Ben Dodson

Freelance iOS, Apple Watch, and Apple TV Developer

Building tools for Kylo Ben

I’ve been running my Kylo Ben website about video games since October 2016 and this year I decided to start doing a weekly update about gaming news and what I’ve been playing. Whilst it is fun to do, it is very time consuming as I need to collate interesting links I’ve found, my articles, podcasts, game releases, games I’ve played, and games I’ve purchased which means an average post will take between 1.5 to 2 hours to write. Being a developer means I can build my own digital tools to help me out and so last week I built a few little tools to help cut that time dramatically.

One of the bigger pieces of the weekly roundup is a list of interesting news that has happened in the world of video games. Initially I would copy and paste the URLs of interesting links I found and save them into the Notes app1 on iOS or Mac. This worked fine but it was a little clunky and getting the data back out took time as I’d need to open each one to see what it was and then add Markdown syntax to each URL I wanted to use. To solve this, I wrote an iOS app and a macOS app that would provide extensions for URLs allowing me to quickly save them to my database.

[app name] would like to access Apple Music

The iOS app is purely a blank view controller with a bundled share extension that looks a little like this2:

class ShareViewController: SLComposeServiceViewController {

    override func isContentValid() -> Bool {
        return true

    override func didSelectPost() {
        if let item = extensionContext?.inputItems.first as? NSExtensionItem, let itemProvider = item.attachments?.first as? NSItemProvider {
            if itemProvider.hasItemConformingToTypeIdentifier(kUTTypeURL as String) {
                itemProvider.loadItem(forTypeIdentifier: kUTTypeURL as String, options: nil, completionHandler: { (url, error) in
                    if let shareURL = url as? URL {
              "save", parameters: ["url": shareURL.absoluteString, "title": self.textView.text], onCompletion: { (error, response) in
                            if let error = error {
                                let controller = UIAlertController(title: "ERROR", message: "That didn't work: \(error.localizedDescription)", preferredStyle: .alert)
                                controller.addAction(UIAlertAction(title: "OK", style: .default, handler: { (action) in
                                    self.extensionContext!.completeRequest(returningItems: [], completionHandler: nil)
                                self.present(controller, animated: true, completion: nil)
                            } else {
                                self.extensionContext!.completeRequest(returningItems: [], completionHandler: nil)

    override func configurationItems() -> [Any]! {
        return []


This is paired with an NSExtensionActivationSupportsWebURLWithMaxCount entry in the Info.plist file so that it will activate whenever I try and share a URL anywhere within iOS. If I’m in an app reading an article that I want to save, I simply tap the share icon and then choose the Kylo Ben app from the list as shown in the screenshot above. The URL and title will then be sent to my server for retrieval later on.

[app name] would like to access Apple Music

I’ve written a number of Safari extensions in JavaScript before but El Capitan added the option to write native extension in Swift via a Safari Extension bundled with your macOS app. To avoid having to make AJAX calls in JavaScript3, I chose to build a simple macOS app with a Safari Extension that looks like this:

class SafariExtensionHandler: SFSafariExtensionHandler {
    override func toolbarItemClicked(in window: SFSafariWindow) {
        window.getActiveTab { (tab) in
            tab?.getActivePage(completionHandler: { (page) in
                page?.getPropertiesWithCompletionHandler({ (properties) in
                    if let properties = properties, let url = properties.url {
                        let title = properties.title ?? "Unknown Title"
              "save", parameters: ["url": url.absoluteString, "title": title], onCompletion: { (error, response) in
                            if let error = error {
                                NSLog("PROBLEM! \(error)")
                            } else {


When I click on the controller icon, the method above is called and the URL and title are sent to my server; once completed, the page reloads to show me it has been successful. I spent a long time trying to just get a simple alert to display on either success or failure but I couldn’t get it to work correctly. It is possible to interact with JavaScript and I was able to log to the console but any alert would silently fail. If anybody has any tips on that, I’d love to know how to improve it.

