As many of you who follow me on twitter may have noticed, last Sunday I was having an incredibly difficult time trying to actually compile, deploy, test and sign my Playbook app from my normal ANT build project (AntPile). I should say that my workflow is a little more complex then most people’s but it is one of the ways I’m actually able to target and deploy to any device such as Web, Chrome Market, Desktop, Android, TV and recently iOS. I also expected, as a person who handles the automated workflow of some very large and high profile sites, to be able to simply modify a few target calls and have my app ready to resubmit for the free Playbook offer ending in 2 days. What happened next was a 5 hour rant on twitter questioning not only why this process was so damn painful but if it was even worth the time I was investing for a free $500 device.
So I figured I would outline 5 key things RIM MUST DO, in order to make their platform appealing to developers.
-
Better Developer Website – Many of you heard my initial rants about how I had to get a notarized letter to be accepted into the program. Looks like RIM has fixed this but still there is no one click solution for becoming a developer like Apple, Google and Microsoft have. When I re-downloaded the new Playbook SDK I had to create a new account even though I was already a registered developer. When I needed to get a certificate, I had to submit a form and wait 1 day to get it back unlike Apple’s instant provisioning process. That being said the final part RIM needs to improve in their dev portal is how applications are uploaded and versioned. It makes no sense that I have to zip up a bar which is a zip and I was totally unsure about what even had to go in the final file since it’s not intuitive. Their tools should give me the exact output I need to upload to make it easy. Google’s app store is a perfect process. It’s clean, simple, easy to follow and instant.
-
Cleaner Tool Install Process – As of now the tools actually copy a version of Flex’s SDK into it’s own directory and installs the packaging tools. This is problematic because as Adobe updates the SDK, the Playbook SDK could find itself out of sync. There have been 3 Playbook SDK releases and Adobe has actually updated AIR twice since then which have rendered my build of the tools and Flash Builder broken each time. I realize this is all still “beta” but RIM needs to be more forward thinking about how their tools play in the Adobe ecosystem. This actually requires me to have two separate ANT build paths for the Flex SDK to target the Playbook and another to target everything else. Their tools should install in it’s own folder like how Google’s tools do and the supporting SWCs should be automatically linked up in Flash builder for you.
-
Follow Adobe AIR’s Conventions – It is kind of strange that I need to create an additional application xml for Playbook apps when I already have a AIR application file ready to go for the other 5 platforms I’m targeting. I realize RIM has no need for the AIR descriptor once it’s on their OS but please just have the tools generate the proprietary one for me from the pre-existing AIR app xml. I shouldn’t have to know anything about how your .bar is structured, as an AIR/Flash/Flex developer I am going to assume I am doing just that, packaging a normal AIR app that happens to be converted into a bar or what ever proprietary application zip is needed to run on the device.
-
Developers Needed Devices Yesterday – So you are launching a new, untested platform that no one outside of a few key people have been able to see and you expect developers to build solid apps for launch day? After two years of mobile development I know that testing on an emulator or VM in this case is not going to cut it. This puts a huge burden on developers who now have to wait for their free device to validate performance and usability while consumers are actually downloading and rating them on day one. Developer must come first! I hope all the free playbook offers arrive at least a week in advance of the launch because I would hate to see my app get 1 star ratings over bugs I can’t fix because I didn’t have hardware. No other company has done this because all other major mobile companies had phone devices before their tablets so devs at least had a basic understanding of what the next device would be like. The Playbook is nothing like a Blackberry phone and the gap of technology is too large to make assumptions on without real world testing.
-
It’s Time To Innovate Not Regress – Unfortunately developing my first playbook app was the worst mobile dev experience I’ve had yet. Mind you the app I ported over was already running on a website, chrome market, iPhone, Android and even TV so this was literally a straight port of an already exiting Flash mobile app. This is the power of Flash and I still don’t understand why developing for Playbook had to be so challenging? It’s time for RIM to really pay attention to what it’s competitors are doing with their dev communities and take it to the next level.
To illustrate the problem RIM will face, take a look at Microsoft. In my opinion they have the best tools, a great language and a solid OS but they came very late to the mobile OS game and are having issues trying to gain market share. They are incredibly developer centric and they even sent me a prototype phone to work with which in turn help the company I work for win WP7 business. RIM can’t hold a candle to Microsoft’s tooling, they have not done a good job with making developers happy and finally they are going up against Apple in the tablet space. Without killer apps it’s going to be a hard sell. RIM needs to step up their developer tooling it entice more innovative apps. After my experience with AIR on the Playbook I would be hard pressed to invest the time porting over my Android apps. If this was the first tablet EVER released I would understand but it clearly shows RIM is not using other companies tools to find ways to innovate what has already been done. It all reeks of a rushed to market product. Unfortunately that doesn’t work so well, take a look at the Xoom for example. Hopefully the Playbook will be a little more stable but I don’t have high hopes because I haven’t been able to actually test anything on it yet.
In the end I want to like the Playbook. When I saw it at CES it was very impressed. In that time however we have seen the release of 2 next generation tablets, the Xoom and iPad2 which will be followed by the Tab 10.1 and Tab 8.9. If RIM wants to even compete they will really need to do something amazing on the software side. Relying on enterprise isn’t going to cut it, they have all adopted iPads by now. RIM will not be able to enable us developers to deliver the software they need to compete without better tooling and workflow. I hope someone at RIM sees this and does something constructive with it. I happen to be their ideal developer, I have a lot of exposure to different languages, I am well connected, I do a lot of development, I work in the enterprise space and people listen to my opinions. I would love to see RIM change my first impression, in fact it’s actually mandatory that they do before I invest any more time in their platform.
Oh and I want to thank all the people who reached out to me on twitter yesterday to try and help. Especially Julian from QNX and Renaun from Adobe who I chatted at great lengths to about my issues. Also Thanks to Joseph Labrecque and Jon Campos as well as every one else who threw their opinions and thoughts in the ring.
