This is an important discussion. Context:

* * *

I have long held that Conde Nast didn't buy Reddit so that they could have an ugly, low-margin website full of jailbait and f7u12 cartoons. They bought Conde Nast because Reddit's nested comment structure should, in theory, auto-optimize comment threads. Which is a cool feature if you own, say, Conde Nast Traveller. Or Wired. Or Glamour. Or The New Yorker. Or Architectural Digest. It becomes a really handy thing if the content you're pimping is your own. Not so useful if you've got a Wild Wild West full of angry basement dwellers who would vote for Ron Paul if only they were old enough.

Further, the "Reddit strategy" is exceptionally well-suited to support forums. Got a question? The answer that works the most goes to the top, along with commentary. Experiments in open-sourcing Reddit and making it portable support the notion that a "Reddit node" for managing your public tech support face was seen as an attractive technology spin-off for one of the largest publishing concerns in the world. It is my well-considered opinion that Conde Nast came down so hard over Sears because Conde Nast needed to show that the platform they were trying to interest B2B buyers in wouldn't come back to bite them. By the time the Prop 19 controversy erupted Alexis and Steve were fully off the leash and Conde was clearly looking for an exit strategy.

* * *

Reddit is now and has been focused solely on growth. Aside with some flirtations with building the appropriate community back in the early days, Reddit Inc. has been all about the poor, the tired, the huddled masses of 4chan. As such, their community is unwieldy and unattractive to outsiders. Somewhere I have a Quantcast report on Reddit; it's far uglier than Reddit Inc. would have you believe. Not only that but it's a truly unwieldy architecture. These two aspects - lack of scalability and a poison community - have left it a poorly-understood "other place" on the Internet, just as 4chan and Usenet were before it.

Hubski does not have that problem.

There is a lot to Hubski's architecture that solves a lot of the problems inherent with Reddit. It's far easier to ignore people you don't want to listen to on Hubski. Hubski is a lot less time-sensitive than Reddit is. I'm reluctant to highlight Hubski's community because communities can and will turn on a dime. In fact, I would say that taking this opportunity to iron out some of Hubski's future problems will help ensure that the community stays.

Porting the comments to an external source has two uses - it ports the Hubski community to that external site and it ports the external community to Hubski. The external site will do what it must but if Hubski wishes to use this externality to attract and retain useful contributors, a few things MUST be ironed out ASAP:

1) an immediately parse-able "welcome to Hubski" guide such that a moron in a hurry can figure it out. Hubski has always been about "aren't we clever with our design" without recognizing that it's pretty much the Stilton of the internet - an acquired taste that's offputting to many and worshipped by few. steve and thenewgreen have been busily podcasting it up; them muthafuckas need to put together a 30-second video that answers all immediate questions about Hubski.

2) A navigable taxonomy. I know you hate tags, MK, but you were wrong about that. I know you want people to only follow people but you were wrong about that, too. Any new visitor to the site is going to look for things that interest her, and she's going to look by subject. You have made no real effort to make this search fruitful or easy. The fact that the excellent material in #writebetterdammit is invisible to everybody new because some chucklehead went with the far-more-reasonable #writing and it stuck is exactly what I'm talking about. That might mean serious tag curation. If so, figure out a system NOW while it's still manageable.

3) Skins. I know you love your "vanilla-on-asphalt" look but when Redditors think your site is ugly you have a real problem. The functionality is there. The usability is there. The ergonomics are almost there. The style looks like something an East German varsity chess club would come up with in 1987. Nobody is going to be happy with everything, but you need some simple and inviting options so that designers who actually know what they're doing can come along and easily create something for everyone else. Style matters. It matters a lot.

4) Utter and total scalability. You don't want to be in a position where, in order to keep the site functional, you're buying large fractions of the Amazon server farm because your code never considered more than 10,000 users at a time. You don't want to have your comments over-run by trolls because your community policing methods are "comment assholes out of the code." You want to build an infrastructure where your community maintains your community and your resources are linear, efficient, and go to eleven.

This is a big decision. It has sweeping implications. Start thinking about them now, before they matter, because someone else will ask. And someone else. And someone else. And before too long, you're native on Slate.

And you want to be able to flip that switch right now, not "yeah, we think it'll work, but we've never tried anything that big before, so bear with us."

'cuz they won't.