The basic template of my weekly update is the same every week and I used to use a number of custom MySQL queries to pull out the various information I needed and then write it up manually. Now that I have my links stored in my database, I decided to write a PHP script to generate as much of my update as possible so all I need to do is fill in some of the blanks that aren’t automatically provided (i.e. upcoming game release dates) and write my own thoughts and opinions around the news articles. I have a basic PHP script which runs a number of MySQL queries and then generates a Markdown document like this:



[Final Fantasy 15's PS4 Pro Update Out Now, Improves Frame Rate And More - GameSpot](

[New PlayStation 4 Pro patch for Final Fantasy XV makes it look worse | Ars Technica](

[This tiny Nintendo Switch feature is already making fans super happy - Polygon](

[Alto's Odyssey awaits, Summer 2017](

[never gonna give you up - What’s In the Box?2?! Take 2](

[Steam Community :: Group Announcements :: Orwell](

[Nintendo tag teams with John Cena for living room-inspired Switch demos - Polygon](

[Look What Mega Bloks Is Doing To Pokémon ](

[Pillars of Eternity 2 campaign clears $3 million - Polygon](

[Take a look at how itty-bitty the Nintendo Switch cartridge is - Polygon](

[Australia Is Coming To Civilization VI](

[Rocket League Original Minis toys expanding with light-up cars - Polygon](

[Hot and heavy Mass Effect pack comes to Cards Against Humanity - Polygon](

And finally, 

###My Posts
- Making the earth move with Stagehand — "I really like the premise of a "reverse platformer" but there simply isn't enough content to keep me coming back when it is stood next to _Tiny Wings_, _Alto's Adventure_, and _Super Mario Run_" [[link](]

- Podcast #xx: Title [[link]()]
- Another Podcast #xx: Title [[link]()]

###Upcoming Game Releases
- _Game Title #1_ (date - platforms) [[link]()]
- _Game Title #2_ (date - platforms) [[link]()]
- _Game Title #3_ (date - platforms) [[link]()]
- _Game Title #4_ (date - platforms) [[link]()]
- _Game Title #5_ (date - platforms) [[link]()]

###Gaming Time
This week I spent 9.6 hours playing six different games:

- **Stagehand** (0.5hrs): Text...
- **Rocket League** (0.6hrs): Text...
- **Pokémon Moon** (0.7hrs): Text...
- **Forza Horizon 3** (1.1hrs): Text...
- **SteamWorld Heist** (2.8hrs): Text...
- **Night in the Woods** (3.9hrs): Text...

This week I added 2 new games to my library: _Crusader Kings II_, _Night in the Woods_.

Details on games I'm planning on playing this week...

Until next time, have a great week!


_Did you enjoy this weekly roundup? Make sure you don't miss one by subscribing to [Kylo Ben Weekly]( - it's this post in email form every Monday!_

The news URLs are simply pulled from the database and wrapped up so that each link uses the title of the page as provided by the macOS and iOS extensions. I will nearly always change the link title (as it’ll be part of a sentence) but it allows me to quickly see what an article is about without needing to open it up and re-read it. The “my posts” section requires no editing at all as it pulls the title, link, and a pull quote directly from the articles I’ve published in the previous week. The podcasts and upcoming game release sections can’t be automatically populated (yet) so I just use placeholder text for these to reduce the amount of effort required. The final section on my gaming time uses a number of queries to get the exact amount of time I’ve spent playing in the past week, adds placeholders for each game so I can write about them, and then lists out any new games I’ve added to my library; all of this is thanks to some scripts I wrote a while back that scrape my Steam and Xbox One libraries to track changes and allow me to render a page showing my gaming time for the past few months.

Once I’ve finished writing, the Markdown file is uploaded to my server and the weekly update will then appear on the website. I then use Byword’s “copy as HTML” feature to generate a HTML version and use that with Mailchimp to write and send out the email version of the update.

With these tools, I can now write my weekly update pretty quickly and only have to focus on what I want to say rather than spending time on copying, pasting, and formatting. If you’re interested in video games, sign up to the weekly email as it is the best way to get a digest of what has been happening over the past week as well as seeing what new games are arriving.

