I've had a persistent problem with static hosting, going back to the beginning of scripting.com in 1995 or so. That's a long time to not solve a problem. While all the incubators start new tech companies with viral business models, one of the most basic ideas about the web, a folder that maps onto a website, still requires a degree in bicycle repair (not quite rocket science) to keep going.
I talk to my friend Chuck Shotton about this all the time. Chuck used to write web servers.
I say to Chuck one day we should knuckle down for a month and do an completely brain-dead simple self-hosting server, running on a Mac probably, cause it comes with Apache built in and turned on. We never seem to get around to it.
Meanwhile, Amazon made it so that if you stand on one foot while touching your nose with the pinky of your right hand and say Wild Beatnik Pie! three times you can host a static site in S3, which seems to have been made to serve static sites. My friend Joe Moreno has mastered this art, and posted a howto about it. I am going to proceed to follow his directions and set up a simple site in a S3 bucket, one that very rarely changes, to see what happens.
Joe Moreno: How to host a static site in S3.
1/2 hour later, I think I've set everything up right, but I get a "Sorry invalid request" when I try to visit www.smallpicture.com, a site I set up following Joe's instructions. Maybe it takes a while for things to show up?
When I tried to go to dsxuvulkqwac7.cloudfront.net it correctly linked to the index file in web.smallpicture.com, but I got an "access denied" from (what appears to be) S3. After changing the permissions the index file correctly displays, but I still get the invalid request message when I go to www.smallpicture.com.
I think the problem is in Joe's step 2. I changed it so that www.smallpicture.com is a CNAME for s3.amazonaws.com. I bet that fixes it and www.smallpicture.com will now correctly show the small Mona Lisa.
Saw this headline roll by and had to remember to breathe. Until that moment I thought of Tumblr as a tiny company operating upstairs in a random midtown office. David Karp, its founder, is brilliant, and the site is growing at an incredible rate, and now the company is valued at $135 million. Oh to be young, talented and rich in NY in 2010.
Instant comments say its more bubble bidding, but it might not be. I hope Tumblr opts for a WordPress-style business model known as "freemium," where they give away the basic site, and charge for the extras. That will keep their interests nicely aligned with their users'. I think long-term, companies like Twitter and Facebook will have big problems as they have to narrow the margins for their participants, always competing with more members of their community, forcing them to either disappear or look for greener pastures. I don't see any clear line of places they can't go. But if Tumblr charges for sites, then its users are customers, and that's something that we're all comfortable with because we're either customers or have customers in the rest of our commercial lives. The user-as-hamster model is growing more uncomfortable all the time.
Anyway, it might not be obvious at first that this is not great news for the publishing business, as it's currently configured. Flush with cash, companies like Tumblr, Foursquare and others are mining tech talent in NYC, and bidding up the already-high prices for engineers. If you're a newspaper or magazine publisher in NYC, and you've been employing technical people, watch out -- you're competing with employers with newly deep pockets. And don't forget Google is (rumored to be) buying a huge office building in Chelsea, near the meatpacking district. It's going to be filled with technical people too.
Of all the tech wunderkinds out there, it's mainly WordPress that sees the media as customers. It probably would be a good idea if they had some competition.