Our BlogTips, Tricks, and Thoughts from Cerebral Gardens

Open Source, the GPL, and the App Store

I am a believer in the idea that a large subset of humanity enjoys helping one another to be successful, and that they become better themselves in the process. One way this occurs is with the Open Source movement. People that write software not for the profit that can be made with the software, but to give it away so that others can use it, as an end product to accomplish a task, or even to use as part of another project. This enables the next person to get a jump start on what they're working on in order to become successful faster than they would have been if they'd had to start from scratch.

Ideally the next person becomes a link in a chain so that others can benefit from their work and become successful themselves. Everyone does their part and everyone benefits.

Not all Open Source products are equal however. Just because someone releases their code as open source, doesn't mean you can do whatever you want with it. There is always a license of some sort attached to the code that governs what you're allowed to do. The Open Source Initiative site lists 67 different licenses that they've approved as meeting the existing community norms and expectations of what an open source license should be. Before we go any further, let me clarify that I am not a lawyer, I don't even play one on TV. So you should always consult your own legal council when dealing with legal matters.

Of all of the open source licenses, there are a few that are used more than others. The GNU General Public License (GPL), the Apache License, BSD License and the MIT License are the most popular. The latter 3 are very similar and impose the least restrictions on users of the code they license. The GPL imposes some very specific implications on users of the code it licenses and is the reason I'm writing this article today.

In simplified terms, the Apache, BSD and MIT licenses allow you to use the code they cover in other products and allow you apply a license of your choice to the new product. It doesn't have to be an open source license and so you're not required to provide the source of your product to others. You do have to include the appropriate copyrights for the source you've used that was created by others, and you do have to accept the fact that the source is provided to you AS IS, that there's no warranty of any kind given by the original creators of the code you're using.

The GPL however forces you to use it as the license for your new product. In this way, it has been described as a virus license since it effectively reproduces itself onto your product, and limits what you're allowed to do with your product. That is, you're forced to release the source code of your new product, and you must allow others to redistribute your product for free, in both source and object code. The philosophy behind this is that if you're benefiting from the work of others, you should be forced to share your work in the same way so that others can benefit. These restrictions make the GPL incompatible with the App Store, which means, as an iOS developer, you should avoid using any code that is covered by the GPL.

The reason that GPL code is incompatible with the App Store is the requirement that someone receiving the product be able to redistribute that product to someone else. Since all apps in the App Store have Digital Rights Management (DRM) applied to them, users are unable to directly share the apps with others. Because a developer is unable to give that right to the user, they are effectively barred from distributing their app through the App Store.

Just yesterday, this became an issue on the App Store as the popular app VLC was pulled by Apple due to complaints that it's presence on the App Store was a violation of the GPL. Even though the app was available for free, and the source code to the app was made available at http://git.videolan.org/ for anyone that wanted it, the DRM issue overrules and thus, no VLC for you.

As a user this is disappointing since the app allows you to play media files that no other app on the App Store, or built into iOS devices can play. But as developer, this is extremely alarming and should cause you to audit all the libraries you've used to ensure none are covered by the GPL, else your apps could suffer the same fate as VLC.

What's most disturbing about all of this, is that it seems as though, this is really just a silly technicality. The fact that a user is unable to share the exact binary they downloaded from Apple doesn't mean they can't share the app with a friend. The user can share the app just by sharing the iTunes link to the download (which is free), or alternately, provide the full source code. So the spirit of the GPL is still possible with the App Store, just not the letter of the GPL.

It would be possible for the Free Software Foundation (creators of the GPL) to release an updated version of the license to make it compatible with the App Store, but since they feel that all DRM is inherently evil, it's unlikely that will ever happen. Arguably, it might be possible for Apple to make DRM optional for certain apps, but that would require fundamental changes to the iOS architecture and might introduce security holes that make it easier to pirate apps. I don't see either side making any compromises here so your best bet, is to just avoid all GPL code and find (or create) alternatives that are licensed under the Apache, BSD, or MIT licenses.