  1. I could have used a service like Pocket to do this but then I’d have to either use two Pocket accounts or fill my personal account with links that I don’t want to read later. ↩︎

  2. This is not what I would call production code quality so don’t just wildly copy and paste this into an app or you’ll likely regret it. Works well enough for my own personal use though! ↩︎

  3. Which is a nightmare when you start hitting cross domain restrictions. ↩︎

Proposal for an Erase Data Passcode

Last month, US-born NASA scientist Sidd Bikkannavar was detained by Customs and Border Patrol agents and told he would not be released until he gave the agents the passcode to his phone. They then took his phone (containing sensitive information from NASA) for 30 minutes before returning it and letting him go. He doesn’t know what information was taken at that point although popular consensus is that the entire device could be cloned within that time period.

Many articles have been written about this but the one that caught my eye was by Quincy Larson of freeCodeCamp entitled “I’ll never bring my phone on an international flight again. Neither should you.

When you travel internationally, you should leave your mobile phone and laptop at home. You can rent phones at most international airports that include data plans.

If you have family overseas, you can buy a second phone and laptop and leave them there at their home.

If you’re an employer, you can create a policy that your employees are not to bring devices with them during international travel. You can then issue them “loaner” laptops and phones once they enter the country.

Since most of our private data is stored in the cloud — and not on individual devices — you could also reset your phone to its factory settings before boarding an international flight. This process will also delete the keys necessary to unencrypt any residual data on your phone (iOS and Android fully encrypt your data).

This way, you could bring your physical phone with you, then reinstall apps and re-authenticate with them once you’ve arrived. If you’re asked to hand over your unlocked phone at the border, there won’t be any personal data on it. All your data will be safe behind the world-class security that Facebook, Google, Apple, Signal, and all these other companies use.

Is all this inconvenient? Absolutely. But it’s the only sane course of action when you consider the gravity of your data falling into the wrong hands.

I’ve seen similar responses on Twitter including one that you should use a burner phone with a different sim. This is all massively inconvenient, even if you follow the “wipe everything and reinstall once you’ve landed” method; bear in mind that the average iPhone takes hours to re-download all of its data1 at a point when you likely need to get maps, book transport, etc.

My suggestion is much simpler; Apple (and other handset manufacturers) should introduce an Erase Data Passcode. This would be a user-defined passcode2 that when entered immediately performs a secure wipe of the device in a similar way in which the existing “Erase Data” option works3. It would be expected that the device would disable power-off options during the secure wipe so that the only way to stop it would be to remove the battery (which in most circumstances would take considerable time at which point the data would be erased).

This is a solution that would also work in other cases such as theft, muggings, or a jealous partner. Whilst Apple have long had the option to remotely wipe your device via this has become far less easy to do quickly if you have 2-Factor Authentication enabled4 as you may not have access to your own devices.

I’ve filed a Radar on this issue (rdar://30553231) and would urge any other Apple customers that deem this to be a good idea to duplicate it. Apple goes to extraordinary lengths to protect user data and fight for the privacy of its customers but all of that is pointless if you are compelled to give up the keys to your device5. It is also pointless to have such powerful devices if we need to reset them every time we travel.

  1. This is especially true if you are data roaming as you usually get the slower speeds not to mention that airports generally have congested networks due to the volume of people. Finally, iOS 10 does a load of additional stuff during the first few days of a new device (like Machine Learning on your entire Photos library) which will cause further battery drain / wear and tear on components. ↩︎

  2. And optional fingerprint for TouchID devices (i.e. my right thumb unlocks the phone, left thumb wipes it) ↩︎

  3. This is an option within Settings > Touch ID & Passcode that will trigger an automatic secure wipe of the device if your passcode has been entered incorrectly ten times. I’ve always wanted an option to reduce this to three times. ↩︎

  4. You have 2-Factor Authentication enabled, right? No!?! Go do that now. ↩︎

  5. As usual, XKCD sums this up nicely. ↩︎

The Checked Shirt #1 - Lost AirPods, iOS 10.3 beta, App Store changes, and Invoicing

I’m happy to announce a new podcast I’m doing fortnightly with Jason Kneen; The Checked Shirt.

Every fortnight we’ll be producing a 1-2 hour show around the topics of freelance life, technology (specifically Apple), and gaming. Our first episode is available now in which we discuss the AirPods (and how easy it is to lose them), the new changes in iOS 10.3, the ability for developers to leave reviews on the App Store, invoices with Cushion, and lots of other fun stories and anecdotes.

You can get The Checked Shirt from these fine outlets:

Don’t forget to leave a review on iTunes and follow us on Twitter via @thecheckedshirt.

