We Did It That Way for a Reason

June 3rd, 2010  |  Published in etc  |  1 Comment

I went on a “find news apps” troll of the app store tonight. Some observations:

First: The part of the iPhoneOS SDK provides an RSS parser? Apple should ban any apps that use it for more than 50 percent of their functionality. You know what it gets end users? A bunch of half-baked apps that can read RSS feeds and little else because—hey—the RSS library was free with the SDK. We’re being asked to cough up $0.99 for someone’s RSS-enabled version of “Hello World.” I guess it’s a small mercy that Apple is standing at the ready to make sure it’s someone’s RSS-enabled version of “Hello World” written in Objective C.

Second: It might have been good of Apple to provide more structure to magazines wanting to sell individual issues through a free app. The reviews for apps on a number of flagship magazines are abysmal. I wouldn’t mind having “The Nation” on my iPad, but when I see the app getting plastered by reviewers because readability is poor, I figure I’ll wait a while.

Joe Clark wrote of the iPad edition of “Wired”:

There’s no live text, meaning there’s no search. It also means there’s no accessibility on the first computers that are accessible by default if you the developer do no extra work at all. (Follow the spec exactly and your app is accessible right away.) Think of how much effort it takes to blow an opportunity like that.

Jonathan Hoefler designed the typeface and gamely sticks up for his client, but surely he realizes he’s defending a turd. Anyone with a copy of InDesign and good type skills, which admittedly take a decade to acquire, can produce a PNG export of type that is set as nicely as Wired.app’s. But it’s still a picture.

I don’t know what Apple can do: It provides a decent SDK, it has a decent HTML renderer, it provides the mechanisms a publication needs to sell copies through an app. But it feels like the the early arrivals in the iPad space are screwing up and stumbling around. My first thought is: Provide some sort of “Pagemaker for e-Mags” app that allows people to lay out a HTML/CSS-based, iPad-optimized edition of their content and spares them the embarrassment of shipping 250MB PNG files even if it costs them the opportunity to go bananas with the design. But that brings me to my third observation of the evening:

A number of developers have decided to “recreate the newspaper experience” by writing RSS readers that slather articles and excerpts around the screen in a manner designed to imply a newspaper page:

Early Edition

The problem with that picture is that the “page” was “designed” like that based on the order in which the articles in the feeds in the Business section were published. It doesn’t have anything to do with which content is most important, most potentially interesting, or anything else … it just kind of looks like a newspaper page, which caused a reviewer to gush that the app is “the future of news.”

The app provides an atrocious reading experience. The eye darts around to each of the zones in the grid in search of organization: Why do some items get an excerpt and others don’t? Why are those four heds stacked like that down in the lower left? Are they related to each other? What about the story in the lower right? Why is its dek longer?

After about a minute (and checking a few more sections of the app out), you realize “Oh … no reason … it’s just that there are ten buckets in the design and the articles go into them in the order they appeared in the feed.”

It’s not even clear (because there are no publication dates or times), which order the articles arrived in. It doesn’t even merge the different sources for each section, so you’re just getting the articles in chunks from each source, in no overarching chronological order.

As a horribleness side note, each of the articles can scroll within its tiny zone on the grid, so if you accidentally put your finger over an article and try to do another motion, it might decide you’re trying to scroll the article instead of the other motion. The developer has successfully reintroduced the chaos and confusion of frames as a way to complement the confusion of trying to make sense of the day’s news with his app.

I kind of get why the app seems so striking at first. It does the whole grid thing, and grids are … well … were sort of cool and all, and they represent one of those design trends nerds swarm all over because they involve math. If you can do math, you can make a grid. If you can lift Apple’s RSS library and pour the contents into a grid, you can make something that looks sort of like the New York Times, and that elevates your app from a $0.99 RSS-enabled “Hello World” to a $3.99 RSS-enabled donkey in a lion skin.

Anyhow, the last thing I want to do is present myself as some sort of design expert. I put together a few newspaper pages years ago and it was mildly excruciating. But even on the first day I stood in front of a light table with X-acto knife and wax stick in hand, I knew that the important story was supposed to go on top.

You know what would be sort of cool?

If the app started out just splashing random articles across the page in order of publication your first few days of use, but quietly kept track of what you were tapping to read more of. Over time, as it gathered that attentional data, it could analyze it and develop a set of assumptions about your interests and put together something that not only looked sort of like a real newspaper, but worked like a real newspaper, only “edited” by you in a natural, indirect manner. It could assess keywords in headlines and create a package of articles built around a topic (maybe down there where four heds are stacked for no good reason). It could even let you opt in to an anonymized pool of fellow readers who share attentional data and pull stuff in to your personal edition based on a recommendation engine. It could provide a customized ticker for the headlines from feeds you like to keep an eye on but don’t read closely.

Kind of like Google News? Maybe, but without the dependence on Google’s massively distributed popularity contest. Truly personal.

Something that occurred to me as I typed this was that I could put something like this together on my desktop.

From January of last year, I’ve got an entry on how I graphed attentional data from my NetNewsWire use with Ruby with Ruby. And there’s also an entry about how easy it is to dump all of Safari’s history with Ruby .

Attentional Chart

So on the one hand, we’ve got a big lump of data in the form of my existing NNW subscriptions and the relative attention each gets. On the other hand, we have the stuff I’m reading in Safari. The Safari stuff is more attentionally juicy because it’s usually stuff I went looking for. My NNW subscriptions are less juicy because they represent a collection of habit, duty and professional obligation on top of genuine interest.

But the Safari stuff … subjecting my Safari history to a nightly trawl could seed a database of domains and page titles. As certain keywords and domains bubbled up in prominence, feeds could be autodiscovered and added to my NNW subscriptions. As feeds proved less interesting over time, maybe as my interest in a topic faded, they could be quietly retired from my active NNW reading list. As a way to reinforce the engine, I could even create a slush list of boomarks for retired feeds: Wish something hadn’t been dropped? Prove your love by visiting it in Safari now and then until it can earn its way back in. It’s in the slush list but not in NNW a period of time after being added? Gone.

Just an idea.


  1. Density :: dot unplanned says:

    June 16th, 2012 at 11:31 am (#)

    […] I recently started using Tweetbot to follow my Twitter lists, and noticed the same thing going but felt  more relaxed when it happened. Probably because with Tweetbot it’s possible to rip down a list, nine or ten posts at a time, and there’s no deceptive visual cueing going on, which is something I was irked about  two years ago: […]

Leave a Response

© Michael Hall, licensed under a Creative Commons Attribution-ShareAlike 3.0 United States license.