IBM

Is RIM doomed to repeat history?

by alec on February 11, 2011

I invite you to cast your memories back to 1992.  No, not the election of Bill Clinton, but the IBM launch of OS/2 2.0.

OS/2 2.0, IBM’s multi-tasking OS with the ability to run Windows applications in virtual machines was widely touted as “a better Windows than Windows”.  And indeed, compared to Windows 3.1, it had many advantages.   It could run Windows applications and OS/2 applications.  It could isolate Windows applications in separate virtual machines so that when one instance of the notoriously unstable Windows OS died, you could keep working.  And it multi-tasked oh-so-smoothly as IBM’s pre-eminent OS/2 sales pitcher David Barnes demonstrated over and over again to rapturous audiences.

OS/2 was a huge hit with technophiles.  Developers, however, didn’t take to it as readily.  Why write an OS/2 native application, they reasoned, when a Windows application would do?  The OS/2 market was small, the Windows market huge, and the extra expense for a native OS/2 application wasn’t justifiable.

In 1995, Microsoft released Windows 95.  Technically it closed the gap with OS/2.  More importantly changes in the Windows OS made it more difficult for OS/2 to virtualize.  Without a developer base of native OS/2 applications, OS/2 withered on the vine and died.  In 2006 IBM stopped building new versions.

The lesson of OS/2 vs Windows is that there is no substitute for building a strong native application ecosystem.

It has been widely speculated since the RIM Playbook announcement that RIM might include some kind of Android compatibility in the Playbook.  The rumours, it seems, just won’t die.  This morning’s Postmedia Network papers contain a piece by Hugo Miller and Olga Kharif making this claim again, BGR reported January 26th that RIM might be choosing the Dalvik Virtual Machine for Playbook (which Android is built upon), Fortune wrote about this rumour in December and so on and so on.

The strategy under discussion is the addition of an Android compatibility mode to the Playbook’s QNX operating system.  The thinking is that enabling Android applications to run natively on QNX would allow customers to use the 200,000+ native Android applications that are out there already.  Instant ecosystem.  QNX would become “a better Android than Android”.

This is a risky strategy.

Bringing Android apps to Playbook risks ceding the developer ecosystem to Google. Developers wouldn’t need to write Playbook apps – an Android app would allow them to cover both bases.

Bringing Android apps to Playbook risks giving control of the Playbook user experience to Google. In order to have the best Android experience on Playbook, RIM will need to devote a team to just ensuring that their VM retains compatibility with the Android VM.  And what should developers target?  A native Android experience, or a native Playbook experience?

Small changes in Android could cripple Playbook. What if changes in the Android VM made it harder to virtualize on Playbook?  What if Google decided to pull Android Market support from RIM?

In short, RIM’s success could become tied to Google’s good will, and over time there may be fewer and fewer native RIM applications as developers put their energy behind Android’s momentum confident in the knowledge that their applications will run on RIM platforms.

We still don’t know what RIM’s actual strategy is.  Let’s hope that their strategy is more than just hitching themselves to the Android band-wagon.

{ 6 comments }

Apple’s naked self interest

by alec on April 12, 2010

Any vendor in the platform business knows that their primary product is programming interfaces – the so-called APIs that developers depend upon in order to deliver applications.  The API exposes features of the platform, and differentiate applications running on that platform from all others.  Lose control of the API, and you will lose control of the developer.  Developers are the leading indicator for platform success.  Ergo, lose the developer, lose the platform.