Syncing Apple Music with Spotify

Before Apple Music launched in April 2015 I was a longtime Spotify user and subscriber. I maintained a playlist I affectionately called Ben Dodson’s Definitive Hits Collection which contained nearly 45 hours of songs I thought were particularly good1. On most Tuesday nights, my friend and podcast co-host John Wordsworth and I play a few rounds of Rocket League and we will regularly have the Definitive Hits on whilst we play. There are two issues with this:

  1. As I use Apple Music now, I don’t pay for a Spotify premium account and so I have to put up with adverts (which are utterly terrible).
  2. They aren’t in sync so we might be humming (or badly singing) along to a song that the other person isn’t listening to.

Now I could just recreate the playlist in Apple Music to solve the Spotify ads issue but we still wouldn’t be in sync. As we’re both developers, we decided to remedy this problem with a fairly convoluted solution…

The basic idea is that John acts as the host with the playlist on Spotify (on macOS) playing into his headphones. He has written an app that checks if the track has changed and, if it has, sends the track information to my server. I then use the iTunes Search API to look up the song and find the correct identifier which is then sent to an app on my iPhone via push notification to start the song playing on Apple Music.

I’ll run through each piece and go over the challenges that were encountered.

Retrieving track details from Spotify

I hadn’t heard of it before but Apple has provided a tool called Scripting Bridge since macOS 10.5 which allows you to interface with AppleScript from other programming languages such as Python and Ruby. With this, John was able to write an app that polls Spotify regularly2 to see if the track has changed. If a change is detected, the title, artist name, and album name are all sent to my server so I can begin the process of matching the song on iTunes. In future, we may add more information (track number on the album, duration, etc) in order to try and match better but this is working well enough currently.

Finding a track on Apple Music

The next step is for the server to take the information that has been sent and use the iTunes Search API to try and find a match. This is fairly straightforward and a first draft would send a request like this:

Unfortunately the iTunes API does not allow you to search multiple terms (i.e. artist=rick+astley title=never+gonna+give+you+up) so everything has to be concatenated together which leads to an issue; sometimes the song you expect is not the one you get. For example, consider the song She Looks So Perfect by “5 Seconds of Summer”; If you search for this, the first result on the iTunes API will actually be a the “Ash Demo Vocal” version of the song which is not the one we want. To resolve this, we started sending the album information (in addition to title and artist name) so I could match that manually by iterating through the results; I then only choose the first result if there isn’t a song in the list with the same title and album name.

The next issue I encountered involves the Romanization of Belarusian; the track Solayoh is rightly attributed to Alyona Lanskaya on Spotify but Apple Music uses Romanization so it becomes Alena Lanskaya. If you search term=solayoh+alyona+lanskaya then you get no results. To fix this issue, if no results are returned from the iTunes API then I then do a search for the title alone and return the first result as that works in 99% of cases.

The final issue on the API side revolves around remastered tracks. The song A Horse with No Name is listed as A Horse with No Name - 2006 Remastered Version on Spotify but Apple Music doesn’t include that suffix even though they have the exact same version of the album. To fix this, if there are no results returned (again) then I split the string by non-alphanumeric characters and just try the first part in the lookup. Again, this seems to work in 99% of cases.

Once I have a track, I take the identifier and send it within a push notification along with the server time before I started making API calls (you’ll see why shortly). I use a silent push notification via the content_available flag as I want to wake the app up and run some code but not actually display anything to the end user.

The iOS App

The final piece of the puzzle is an iOS app with a fairly minimal interface3

The key thing for the iOS app to do is to play the track that comes from the push notification. This is fairly easy with an MPMusicPlayerController but we run into problems when the app is in the background as whilst the app will wake up from the silent push it isn’t allowed to play music.

That said, we can enable the background audio capability that allows us to control audio from the background but it only works if audio is already playing. To remedy this, I play a 5 minute track from the album “Silent Tracks of Various Useful Lengths” (id #366737838) on repeat so that the app is continuously playing music… it just happens to be silent music4.

Once a silent push is received, it starts to play the track but it also adds the 5 minute silent track to the queue. This is important as it prevents the background audio from terminating should I have a different length of music to Spotify or if a push is delayed due to network reasons. In essence, a normal track will be played followed by a track of silence whilst it waits for the next notification.