Discussion - Reducing App Store Piracy

This week, I'd like to throw something out for discussion. These are ideas that have come up in the course of real life discussions with other app developers. I need to preface this by saying that I have not implemented any of these ideas, and that I'm only putting the ideas out there to encourage you to think about and discuss alternatives.

It surprises me that in an ecosystem of 99 cent apps, piracy rates are still incredibly high. For this article the assumption is that as an iOS developer, you are able to detect at runtime, if your app has been pirated. There are multiple ways to do this, but the technical details of which are not required for this article so I'm going to skip them.

One of the requirements to install a pirated version of an iOS app on a device, is that the device needs to be jailbroken. If the device has been jailbroken, then certain security features will have been disabled. This makes it possible for apps to do things on the device, that Apple doesn't allow under normal circumstances.

This means, that a normally well behaved app, could be made to 'go rogue' when it detects that it has been stolen by the user. For instance, a stolen version of your app, could make a phone call to a charge per call phone number that charges the cost of the app to the users phone bill, therefore recovering the cost of the app for the developer. Similarly, the app could send a premium text message out, also charging the cost to the user's phone bill.

Now, I'm not proposing that you actually build this into your apps, as it's almost definitely the wrong way to go about building up your business1. As iOS developers, is it not part of our job to educate users on the dangers of using pirated apps? The above ideas can be used in the same way we've warned against pirating desktop apps due to the dangers of viruses and other malware. Common users should be encouraged to live in Apple's walled garden as is indeed a great place for users and developers to be.

1 A user that has pirated your app is still a potential customer and needs to be treated as such.

Discussion - Apple's App Store Policy Against Name Squatting

Today I'd like to discuss Apple's recent policy change with regard to app name squatting. If you're unaware of the policy, Apple says you must now submit your binary for an app within 120 days of reserving the name. If you don't submit a binary, you'll receive warning emails with 30 and 7 days remaining in your 120 day grace period. At the end of 120 days, if you still haven't submitted a binary, the app is automatically deleted by iTunes Connect and you're forbidden from using that app name again in the future.

The idea behind this policy is likely to prevent the ridiculous name squatting environment the exists with domain names. Personally, I don't believe that Apple has come up with a great solution to the problem though. 120 days isn't long enough for an independent developer working in his/her spare time to create a good quality app. And there's a known work around anyway1, that just forces you to do an extra 5-10 minutes work per name to reset the 120 day count. So what has been accomplished? It makes it awkward for someone to register a tonne of app names and just sit on them, since the 5-10 minutes add up fast.

I propose that there is a better solution. Instead of an arbitrary time period to submit an app, why doesn't Apple limit each developer account (that is, per $99 fee), to an arbitrary number of incomplete apps. Lets say 10 for example. In my hypothetical world, you can squat on up to 10 names for each $99/year. Making app names cost about $10/year, similar to domains. But the domain ecosystem is a disaster you say. Well the difference is that with domains, there's a whois database. Everyone can find out who has registered a domain, and contact them in order to generate a sale. There is no current way to find out who has reserved an app name, and thus no way to buy the rights to a name. Which means, there is no market for buying and selling app names; crazy domain name ecosystem averted. The only people reserving app names will be those who plan to use them.

I'm sure there are problems with my proposal also, or that you have an even better solution. Please add to this discussion here.

1 Rename the app that's about to expire to some gibberish, and recreate a new app with a new SKU and the original name you're reserving. Credit: Daniel Jalkut

Three’s Company: Multiple XCode Versions Living Together Peacefully

Apple is making fantastic progress with iOS. They continue to release firmware updates and betas at a frequent pace, and with each new release, developers must download and install a new version of XCode compatible with the new firmware. Managing the various versions and their associated firmwares can be a challenge. In this article I'm going to give you a couple of tips that will hopefully help.

It's not uncommon for developers to have multiple versions of XCode installed on the same system in order to support development/testing across a wide range of devices and firmware versions. XCode by default installs in /Developer, and it is common practice to install the latest beta version in /DeveloperBeta. Those testing XCode4 know its default install folder is /XCode4. Personally I found this to be messy and inadequate.

