Home > Archive >  2009 >  May >  28

Google's killer app

Thursday, May 28, 2009 by Dave Winer.

In May 2001 I went for a visit to Google, which was then a very young company. I offered a list of ideas that we could work on together that would help bring their search engine closer to the blogging world, which was also then, quite young. 

Then as now I saw the challenge as shortening the distance in time between a post and its appearance in the search engine. I called the idea then just in time search, it's the same idea that people call "real time" today. The term "just in time" was borrowed from the manufacturing world, where the goal is to have the components needed to build a product ready at the exact moment they are needed.  

The meeting was cordially breathless, we loved Google then, it felt like it was our company, they were on our side in keeping the web open and the playing field level. But the ideas discussed that day weren't implemented. I don't think this was anyone's fault because then it wasn't so easy to see exactly what was necessary. This was quite a bit before RSS solidified behind version 2.0, and weblogs.com hadn't fully bloomed yet. It may not have been obvious how to do it then, but today -- I believe it is.  

I described it to Michael Gartenberg in a phone conversation this morning. It took me about five sentences and less than a minute. That deserves a blog post, I thought. 

Here's what I'd like to see them do: 

1. When my blog updates I ping their server with the address of the RSS feed. 

2. They read the feed, note the changes from the last time they read it, update their index. 

3. On the other side, they allow users to subscribe to a set of RSS feeds they maintain, either on Google or elsewhere (using OPML subscription lists). Google provides an RSS feed that can be read every minute to get all changes to the aggregated feeds.  

I know they already have elements of #3 implemented in Google Reader, that's good!  

The whole thing needs to be tested, burned in and tweaked with clients that do more or less what Twitter does. How it works on the back-end is Google's business. 

If they want to do more, I don't have a problem with that. This is the part my software will use. 

I would like to see #1 based on the weblogs.com ping protocol, which included a REST interface. I'm not wed to any particular format other than it be as simple as it possibly can be and not dependent on toolkits or having any particular software installed. No opportunities for lock-in, nothing that isn't completely transparent and simple. 

Then Google gets to do what Google does best, and what we depend on them for -- run a great search engine. They get all the updates flowing through their servers. Believe me, everyone will support this. And the advantages for users are manifold. Wide choice of software to use. You can have a 140-char limit if you want, or not, if you don't, again -- your choice. You can include media objects, RSS fully provides for it, or you can choose not to. The whole point is to keep it super lightweight and give you all the choices. Let the market figure out where it goes, so we don't have to wait for any one person or group of people to figure it out. That's the Internet way of doing things. 

I can be happy that at least I've written down the idea where all can see it. I don't know if anyone from Google reads my blog, or cares. Microsoft or Yahoo could do it, I would be happy to work with anyone with the search and scaling ability.