The final issue to solve is one of latency; there is a lot of latency inherent in this setup as we are polling Spotify, sending data to a server, doing one or more lookups against the iTunes API, relying on a push notification, and finally buffering the song in Apple Music! In order to keep us roughly in sync, the app will connect with my server when enabled and fetch the server time so that it can keep time5. When the push notification comes in, it contains a timestamp from when the server was first hit by the macOS app and I can then calculate the offset in order to skip into the track a bit.

For example, lets suppose John starts listening to C’est La Vie by B*Witched6 and his app hits my server at 1485359762 seconds from the unix epoch. This is recorded and sent in the push notification along with ID #298026101. If that process takes 3 seconds, then the iPhone app will know the server time is now 1485359765 and can work out that it needs to skip forward 3 seconds in the song in order to keep me in sync.

Amazingly, this crazy system actually works and we are able to have our playlist synced and ad free on two completely separate streaming services. I built my portion of the project as an iOS app as Windows does not have access to Apple Music yet I play Rocket League on the PC; in order to actually hear the audio, I wear a single AirPod in my right ear underneath my Turtle Beach X10 Headset so I can hear the music but still get the audio from the game and Skype.

It was only after we’d got it working that we realised we could have just set up some form of streaming radio server but that likely wouldn’t have been as much fun…

  1. The actual criteria to add songs is simple; I either have to use the phrase “it’s a classic” to be able to describe the song or it has to be “catchy as f**k”. ↩︎

  2. An improvement would be to hook into some sort of notification so that the app can be told when Spotify changes track rather than polling every second but this works well enough for now. ↩︎

  3. The switch simply activates the app as I’m using various background modes and don’t want my phone to randomly start playing music if John is listening to Spotify whilst we aren’t gaming! ↩︎

  4. I was originally planning on using the track 4′33″ until I found the album of silent audio. ↩︎

  5. Originally I would get the timestamp and then start counting it up with an uptime C method. This had some issues when the device was in standby so I made it simpler and I just work out the offset between the system clock and the server date; then, when I want to know what time it would be on the server, I can just add the offset to the system clock. ↩︎

  6. It’s a classic. ↩︎

Some thoughts on Apple's new Alternative Tier A and B pricing strategy for apps

Apple increased their prices for apps in the UK yesterday due to the current changes in the GBP vs USD exchange rate. One of the interesting changes that came with this were two new “alternate tiers” which would allow developers to maintain the previous 79p price point or drop it down to 49p. I was asked about this by BBC News and gave this quote:

I don’t think many publishers will respond to that change. It’s just throwing money away and there’s no reason to give people in the UK a discount. I certainly won’t be discounting my own apps.

I stick by that today as I don’t see why developers should be penalised for Apple adjusting to economic conditions; it is a change in price that UK users will get used to in the same way they did when the lowest tier changed from 59p to 69p and then again to 79p. This is especially true bearing in mind that the US App Store doesn’t have sales tax included whereas the UK does have VAT (so at current exchange rates a $0.99 app would work out around £0.96).

The key thing though, and something I didn’t clock onto fully until looking again at the charts today, is that Alternative Price Tier A and B doesn’t apply just to the UK; they change prices in many places such as China, Australia, and Canada but maintain them in other places such as Japan and Denmark. If you went for Tier A you’d be reducing the price of your app by 50% in the UK but reducing it by a whopping 83% in China! There are also a few odd choices such as Canada’s pricing being $1.39 at Tier 1 (which is $0.99 in the USA) but it drops to $0.99 for both the A and B Alternate Tiers.

If you are a worldwide developer wondering about which tier to choose, I would urge you to stick with the basic Tier 1 pricing. If you choose something like Alternate Tier B to maintain the old price for the UK, you could be dropping prices dramatically in other regions as well. I’m not sure why Apple decided to introduce these tiers but they are not helpful to developers, especially at a time when we should be trying to get away from the bargain basement $0.99 apps that are so prevelant.

If Apple want to reduce prices further, perhaps they should start by lowering their 30% cut?

Sporta 2.0

I’m pleased to announce the release of a new client app I worked on late last year: Sporta.