My preferred setup now is to create a directory /XCode and install each version of XCode underneath using the version info as it's base folder name. Currently this looks like this:

/XCode/3.2.3_4.0.2
/XCode/3.2.4_4.1b3
/XCode/xcode4_dp2

When installing XCode in this fashion, you’ll find that in some cases, you won’t be able to install/debug a build on one of your devices because of a mismatch in firmware versions. For example, the XCode 4 Developer Preview 2 was released before firmwares 4.0.2 for iPhone and 3.2.2 for iPad. So if you’ve updated your devices to those firmwares, you’re now unable to use XCode 4 to directly install or debug.

There is a simple fix however. You just need to create symlinks from one version of XCode to another. Assuming you’re using a layout similar to what I’ve detailed above. If you look under /XCode/xcode4_dp2/Platforms/iPhoneOS.platform/DeviceSupport you’ll see folders for each version of iOS that you can install/debug on. 3.2.2 and 4.0.2 are obviously missing. If you’ve installed the latest version of XCode 3 with support for those versions, you can make XCode 4 work with them.

In terminal, issue the following commands to create the required symlinks:

ln -s /XCode/3.2.3_4.0.2/Platforms/iPhoneOS.platform/DeviceSupport/4.0.2 \
	/XCode/xcode4_dp2/Platforms/iPhoneOS.platform/DeviceSupport
ln -s /XCode/3.2.3_4.0.2/Platforms/iPhoneOS.platform/DeviceSupport/3.2.2 \
	/XCode/xcode4_dp2/Platforms/iPhoneOS.platform/DeviceSupport

Similarly, If you’re using the 4.1 beta 3 firmware on your iPhone, you’ll need XCode 3.2.4_4.1b3 installed, and can then enable support for that firmware version in XCode4 also:

ln -s /XCode/3.2.4_4.1b3/Platforms/iPhoneOS.platform/DeviceSupport/4.1\ \(8B5097d\) \
	/XCode/xcode4_dp2/Platforms/iPhoneOS.platform/DeviceSupport

If you have your XCode versions installed under a different directory structure (/Develper, /DeveloperBeta, and /XCode4 etc), you’ll just need to tweak the above lines to point to the correct folders.

A similar trick can be used to allow you to use a base SDK of 3.1.3 or earlier, which is no longer included in the latest versions of XCode. Just create links under the /XCode/Platforms/iPhoneOS.platform/Developer/SDKs to an older XCode that does include support for the SDK you need.

If you have any tips to improve upon this, see an error in what I’ve described, or otherwise have anything else to contribute, please let me know.

Some Thoughts on Tweetie

Everybody else seems to have a blog, so I suppose Cerebral Gardens ought to have one too. This will be written by me, Dave Wood, the founder and developer.

The issue that has compelled me to finally write something is the whole Tweetie upgrade pricing issue. First, I want to say that I always try to look at an issue from every side, though I'm not always successful at that. In this case it's pretty easy, since I'm both a developer and a user. Next, I'll admit that while I did buy Tweetie v1 for iPhone, I barely use it. I use Tweetdeck most of the time because of it's quick account changes. And Twitterific the rest of the time. Bought Tweetie just because it worked during the twitpocalypse. I do use Tweetie on my Mac exclusively (with the ads 'cause I like the ads, very relevant to me).

So, the issue with Tweetie v2 being a fully new version with no upgrade path is a tough one. Largely because developer’s choices are limited due to some of the restrictions imposed by the App Store. Enough people are writing about how to fix the store, so I'm going to present options that may work with the current setup of the store.

There are two main problems to deal with here (again, not counting the App Store issues). 1. Users expectations for free updates. 2. Developers needs for a sustainable business.

Regarding users expectations for free upgrades: I believe there should be an expectation for free upgrades for some length of time whenever someone buys software. That time is up for discussion. With a $1500+ Adobe Suite purchase or even Apples $69 iWork product, I believe that time period should be at least a year. With a $3 app, perhaps it should only be 30 days. Although I believe current users are expecting longer. Anyone expecting, or offering, lifetime updates is insane. That's obviously not sustainable.

