the mobile Facebook app.  “We burned through two years on that.  It probably was the biggest strategic mistake we made."
Then, less than 2 weeks later, one of the Facebook engineers followed up with a post detailing the shortcomings of HTML 5, especially around performance, adding fuel to the fire.
But Benioff is betting the farm on HTML 5
In the other corner, however, salesforce.com made a very public showing of support for HTML 5 at Dreamforce.  CEO and pitchman extraordinaire Marc Benioff launched salesforce.com’s new Touch capability to great fanfare.  I sat in on a couple sessions to see Touch demoed (minus a crash or two) and hear how they built it.  Benioff and company bet their farm on HTML 5 to create Touch.
The diametrically opposed feelings about HTML 5 weren’t lost on Jeff Hammond of Forrester or James Governor of Red Monk.
So why the mixed signals about HTML 5?  And what can an enterprise can learn from it?  Both data points actually lead to a similar conclusion:  you better know your use case.
What Facebook learned about HTML 5
HTML 5 has been touted as a panacea for mobile application development.  Write once, run anywhere -- on the Web and on your mobile device.  The problem, as Facebook found out, is that things are different across different environments.  You certainly get to leverage the development skills you already have, which is a great financial benefit, creating one environment for both Web and mobile platforms.
However, the user interface you end up with is lowest-common denominator in many respects.  And HTML 5 has been shown to perform differently across different platforms, so if it’s speed you’re looking for, you’re not guaranteed to get it.  In case you missed it, there’s been lots of grousing about Facebook’s mobile app performance (I admit I'm one of those).  And even the salesforce.com Touch demos last week showed sluggishness and clocking here and there.
One more thing that Facebook and salesforce.com may or may not have figured out:  many security questions don’t get solved by HTML 5.  This is another potential gotcha for enterprises.  (You’d hope this would be a big issue for salesforce and Facebook, too, though Facebook doesn’t seem too bothered by privacy concerns in general, frankly.)
Overhyped
Over the past year, HTML 5 has been the darling of the industry, getting the hype I remember being reserved for things like server virtualization and web app servers.  Analysts at Gartner and other firms have been pronouncing HTML 5 the Next Big Thing.
Of course, when you are at the top of the hype curve, the trough of disillusionment beckons.
In fact, Facebook’s very public comments aren’t the first beating that HTML 5 has gotten recently.  There was a bit of public discomfort about a split in the HTML 5 standard.  If something is supposed to be a standard you can use across all environments, the industry at least needs to agree about what it is.  Otherwise, you’ll have to start asking “Which HTML 5?”
SFDC: happy on the surface, but some worries, too
At Dreamforce, salesforce.com was very excited about Touch.  Of course, is Benioff ever not excited about something he’s selling?  However, two things raised a few red flags.
First, performance.  In their session on how they built Touch, the development team acknowledged it as one of the biggest issues they had been grappling with.
Second, keeping up with the functionality that people need.  In the mobile roadmap session, Clarence So talked about delivering part of their functionality now via Touch, and progressively more and more as time goes on.  They gave no specific dates for the mobile versions of their apps being fully capable.  Unfortunately, this approach means the salesforce mobile solutions might be second-class citizens for quite a while.
What should an enterprise do?  Know your use cases
So what are the alternatives to HTML 5?  And should Facebook, salesforce.com, or the average enterprise choose differently?
The traditional thought is that there are 3 alternatives: HTML 5, native, or some hybrid.  At Disrupt, Zuckerberg said he regretted using HTML 5 and not doing device-native rewrites of their application (which they have since done and seem to be getting improved performance).
Enterprises listening to that should be wary:  Zuck has a different use case from many.
Facebook’s needs are for a single consumer app, big on user interface and local performance.  Many software companies that have one main application that they are selling are in a similar boat.
So, should you side with salesforce.com instead?  Not necessarily.  Their use case is also very specific.  In the “How We Built Touch” session at Dreamforce, they outlined their reasoning for using HTML 5: many of their customers had spent a great deal of time and effort customizing their solution.  Writing a new, native mobile version of their app would require reimplementation of everything that salesforce had already done on the Web version of their products PLUS would require customers to re-implement all those customizations.  Ouch.
Enterprises may need something different
So end user enterprises may not be in the same boat as either of these vendors.  We at Framehawk have been talking with folks about use cases around mobile access to multiple applications, with a big requirement for security and performance, and an interest in a branded workspace with a unified way to get to multiple apps.
In those cases, something like what we’re providing gives all of those benefits, without regrets over cost, performance, or time to market.  Our approach is to let organizations keep their apps as they are and send only the screen images down to the mobile devices.  Our protocol lets you do that at very high speeds without putting data on the device, and still use native, touch-based interaction methods with the tablets.
So, what’s the right lesson for an enterprise to take away from all this?
Evaluate any “silver bullet” answer like HTML 5 very closely.  Make sure it matches the requirements you and your organization actually have before assuming it has all the answers.
Oh, and think very carefully before buying Facebook stock.

This post also appears on the Framehawk blog.

No comments: