Monthly Archives: September 2014

The Future Library

The Future Library project will collect writing to be locked away and published for the first time in 100 years:

“Each year, the Future Library trust, made up of literary experts – and Paterson, while she’s alive – will name another ‘outstanding’ writer who will be contributing to the artwork. The trust is also responsible for the maintenance of the forest, and for ensuring the books are printed in a century’s time. A printing press will be placed in the library to make sure those in charge in 2114 have the capability of printing books on paper.”

Love this idea. Although for a different goal, I think we need a similar set of trusts to maintain electronic publishing. Domain names and hosting are much more fragile than the paper Margaret Atwood will print her story on for this project.

Using Ember.js

Brent Simmons isn’t totally convinced about the new crop of JavaScript frameworks:

“Part of me thinks those frameworks are overkill (even jQuery), and that writing regular-old JavaScript to do what you want is not that onerous a thing, and will make for leaner, better code.”

This was how I felt when I built Watermark. It uses jQuery and Bootstrap, but otherwise it’s pretty old-fashioned JavaScript. Even the parts that are Ajax just fetch and insert HTML that has been rendered by the server. There’s not much to do in the client.

For my latest project I’m using Ember.js. I want this app to be very fast, and I think putting more work on the web browser is the way to do it. I only know the basics of the framework so far, but already I like it. It feels lightweight to use, even if the actual JavaScript include is fairly large.

And 100k is really not that big of a deal anymore. You don’t want bloat for no reason. But look at a popular site like and you’ll see several JavaScript files between 200k and 500k each. They get cached and no one complains about performance.

I used my old iPhone 4S again yesterday for the first time all year. Still think it’s one of the best things Apple has ever designed.

→ 2014/09/05 8:50 am

Twitpic and what’s next

After being threatened by Twitter over trademarks, Twitpic has decided to shut down in 3 weeks:

“We encountered several hurdles and difficulties in getting our trademark approved even though our first use in commerce predated other applications, but we worked through each challenge and in fact had just recently finished the last one. During the ‘published for opposition’ phase of the trademark is when Twitter reached out to our counsel and implied we could be denied access to their API if we did not give up our mark.”

Twitpic is not a small hobby site. It grew to a $3 million business at its height in 2012 according to this Mixergy interview. Founder Noah Everett also attempted to launch a Twitter-like service called Heello, though it never gained much traction and appears to be offline.

John Gruber asks:

“Why not just change the name to something original?”

While I wonder if that comment may have a dual purpose, aimed as much at Standard Markdown as Twitpic, I’ll answer it anyway. Because Twitter has a well-documented history of stepping on developers. This trademark fight is just the latest, and at some point, I have to assume that Noah was fed up and called it: enough is enough.

Enough with building apps in a toxic ecosystem. Developers who care about microblogging should take it back. Let’s build tools for the web that will matter, that will move the web forward and make our writing last, not locked away behind APIs and ads.

Owning the microblog

I mentioned earlier this year that I really liked how Noah Read posted tweets and photos to his own site, calling them snippets. I’m borrowing that name and implementing something similar here. If you visit my home page, you’ll see these short posts already interspersed alongside regular, longer posts.

These new snippets don’t show up in the normal RSS feed. There’s a separate feed at /snippets.xml for them, and they get their own category for browsing on the web. I’ll also be cross-posting them to

Like the posts in Dave Winer’s new Radio3, these RSS items don’t have titles. They’ll be short like a tweet or post: usually 100-200 characters.

This is the first step in a larger project that I’m working on — something fun that reminds me of the early days of blogging. But if nothing else, it’s nice to control my own content again. I should have kept all these microblog posts on my own domain years ago.

Sunlit 1.2.1 is now available in the App Store, fixing some problems communicating with Flickr after they switched to SSL-only. We fixed this months ago but I, uhm… forgot to submit it to Apple until recently.

→ 2014/09/04 12:16 am

I registered a new domain name tonight. It uses .today, one of the new crazy top-level domains. I think it’s a good fit, and a reminder to have fun with the project.

→ 2014/09/03 11:29 pm

Thank you Movable Type

10 years ago I switched this blog from Radio Userland to Movable Type. Looking back, it’s incredible how mature the software was at the time that it could last this long. I didn’t stay current with updates; the version I was running was ancient by any standard.

I’m planning some new blog-related experiments soon, so it was time to migrate to a new blogging app. While I still prefer static publishing, after evaluating a bunch of options I ended up going with WordPress. I don’t love it, but it has one feature that is more important than anything else: I’m fairly confident it will be around and well-supported 10 years from now. Rands recently made a similar choice.

I preserved every blog post URL from the previous version of the site so that links don’t break. This wasn’t trivial. For a time it had me lost into the darkness of writing WordPress plug-ins, which didn’t work anyway, before finally coming to my senses and writing a script to post-process all the Movable Type titles into custom slugs.

The RSS feed is also the same, but not the RSS item GUIDs. This means you’ll probably have a one-time “mark all as read” in your favorite news reader. Apologies, but it was either that or never ship the new site.

There’s a little bit of cleanup and design work still to do. The site already looks much better on mobile devices, though. I’m excited to get back into regular blogging.