Steve Jobs’ protestations about quality aside, Apple’s recent moves to bar developers from using any but Apple approved technologies to write iPhone applications is naked self interest… nothing new. For example:

  • In 1992, Borland was kicking Microsoft’s rear with Turbo C++ and a development framework called Object Windows Library (OWL) which abstracted the underlying Windowing system out of existance, allowing developers to write applications that run on several OS platforms, including IBM’s competing OS/2. In Canada, Microsoft’s developer market share was below 30% in 1992.  They were about to lose the entire Windows franchise to Borland – a company that made programming languages. Microsoft responded by releasing C7 with the Microsoft Foundation Classes (MFC), Visual Basic, and a number of other tools that eventually became Visual Studio.  Within less than two years, market share amongst developers had climbed back over 70%.
  • In 1993, IBM made their famous statement that OS/2 was a better Windows than Windows.  They allowed Windows applications to be run inside a virtual machine on OS/2, giving them instant access to the thousands of Windows applications on the market.  Having done that, however, developers chose to write applications for Windows instead, knowing that they would run well on OS/2.  IBM never made significant inroads into the developer community, and OS/2 never gained traction.
  • In the telecom industry, Nortel (and many others) routinely asked developers to sign NDAs before allowing access to the APIs for their products.  The NDA was a legal mechanism by which vendors could prevent their APIs from falling into the hands of competitors, and thus hinder the development of compatible knock off technologies.

Controlling APIs by tying them to the use of specific development tools is simply a bare-knuckled way of retaining control of the platform, which is critical to the health of the iPod/iPhone/iPad franchise. It’s not about screwing Adobe or developers who use Flash.  It’s about not allowing Adobe and other cross platform vendors to screw Apple.

As Microsoft learned from their experience with Borland, Jobs and co need to be mindful that developers will naturally migrate to tools which offer advantages such as productivity improvements, cost savings, or user experience benefits.  Prohibitions, unfortunately, will only work for Apple for a short time.  Ultimately, they must choose to compete for the hearts and minds of developers.  I expect that they already know that, however, and are planning one or both of:

  • a licensing program to allow third parties to build development tools that use the iPhone API natively.
  • a suite of modern development tools and languages that give productivity benefits in line with what developers are used to in the web development world.  MacOS already comes with Ruby built in.  Why shouldn’t iPhone OS also support Ruby or something similar?

Either would also be in their rational self interest. 

{ 0 comments }

JAJAH.Babel real time translation launches

August 7, 2008

Image via Wikipedia JAJAH.Babel just launched a few minutes ago.  It’s a simple service that puts a translator on the other end of the phoneline.  Developed in conjunction with IBM Research, JAJAH.Babel lets you call a local number in China, the US, the UK and Australia, speak a message in Chinese or English, and hear […]

Read the full article →

VoIP Magazine on MS Unified Communications

June 30, 2006

Bryan Richards, editor-in-chief of VoIP Magazine.com, has a go at Microsoft’s Unified Communications Strategy.  In main, he wonders about the value, writing: “It’s all very interesting technology but seems a bit much to make a phone call even if it is just to leave a voice mail.” and… “making a phone ring when someone is […]

Read the full article →

Mail from Microsoft’s Kyle Marsh

June 29, 2006

A couple of days after I wrote Microsoft’s Unified Communications “Strategy”, I got a nice note from Kyle Marsh, unified communications evangelist. He gave some additional details about the products, as well as correcting a mistake I made. I understood that Exchange would be renamed Microsoft Communcations Server, which in fact is not the case. […]

Read the full article →

Microsoft's Unified Communications "Strategy"

June 26, 2006

For the last week, I’ve been wondering what Microsoft would announce today, at their Microsoft Unified Communications Group Strategy Day (what a mouthful, eh?). You see, ever since Ina Fried wrote her piece titled Microsoft Aims to End “Phone Tag“ on News.com, people have been phoning me and asking what I thought of it. My guess […]

Read the full article →

Free Software, to a Point

September 23, 2002

Free Software, at Least to a Certain Point.  The free software foundation holds a dinner and passes the hat, collecting $6,000 to keep the thing going.  Why? "Mr. Stallman’s point, now supported by many thousands of programmers worldwide, is that software becomes better when more people can work on it. Recently, several companies, including I.B.M., […]

Read the full article →
Alec on LinkedIn Alec on Twitter Alec on Facebook Calliflower on Youtube RSS Feed Contact me