Sporta is an app to keep football fans up to date with the latest matches via local notifications in advance of a game. You simply select which teams you are interested in from the Premier League, Sky Bet Championship, The FA Cup, the Scottish Premiership, or international matches featuring England, Scotland, Wales, or Ireland. You can then choose a reminder time prior to the match (from 5 minutes up to 1 week) and the app will automatically send you a local notification to remind you about the match with details of kick off time, the teams involved, and the venue.

I was originally contacted by the creator to update the existing version of Sporta but I found that the code wasn’t suitable for me to work with due to a combination of out-dated code and offshore developers with poor documentation. Instead, I proposed that re-building the app from scratch with Swift 3 and modern iOS updates (such as AutoLayout to work on all sizes of iPhone and the new permissions systems for notifications introduced in iOS 10) would actually be faster than trying to untangle what was there. I was quickly able to recreate the app and scale it for different iPhone sizes whilst adding an intelligent system for keeping the app constantly up to date; the API for the match updates is called periodically using a combination of Background App Refresh and silent push notifications to ensure that data is always up to date and notifications will always fire at the right time even if the phone happens to be offline1. Finally, I used the Realm mobile database in order to store the match data and be able to quickly execute queries such as “fetch all fixtures that contain teams x, y, and z as either the home or away team”.

This was a really fun project to work on and I’m particularly pleased with how reliable the notifications are. If you are a football fan, you can download Sporta on the App Store.

  1. The naïve way of doing this would have been to use standard push notifications from a server at the appropriate times but these are not guaranteed to be delivered quickly or at all. Also, you’d need to send data from the phone if you changed your team selection or notification period whereas with the local database system I created you can do that entirely offline and it’ll just work (as I store all upcoming fixtures locally on the device). ↩︎

Booktrim for iPhone

I’m very pleased to announce the launch of a recent client app I developed; Booktrim.

Booktrim is a hassle-free app that allows customers to quickly browse barbershops, book a haircut, and pay right from their iPhone. I built the app in Swift and made use of AutoLayout to ensure that everything scales beautifully across the various iPhone Sizes and I worked closely with Trim Ventures’ API developer in order to ensure best practices. I was also instrumental in the design of the application and making sure that the user experience was optimised for making quick appointments on the go.

Once the app was completed, I was asked to build TRIMbook, an app for barbers which allows them to manage their appointments and availability. Once again, the app was built with Swift and I was key in both designing the UI and ensuring the API was being used efficiently. I worked closely with the Trim Ventures team to publish both apps on the App Store and also to run wider beta tests via TestFlight prior to launch.

This was an interesting project to work on and I think the end product works well. You can download Booktrim from the App Store or learn more on their website.

Don't be the idea person

I’ve been a fan of Dilbert since the mid-‘90s1 and this week a lot of my friends (and some clients) sent me this strip from November 13th:

Like most things, it’s funny because it’s true. I get several of these types of pitches per week and they never fail to amuse me. I’ve written about this in detail previously but I still get people who believe they have the next big idea and that it is worth something.

Ideas are worth nothing. Absolutely nothing. It is implementation that is worth something.

The analogy that I always like to use is that of an architect and a builder. Most software developers are builders that take some detailed plans and create to that specification. The “idea person” should be an architect with a clear vision and enough knowledge to know what can and can’t be done. Unfortunately this is rarely the case. Your typical “build my app idea” guy will have no knowledge of what can and can’t be done and therefore no understanding of the time and complexity there is in turning it into a reality. They just assume that they were the first person to think of a life changing idea and that it will be easy money for whoever builds it. It would be like delivering your kids drawing of a house to a builder and expecting them to have it built, for free, by next week2.

If you want to be taken seriously, do your research. Speak to developers and find out what is and what isn’t possible as the vast number of ideas I receive just aren’t possible given the constraints of software on iOS3. Then, once you’ve understood the process, either pay a developer for a basic MVP4 or pitch to investors to get enough funding to build one.

Don’t ever be “the idea person” and don’t ever expect other people to work for you for free. The world does not work that way.

