By John Gruber
Atoms: We are mostly sold out... but there is more!
Most of the reaction I’ve seen regarding the rumor of a Microsoft acquisition of Yahoo centers around the business aspects. Combined search traffic versus Google, that sort of thing. Given that Microsoft views Google as their new archenemy, a Yahoo acquisition makes some sense.
But at a technical level, it would be quite a thing, because Yahoo’s software isn’t written using Microsoft technology.
One of the least remarked-upon remarkable things about Microsoft is just how much software they’ve written themselves. Not just their software products — the stuff they sell in shrink-wrapped boxes — but all of the developer technology used to build those products. Their own entire operating systems. Their own file systems. Their own compilers. Their own IDE. Their own web server. Their own scripting and programming languages. Their own SQL database server.
There’s no other company that does anything even remotely like this today.
Long ago, Apple did. Apple wrote the entirety of the classic Mac OS. They had their own compiler, their own networking protocols, built computers using Apple’s own proprietary peripheral ports. Originally, they had to, because the ’80s were a proprietary era. But by the ’90s, that “let’s do everything ourselves” mindset nearly killed the company. The biggest difference between the old Apple regime and the new post-NeXT-merger Steve Jobs regime is that Apple now focuses on just a few things. Wherever possible, Apple now builds on open source. BSD and GNU userland tools in OS X. GCC compiler, Apache web server, open source scripting languages like Perl, Python, and Ruby.
Google writes a ton of their own code and tools, but they do it using open source languages, databases, and operating systems.
That’s what Yahoo does, too. In fact, that’s what most web companies do. Web companies built on Microsoft technology are few and far between. When was the last time you saw a new hit web site developed using Microsoft’s web stack? This is what Paul Graham was talking about when he wrote that “Microsoft is dead” — there’s an entire generation of developers who are growing up without ever even considering Microsoft developer technologies.
And so I wonder what Microsoft would do with Yahoo. It would seem utterly insane to acquire Yahoo and then waste years of time rewriting Yahoo’s existing code to use Microsoft technology.1 The whole point of such a merger is that both Yahoo and Microsoft feel the need to act swiftly to counter Google’s growth.
Acquiring Yahoo could be Microsoft’s way out of the corner they’ve painted themselves into. The widespread adoption of Microsoft developer technology, historically, has been very good for Microsoft overall. But while technology matters, products matter more. If Microsoft wants killer web products, they need to embrace non-Microsoft web technologies. Acquiring Yahoo would give them a way to do this without directly acknowledging just how poorly their current web strategy has fared.
Remember Microsoft’s acquisition of Hotmail? It was founded in 1995 and was the first mega-popular free web-based email service. It was built using FreeBSD. Microsoft purchased Hotmail at the end of 1997 for $400 million. Microsoft then spent three years porting it to Windows 2000. Imagine trying that with all of Yahoo. ↩︎