Tuesday, June 29, 2004
The Taligent Effect
Don Box coins a new term
The story is somewhat more complicated than that. Taligent was started as a joint Apple/IBM venture to build a brand new operating system from the ground-up using object-oriented technology. The core team of 150 software engineers came from Apple's "Pink" project which was developing the next generation Macintosh operating system.
As time went by, it was decided that people didn't really want a new operating system, but that rapid application development was still important. Taligent software became a layer on top of existing operating systems such as AIX, Windows NT, etc. This set of common application frameworks was called CommonPoint.
As Chris Double points out, Taligent produced good work and CommonPoint was pretty nice. Unfortunately, Taligent ended badly and the only tangible results are those three books. (This isn't entirely true, the incredibly useful ICU package was developed by Taligent and is widely used).
Apple never benefited from Taligent. After a few failed attempts to produce a next generation Macintosh OS themselves, they bought NeXT and morphed NeXTStep technology into Macintosh OS X. OS X has a much more pragmatic design. It's based on the Mach kernel and BSD Unix. OS X includes OO frameworks such as Cocoa but it's otherwise a conventional OS.
Why did Taligent fail? I didn't work there so I can't say for sure. I don't think it was solely because the team was adhering to a software trend. It's more likely that a combination of bad decisions and missed opportunities contributed to their failure. Here are my thoughts:
Update: According to Marc Canter, the other joint venture started at the same time as Taligent, Kaleida Labs, failed in a similar way. Marc uses an interesting term to describe the experience. The UrbanDictionary defines the term as having a military origin. In radio communication or polite conversation it's often replaced by the NATO phonetic acronym "Charlie Foxtrot".
Don's claim is that all Taligent ended up creating was three beautifully produced books from Addison Wesley, no useful software.The Taligent effect is what happens when a group of people put adherence to a software trend first and lose sight of the value of shipping software that people will actually use.
The story is somewhat more complicated than that. Taligent was started as a joint Apple/IBM venture to build a brand new operating system from the ground-up using object-oriented technology. The core team of 150 software engineers came from Apple's "Pink" project which was developing the next generation Macintosh operating system.
As time went by, it was decided that people didn't really want a new operating system, but that rapid application development was still important. Taligent software became a layer on top of existing operating systems such as AIX, Windows NT, etc. This set of common application frameworks was called CommonPoint.
As Chris Double points out, Taligent produced good work and CommonPoint was pretty nice. Unfortunately, Taligent ended badly and the only tangible results are those three books. (This isn't entirely true, the incredibly useful ICU package was developed by Taligent and is widely used).
Apple never benefited from Taligent. After a few failed attempts to produce a next generation Macintosh OS themselves, they bought NeXT and morphed NeXTStep technology into Macintosh OS X. OS X has a much more pragmatic design. It's based on the Mach kernel and BSD Unix. OS X includes OO frameworks such as Cocoa but it's otherwise a conventional OS.
Why did Taligent fail? I didn't work there so I can't say for sure. I don't think it was solely because the team was adhering to a software trend. It's more likely that a combination of bad decisions and missed opportunities contributed to their failure. Here are my thoughts:
- Corporate politics must have played a role. Who was going to use this OS? When the Pink team left Apple to join Taligent where did that leave Apple's core business? Apple's value is a combination of software and hardware innovations. If they allowed a jointly-owned third-party control the OS (which could also run on non-Apple hardware) where was Apple's unique value?
- In hindsight, the choice of C++ as an implementation language was a mistake. More dynamic OO languages such as Java, Objective-C, C#, etc. would be more suitable for what Taligent was building but didn't exist at the time (except for Objective-C which NeXT used). There's a long thread on Hack The Planet where people (including ex-Taligent developers) debate this issue.
- Another factor in Taligent's failure was that it was too ambitious. There were no early public successes to build on. There were lots of good ideas in CommonPoint but few people could see them. You needed hefty hardware to develop with CommonPoint and it was an all-or-nothing world.
- Apple and IBM have very different corporate cultures. They both can produce amazing products but in very different ways. Combining developers and management from these two companies was probably very difficult. Since I work for IBM and only known second-hand what Apple was like then I won't comment beyond that.
Update: According to Marc Canter, the other joint venture started at the same time as Taligent, Kaleida Labs, failed in a similar way. Marc uses an interesting term to describe the experience. The UrbanDictionary defines the term as having a military origin. In radio communication or polite conversation it's often replaced by the NATO phonetic acronym "Charlie Foxtrot".
RSS 0.92 Feed