  1. I have around 14 compendium books of the strips; my favourite is entitled “Still Pumped From Using The Mouse” (a reference to a calendar they were making about software developers). ↩︎

  2. “Oh, and by the way, there is a pool at the back and another floor but I didn’t draw that part” ↩︎

  3. “Can you give me the details for Apple then as this should really be installed by default on every phone - it’s a billion dollar idea”. Genuine email I received when I pointed out an idea wasn’t possible. ↩︎

  4. “Minimum Viable Product”, basically a step above a prototype that you can use to gauge interest from consumers and investors. The bare minimum you need to launch. ↩︎

Updates to iTunes Artwork Finder

Last week, my iTunes Artwork Finder script stopped working due to some changes by Apple with regards to rate limiting. In the past, searching for something on my site would cause my server to send a request to Apple and get the data back but with so many people using it my server would get blocked within a few seconds.

After a brief shutdown1, I’m happy to say that my iTunes Artwork Finder is now back up and running again thanks to a few tweaks. Now, instead of sending all requests from my server, it works like this:

  1. When you enter a search term (i.e. “Fall out boy” albums in United Kingdom), that is sent to my server where I’ll generate the correct URL that is needed for Apple’s servers
  2. Your browser then takes that URL to make the request directly to Apple
  3. When the data is returned from Apple, the browser then sends it to my server for processing
  4. Results are then displayed!

This means that there are now 3 network requests instead of 1 but the crucial part is that all requests to Apple’s servers are now made from your own browser so the rate limiting shouldn’t be a problem2.

Whilst not an ideal solution, it does work and means I can keep everything running for a bit longer. I’m hopeful that Apple will alter how their rate limiting works as at the moment it seems a bit broken, especially with it being required for several apps with the new Apple Music APIs.

The code for the artwork finder is available on GitHub although this uses the old PHP request system rather than my new version as that should be more than good enough for personal usage3.

If you run into any problems, please get in touch.

  1. During which I received around 80 emails hoping that it would come back online - thanks! It’s nice to know that so many people use the site; I have no analytics on my website so seeing so many people get in touch made me realise how big it has become. ↩︎

  2. Unless you try and do around 50 searches in short succession in which case you’ll need to wait a bit before you can make more requests. ↩︎

  3. And if you want to do more than just personal usage you should speak to me first as I don’t really want people to create entire duplicates of my own project. ↩︎


Update (9th Nov 2016): I’ve updated this article to reflect the new name for the app, Vortech. Whilst submitting a v1.0.1 update (adding support for some extra figures), Apple took issue with the name Dimensions so I had to change it…

I’m very happy to announce that my latest self-published app, Vortech, is now available on the App Store.

Vortech is a character and ability tracker for the videogame LEGO Dimensions. It allows you to easily see all 52 currently available characters (waves 1-6) and 73 abilities and track them in your own collection. Whilst there are other tracking apps available (namely the official LEGO Dimensions app), these tend not to be as easy to use and they don’t allow you to cross-reference the special abilities and powers of characters.

I built the app as I would occassionally come across something that required a special ability and the game might show me a character with that ability. However, as I have a fairly large collection already, it is likely I have a character with the ability and just need to look up which one has it. For example, I might find an area in a level which requires the Sonar Smash ability and the game might show me the character Benny (from The LEGO Movie) as someone with this ability; with my app, I can quickly see who else has that abilty and find I have Marty McFly (from Back To The Future) who will do just as well!

The entire app is built using Swift 3.0 and I use Realm as a mobile database containing all of the character information that I scraped from the official LEGO Dimensions website1. I added a few nice touches such as Peek & Pop support for 3D Touch enabled devices and taptic feedback for the iPhone 7 models. The app is completely free and contains no in-app purchases, adverts, or analytics tracking – it doesn’t even require an internet connection!

You can download Vortech on the App Store and learn more about it on my Dodo Apps website.

If you find Vortech useful, I’m always grateful for a review on the App Store. If you run into any issues or have feature requests, you can contact me directly.

  1. I built a custom scraper using PHP to periodically check the LEGO website and extract the image, abilities, and other information about each character. This then sends me a Slack notification at which point I can regenerate a JSON file and image bundle which I then include within the app. I’d thought about doing push notifications and remote updates for the app but as this is just a free hobby app (and the information only updates every few months) I figured doing it through the App Store will be fine for now. ↩︎

« Older Entries