Webreakstuff's blog on design, development and strategy. Click here to subscribe.

Clone the Google API

Fred Oliveira on November 3, 2005

Dave Winer speaks. Someone needs to listen. Let me quote what I believe is a genious paragraph from an inspiring essay:

Google

Today, in late 2005, Google is still the leader in search, but either Yahoo or Microsoft, or even better — both — could leapfrog Google by allowing developers to build unlimited applications on their search engine. Or, if unlimited is not possible, make the limit practical for serious Internet applications, perhaps 1 million queries per day? Let’s work this out.

Read the full thing. If you’re a web developer, if you think you have something to contribute to what some call the “web 2.0″, this might just be it. The point is we don’t really need to wait for Microsoft or Yahoo to give the next step in the competition for the development of the web as platform - we can and should do it ourselves.

I still remember the last party we threw at Techcrunch where Dave keynoted and said “someone get out there and give google some competition”. I remember hearing someone behind me:

- Would you like to?
- Like to what?
- Throw down Google.
- Hell yes.

Go get them, is what I say.


Comments on this post

Richard MacManus

What I don’t get about this whole thing is: why *clone* the Google API?? where’s the innovation in that. Opening up the API and making it “unlimited”, that I can understand (nevermind the business issues for now). But what is all this talk about CLONING about???

Read/Write Web

Attack of the Clones

So the latest ruckus in the tech.blogosphere is about Dave Winer’s call to Clone the Google API (note the URL name). Robert Scoble wrote an enthusiastic post entitled Yahoo’s new pretty maps are doomed (and so are Microsoft’s), which understandably…

David Luebbert

Reasons to clone:

1) The features available in that API are sufficient to enable interesting interaction with a search service. Enough experiments have been done on the Google API to know what’s possible to do with it.
Re-expressing that functionality in a different syntax would make folks re-explore the new way of expressing things and waste a good deal of developer’s time with no advance in existing capability.

2) If the API is cloned with the usage restrictions relaxed or removed, all of the existing experiments get new resources and can develop a larger user base. If you make random changes to the API, all of the previous experiments must be redeveloped. If the cloning is done well, all of the existing stuff works instantly with new API providers.

3) If new invention is needed, the servers that support the API can extend existing capability while leaving the existing core API in place. The costs of extending capabilty in this way are likely not terribly high and makes it easy for existing developers to use a new search service. The new search providers won’t have to evangelize on behalf of their new API, in order to find the developers who will support their efforts.

4) It’s in developer’s interest to have multiple services that support a core API, as insurance against lock-in attempts by the service providers. Once another outfit provides the API with relaxed usage restrictions, you would expect that Google would have incentive to also relax their restrictions. If Google ends up finally doing the best implementation of these services, service developers would appreciate being able to use the best stuff without having to support multiple, possibly diverging APIs. Providing this insurance, increases the liklihood that developers will use a new search service that uses the API along with or instead of Google.

GEEKS.OPML.ORG

Dave Luebbert on the reasons to clone

Dave wrote this in a comment on Frederico’s blog , but it was worth calling out in a post of its own.

CrunchNotes » Unrestricted APIs

[...] I agree with Dave on this. And Robert. And Fred. And (partially) Richard. Companies, even big companies, especially big companies, need to open up their data without usage limits, or else they will make their services and themselves irrelevant. [...]

Something to say?