Google AppEngine just introduced an interesting twist to this story. It's been widely pointed out that if a "win condition" for your Web-based Idea getting popular is "get bought by Google", then there's an advantage to already being on AppEngine - you already fit their deployment and scaling model, and the mechanics of the move have much lower friction. (Expect the AppEngine library to sprout many more "mid level" components, like user messaging, so you can just "turn on" features instead of reinventing them - think in terms of what the Mac APIs did for Mac apps - and see how that makes it easier for something to be "part of the family" than to be pointlessly different.)
Given that, what does the feature set look like?
This also slam-dunks the scaling question :-) It doesn't cover any of the more real-world parts of the service - business/management advice, payroll... thinking very far ahead, could google do a SalesForce.com or AdminStaff kind of play? Even in the narrow startup space, there's plenty of duplication possible (that's one reason that experience in "startups" is useful in its own right.)
Clearly there's a lot of Instant Startup 2.0 already available in a google-shaped box - but the gaps are significant too...
In 2001, Sam Hartman and I had a little company called "Mekinok", which did a little systems consulting, but had a Big Idea of doing "Startup In A Box." (More people probably remember "Boxed Penguin", the open source side of the story.) One of our motivating stories was "build the tools that we want to have for our next startup", and we succeeded at that... a large chunk of the original Debian OpenAFS packaging was done by Mekinok (well, by Sam really, but it was specifically to make our jobs easier :-)
We had "big" stories too - suppose you had a software startup idea. To actually get started working on it, you needed some infrastructure to put it in:
The boxedpenguin pages tell the story pretty well, but the key point here is we actually built everything I've listed above, and even delivered parts of it to a customer or two; enough of it got properly integrated into debian that I've certainly found it useful in my latest startup (though we're actually big enough to have actual sysadmins managing it all.)
So clearly that wasn't the big story - after all, we achieved it :-) We had plenty of other concepts:
Basic concept there was "Ok, startup-in-a-box, what do startups do anyway?" and expand outwards. That was 6 years ago; today, Y-combinator takes the reverse approach and points out "you don't need all of that company infrastructure - leave that to whoever buys you. Stay sharp, stay lean, work out of your apartment, and avoid needing all of the trappings..." That's worked out pretty well for them. But what if you do have reasons (based on the kind of people you want to attract, or the market you're in) to build up a more "conventional" organization? How has the space changed in the last half-decade? Can you get "all the toys" without having "all the burn rate"?
Well, the Big Thing to come along is "Web 2.0" and the more sober subset of "Software As A Service". If you take the same list from above, and look at it from the perspective of "assuming you aren't that paranoid, what can you hook up with"? Here are some examples, not intended to be exhaustive, just to show that you could...
One flaw is that most of these are online only, and without an FDIC-for-data, you run a lot of risk - whatever you've got on the service isn't easily migrated or even recovered. (Zimbra is one exception, in that you can use a hosted commercial version or run it yourself.)
There's enough to chew on, at first glance, that maybe it's time to put some attention into "Boxed Penguin 2.0", or "Unboxed Penguin". After all, everything in boxedpenguin was something you could put together yourself with time, sysadmin skills, and development skills; the point was to have it all in one place so you could get on with your Glorious Idea and not waste time with things like library API compatibility issues. Unboxed Penguin would at least be a checklist, with a little vetting of "these play well together" and hints like "be sure to stuff the RSS feeds from this into the external-links on that", though it also needs a story for "how do you scale", "how do you migrate", "how do you control your data/recover from something you weren't paranoid enough about". Also, for the conservative, "here are reasons why this is better than an emacs buffer on your laptop" :-)
(Comments and ideas to penguins at this domain - all comments to that address are assumed to be for possible publication, and may appear here as part of future discussions.)