James Williams' Blog

I'd love for this to be some fluffy quick list of ten or so points but I've really only got one point: Android should have a SocialMedia ContentProvider.

Backing up a bit, applications on Android run similarly to those on the iPhone. Each functions in its own sandbox preventing it from wreaking havoc on the system. This sandboxing is further strengthened by the permission system. One key difference is that Android apps can share information if they store it as a ContentProvider. This is how Android implemented Contacts and the MediaStore(Photos, Audio, and Video). I think the same should be done for Social Media sites. One of the joys and pains of a emerging platform is that in the absence of a market leader, people tend to try out multiple apps centered around the same service to see how they stack up.

For a Twitter-related app, that means giving each app your login details and waiting for it to download whatever it deems appropriate. I would love to be able to provide my details once and have migration be seamless. Adding a TweetDataStore could reduce duplication of data between apps.

As programmers, a species that is prone to practice what I call "Slash and burn programming"(doing a ground-up redesign at the sight of the slightest incompatibility instead of evolving a solution), we could do just that and open ourselves to the OpenID problem. I think it's better if the community comes up with a consensus to make things easier for us all.

Pictorial is an app that lets you manage your Picasa albums from your phone. Not only does it let you browse your albums and integrate search, it lets you upload photos straight from the phone with geolocation. The photos from an album or search appear slide onscreen one-by-one in a similiar fashion to PhotoStream but arrange themselves in a more boring way with basic rows and columns. Having it look like a desk of photos, each slightly askew, would have been cool but the huge amount of polish more than makes up for it. You can even customize the background. This is definitely an application I would pay for.

LocalSpinner is a nice take on the iPhone UrbanSpoon application re-imagined for Android. After giving your location which can be retrieved from the GPS, a search of events or places will present you with a colorful wheel somewhat similar to Wheel of Fortune with its familiar clunking sound. Tapping the center of the wheel takes you to the associated web site if there is one. Not a must have, but an amusing download anyways.

I generally enjoyed PhoneFlix. I found it well-designed and responsive . The app provides all of the functionality of the site besides Instant Watch and rating films. In the comments on the Market, some folks are rating it one because of the lack of Instant Watch. I wouldn't hold my breath for Instant Watch. It would require support in the API and the ability to decode DRM'd WMV files.

The only thing that was a bit jarring was the fact that there seemed to be two different design visions. The application's tabs had an Android feel to them whereas the internal screens looked like a native iPhone app. 

Despite that, PhoneFlix is a solid app to edit your queue on the go.

FatSecret is a nice uni-tasker that tells you the nutrional details of foods. When it starts, the app shows two tabs, one with a search box and a button to type in your food item and one showing common foods. It shows all the information that would show on the bar of your favorite candy bar. My only gripe is that it would be nice to be able to compose a meal instead of each item separately. Daily tracking would a nice touch too.

There are currently two Twitter clients on the Android Market, here's how they stacked up.

Twidroid

Twidroid was an early arrival to the Market. In previous versions it was plagued by instability and speed issues. For the most part, those issues have been resolved and the app is generally responsive. The Look and Feel was very polished and embodied what I would envision for a Twitter app, giving a nod to the LaF of Twitter.com while not being an overt copy. Integration with Twitter search was a nice touch but something is wrong with the background process that monitors tweets. It kept on vibrating the phone saying I had 20 new tweets when I had none. It was also not apparent on how to/if you could see anything older than your 20 most recent tweets. For power tweeters, having only the last 20 tweets is a serious problem. As advertised in the app, the camera function was buggy and had the peculiar requirement that the keyboard be slid out or the image was rotated 90 degrees to the left.

Twitli

Twitli is a more recent addition to the Market and much more basic than Twidroid. Thematically, it takes some cues from Twitter.com giving a sense of familiarity. The menu that lets you navigate between timelines has text so small that people who fat finger might have better luck with the trackball. IMHO, a very ineffective use of screen real estate. The camera function, like Twidroid, is buggy as well and has the rotation issue.

Conclusion

Twidroid is almost there. That is, if you can deal with the erroneous alerts at times. Adding paging (or making it more apparent if it's currently possible) would make it probably 80% done. Twitli, on the other hand, needs a lot of usability work to make it adequate. I would recommend using the regular Twitter site, maybe Twidroid if you're a light Tweeter.

As you may have already read, I published an app in the Android Market(read here). Having had a week to receive comments and tweak the app a bit, here are some of the lessons I've learned in the week since.

1. You will get unjustified 1 star ratings.
I've seen it on apps that did precisely what they said they would do and well. Most users are the type that buy a video game but eschew reading the manual. In their defense, there is no requirement for the app to have a help screen and they would have to go back to the market to see whatever instructions accompanied the description they skipped. By that time, they might have decided they would uninstall it no matter what.

2. Ratings are not to be trusted.
Out of almost 400 installs, I have about 250 active users(about 70%). Only 10 left comments. Statiscally, that's less than 3%. Those are kind of like the high and low scores they throw out in the Olympics. When looking at comments, it's much more helpful to look at comments addressing possible speed and latency issues than the "this sux!".

3. The Market is in beta. I know they said that but unlike GMail, it really is in beta.
IMHO, though it is a great product, I think the T-Mobile marketing made the device seem as what it is not: an iPhone competitor. The number of apps is growing every day but it would be a while until it reaches the depth of the iPhone App store. And because of the current restriction from charging for an app, mixed in with the gems are several duplicated apps. It seems a tip calculator is the Android equivalent of Hello, World!. There is only so much that can be done until you get a real device in hand. I expect the complexity and quality of third-party apps to improve greatly over time.

4. Upgrading is the concern of the developer.
The Market doesn't seamlessly inform users of updates and all comments and stats get erased if you upload a new version.

5. Unlike Alton Brown suggests, unitaskers are preferred.
It's much easier to code and/or find an app that does one thing well than one that does fifteen things satisfactorily.