But the expectations do exist; whether or not they should. Developers (myself included) need to start setting customer expectations to match our plans. Stating what the upgrade path will be in our app description for instance. Just stating something like 'free upgrades for 30 days' etc. As I type that I have to laugh. With a 14-20 day review process 30 days doesn’t seem like enough. 3 months perhaps.

John Gruber of Daring Fireball fame (whom I read religiously) compared buying a $3 app to a $3 cup of coffee. I don’t believe this is a fair comparison. The only thing they have in common is their price. A cup of coffee is a physical item, actual water, coffee beans, cream and sugar maybe, as well as a paper cup, and a plastic top. A person has to physically take your order, prepare the coffee to your taste, and hand it to you. The whole process takes a minute or so of someone’s day. If 1000 coffee’s are ordered at the same time, it takes a lot of people to serve them up in the time expected by the customer. Compare this to apps. There’s nothing physical with an app, not even a disc with iPhone apps. Once the app is made (and of course there is a cost there), it doesn’t cost much more per app ordered. It costs the developer (almost) the same to develop the app whether it ends up selling 10 copies, or 10 million copies. A closer comparison would be comparing apps to songs. We pay about $1.29 per song now. Imagine the uproar if songs were $3 each. Heck, I think $1.29 is expensive, but I pay it.

Continuing to compare to music, consumer’s expectations have been set; people buy a song or CD and they’re done. They don’t expect to get the same song again in a new format if it comes out; people joke about how many times they’ve bought the White album. They don’t expect to get an extended, or remixed version of the song for free if one comes out. The point here is that expectations of users can be set to be very low.

Perhaps this is what Atebits is starting to do. They’re reducing the expectations of their users, drastically. Possibly too fast, hence the uproar.

I've read that Atebits has said they would like to offer an upgrade path that gives current users a discount, but that it's not an option with the App Store. Here's one way it can be done. (This isn't pretty and I wouldn't recommend anyone else do this; further below is an option that could work for the rest of us, with time to plan).

Set Tweetie v1 to $999. No sane person would buy it at that price. If they do, Apple will refund it, (and hopefully not ding you the full refund price as indicated in the dev contracts). Then submit v2 as two apps, an update for v1 users, and a new app for $3 with text in the description to set the users expectations correctly. V1 users get the update they expected and also have their expectations for future updates set to nil. When v3 is ready, v1 and v2 are removed from sa le, v3 is submitted as a new app.

Regarding problem 2, the sustainable business: obviously developers have to be paid for their work, and continue to be paid for building upon that work. I’m going to assume that the incredibly low app prices aren’t going away; even if most developers raise their prices, there will always be bottom feeding app spammers (such as Brighthouse Labs) selling crap for $0.99.  So Atebit’s plan to release Tweetie v2 as a separate app actually makes sense (after setting expectations correctly).  But lets take it a step further.

I believe that we should be selling our software based on the time the user will use the app. Pick a value you want to charge per month of use, in this case, we’ll go with $1/month. Adjust these rates when/if prices in the App Store in general change. If you estimate your app has a 1 month life for the user, perhaps a game that will become boring/be completed in that time period, charge $1. If it's good for 3 months, charge $3. If the app is like Tweetie and useful for a year, charge $12. From here, we get my other possible solution (for the rest of us).

This option requires planning, so it's too late for Tweetie. If you expect a 6 month life of the app from v1 release to v2 release, price the app at $6. Each month drop the price of the app by $1. When v2 comes out, price it at $6 again. V1 goes to free, but is unsupported, or could even seize to function if that’s made clear to the user before they buy it. This is giving your users a set cost per month of usage. It’s not perfect, but much closer to what people expect, and it's a sustainable business model for us developers. I believe this will work extremely well if lots of us start using the model. This could likely be implemented via in app purchasing, by charging for each month of usage.

I’m going to stop here, without a final conclusion since I consider this just part of an ongoing discussion. I look forward to continuing to discuss these issues further and working with others in the community to solve these problems (among others) while creating successful businesses.