I've been thinking a lot recently about the subject of mashups. It's trendy, and we're seeing a lot of it, so that's not surprising, I guess.

We see a lot of openness in applications that handle personal information, things like blogs, or local events, or maps, etc. Our space touches that, but it's a little different - we want to help people create and publish documents that involve daily business and personal tasks. I want to bring meaningful openness (that is, more than just an RSS feed of your calendar, for example) to this kind of data. I think this is possible, and desireable.

When we set out to do this, one of the reasons we didn't try to do a whole office suite was that we didn't think we could push users into it. I still don't. OpenOffice has been around for a long time, and has a really tiny share, and there's a reason for that - you have to add a lot of value to overcome the MS marketing machine, and user inertia. So, we thought we'd focus on a smaller problem where we could really add value, and where folks could start using it without having to make a huge committment. And, to a large degree, I think that's worked well - we've doubled every week since we've been open, with no advertising or marketing at all (finally hired a marketing person this week, though!)

The apps I see coming out that I like follow this philosophy of focusing on the user and a narrow, well-defined problem they can solve well, rather than trying to build out a whole desktop-based suite. I've started talking with a number of these folks about building a common, open, simple API for mashing these apps together. I think the idea would be that once your users get to the boundary of the problem you're solving, you need to have a good solution for helping them work with the next application (so, for us, if a user wants to do a graphical presentation, or a technical paper with Tex, etc, we shouldn't have to code it, and they shouldn't have to move the document, copy data, or sign in again to do it). If you do, they'll stay with you, if you don't, ultimately, you will not succeed.

It's my belief that this is a very powerful architecture - if it catches on, it will be very hard to exist as an application without participating in this API, and if you do participate as one, you'll have to be a very good neighbor, and work hard to solve your problem well. I think it will be a huge win for the user, and a great opportunity for developers to break out and build an ad hoc platform that solves real problems. I think the ease of connectivity on the net makes this possible, and the mashups we're starting to see are proof of concept.

I've never done anything like this before, in terms of talking to different folks about building a standard API. If you're an app developer and you want to participate, please drop me a line. I think there's something really interesting to be done here, and I'd like to get it right. I want to talk to as many folks as possible about this.

UPDATE: here's the working doc on this, email me if you want to collaborate.

http://www.writely.com/Doc.aspx?id=a469zkvzp9r