Posterous

Monday, December 22, 2008 by Dave Winer.

I've noticed that Mike Arrington tends to use Posterous for pictures he posts while traveling, pics of his dog Laguna, random stuff. I wondered why he used it instead of Flickr, which is what I generally use for pictures and small movies, and today he wrote a review and explained -- it's because it takes absolutely nothing to set up. You just send an email to post@posterouscom add an enclosure if you like, and it automatically creates a blog if it doesn't know you (you're identified by your email address) and then creates a post to hold the enclosure and text. This is the way we like our software, easy to get started with, and with instant rewards. Good work!  

I tried it out, enclosing a copy of the MP4 video of Singin In The Rain from 1929, with a bit of text scarfed from scripting.com, sent as an email to Posterous, and sure enough a moment later, it sends back a pointer to a blog with a long weird name, and I click on the link, and there's the text and the movie.  

After that I went back and read the email, it said they were happy to meet me, and I could sign up for an account and my Posterous blog would then have a nicer name. Seemed like a good deal. It suggested "dave" -- but it turned out to already have been taken. I then tried "d" -- that was too short, then "dw" which it approved, and now I've got yet another presence on the www. 

Now come the questions. 

1. Does it have an API? If not, then it's fairly useless as a blogging tool. It should, at a minimum support the MetaWeblog API, so that tools written for WordPress, Blogger, TypePad and all the blogging tools I''ve written (Radio, Manila, lots of one-offs) are compatible. It should also support the weblogs.com ping protocol, which will let it integrate with virtually every service of the "live web" (and as far as I know they do support it). 

2. I reviewed their RSS feed for my site, and it's pretty good! They don't fuss around with multiple versions of the feed, and their RSS is mostly plain vanilla, i.e. really simple, the kind that every RSS processor will understand. Now a few things they could do to simplify even more. 

a. They declare three namespaces at the top of the feed, but only use one. The other two should be removed. 

b. There's no version number on the <rss> element. Since they use namespaces, it must be 2.0, because RSS didn't get namespace support until 2.0. Dan MacTough notes that the version element is there, in the midst of the unused namespace declarations. My bad! :-( 

c. It does no harm to use a CDATA on the <description> element, but it isn't necessary since all the characters are properly encoded.  

d. I don't like that the permalink is encoded in the <description>. Unfortunately this has become common practice in RSS, but the information is already in the <guid> element, which is good. They're presumably replicating it because some reader doesn't display the permalink from the <guid>. I say deal with the problem where it's located, get the reader to display the permalink. Because of this extraneousness, in software that behaves well, the permalink will be displayed twice, unnecessarily. Yuck! 

e. Same with the link to the comments. RSS 2.0 has a <comments> element. I wish people would use it. 

f. Finally, they use Yahoo's Media RSS namespace to convey the information about the MP4 movie I enclosed. I guess some software they want to work with isn't looking for the base <enclosure> element that was designed for exactly this use. In cases like this, I support both, because it should be possible to write a podcatcher or, in this case, a movie-catcher, that conformed to the original spec and knew nothing about Media RSS, which came later and is an optional extension. The way Posterous has coded it, such a catcher app will completely miss the movie. This is the way breakage creeps into a community, and breakage is, of course, bad. 

But on the whole, they did a very nice job, otherwise I wouldn't bother with the feedback.