Saber Forum

Hi,

This has been asked several times without being addressed.
Maybe with  the Obsidian sound board it was not meaningful but now with the Emerald Led driver it means a lot.
With an Android app we could change color on the go.

Is this already planned ? If not, would it be possible to develop such app ?

Thanks.

Unfortunately, this is not the best place to ask this question. With the exception of Deep and Ultra, no one here works for Ultrasabers. You are better off emailing them at [email protected].

I know this is not a place where US staff answers questions. But I would also like to understand the demand from the users. Is it only me ? Is it only a few of us who asks for this ?
Or are we a huge community looking for this type of mobile app ?

So, who's with me ?

while I admit it would be convenient, it would also be pretty difficult to do. first you've have to port the program to android, which is difficult enough, and then you'd have to find some way of connecting it. either you'd have a cable from your phone to your saber - and I've never heard of someone using a cable to attatch a peripheral to a phone, instead of a phone to a computer - or you'd have to go wireless, which would require a fairly major alteration to the Emerald driver, and one which could cost you quite a bit of battery life.

and this is before you consider how tricky it'll be to program the whole thing.

as someone who does programming and such, I can tell you that this would be fun, but pretty difficult to do.

as for whether or not the folk at US have plans to give it a go, that you'd have to ask directly.

hope I helped.

I am an android maniac. I have about 12 android devices in my household.
First of all, nowadays it is very common to connect usb devices to an android device.
Most of the new android devices support usb otg (host mode) and we can connect several devices to the usb port with host mode similar to a PC.
For example;
Usb disks
Usb connected projectors
Usb 3g modems
Usb keyboard
Usb mouse

From my 12 devices, only 2 of them do not support usb otg.

So there is no hardware problems.
Once a software is designed for android we can use it similar to the pc app.
I am not sure if the software can be pported as is. Most probably it has to be redesigned and simplyfied for a mobile os (android).
But it is not very difficult for a person who has already designed the pc app.

^^^ fair enough, I wasn't aware of that, though I'm not honestly all that surprised. give it a few years and phones'll be doing what computers did a few years ago.

again, while it might not be difficult, it could be fairly time consuming, and differences in programming languages can be an issue.

I'd say it could be done, and it wouldn't be extremely hard, but I'm not sure it's something you'd want to do without a fair chunk of time and/or resources.

like I said, I think it's a good idea, and it's certainly doable, but it wouldn't be a few spare afternoon's work, and it's ultimately down to US whether they do it or not.

good idea though.

man, android's really moving on, huh.

Ok. What  I am trying to understand is, are there many people like me who would like an android app for this process.
If we have demand Ultrasabers can address it in one of 2 ways;
- develop the android  app (not more than 1 week for someone who know the protocol used in pc app)
- release the protocol documentation so that any 3rd party developer can develop the app for us (I am sure there are android developers among us - ultrasabers customers)

Of course if there is no demand (or a few demand) then it is not feasible.

I'd b more interested in the program working on a Mac before we make the jump to phones

Mac users at least have a workaround, right ?

Yep, it's called Boot Camp.

Or get a new computer, lol.

I also looked and sadly don't see any 8 inch Windows tablets that have a full size USB port that would allow you to have that portable Windows device and jack your saber in to change colors.

well, you should be able to run it on a windows emulator.

there should be a few for Mac. I've never looked into it personally, but I'm sure there are some out there. same with linux, or any OS with a windows emulator.

"1 week" would be an extremely optimistic guess.  Obviously I don't know the code, but I'd guess that it depends heavily on Windows APIs.  The code probably isn't particularly x86-specific, but I don't know that for sure.  A "port" to Android would be essentially a complete rewrite.  I think it would be much easier to make a Windows RT port than an Android port, but I don't see either of those happening.

One thing I'd be very curious about is how narrow of a target Android is.  Windows has APIs in it that date back over a decade; if you developed something for Windows 98, there's a decent chance you could run it under Windows 7 or 8.  Likewise, hardware compatibility tends to be very good; hypothetically, you can run Windows 8 on a 13-year-old Pentium 4 (very slowly).  The same doesn't seem to be true for Android; inability to run current versions of the OS on 2010-era hardware has been widely publicized.  (I'm sorry, but if you're worse at supporting old hardware than Apple, you officially fail at supporting old hardware.)  I don't know about APIs in Android, so I don't know if deprecation and removal would be a concern.

In summary, I suspect that an Android app for Emerald and Obsidian would require both significant development effort and increased maintenance compared to the Windows version.  The gain as a user would be minor; personally, if I'm already carrying one or more sabers, I can find space for a 12" Windows laptop.  I don't think US would see much in the way of return on investment for this; I'd guess most people who want this app would have bought an Emerald and/or Obsidian saber anyway.

That leads me to question for you: how much would you be willing to pay for an Emerald/Obsidian app?

I am not a developer but I know a little about the Android SDK and I have written (very simple) network apps (for self use) for Android.

I never said that a port would be easy from Windows. I don't think so. Here's what I wrote on that:

But a new app for Android would not be that much trouble. It would be just about sending some commands (which we don't know) through the USB port.
A simple app would be just getting 4 values (1 for each led between 0-255)  from the user and sending appropriate commands for the Ultrasaber Emerald driver through the USB port. If I knew the commands used in the original windows app, I would be able to develop such simple app in less than a month. (well because I am not a skilled/experienced developer)

So, to your question:

As the Windows app is free, why pay for Android app ? If I am buying a 400$ item with a USB port, an extra payment for a mobile app is unnecessary.

I would pay 0 (zero) to Ultrasabers for such app because in my opinion this app is very easy to develop if you know the commands that should be sent through USB. And US has that knowledge.

But as they will spend some efforts, a symbolic payment of 1$ per user is considerable. (Most of the mobile apps are within the 1-5$ range)

On the other hand, if a 3rd party developer supplies such app, I could pay up to 10$

If US decide that they should not spend effort for such an app, they could really release their API documentation.

You're looking at things the wrong way.  The question is, "What benefit does Ultrasabers get from developing this app?"  They had to develop an app since they needed something to interface the saber with, and Windows is the obvious platform since about 90% of their buyers will have a Windows computer (and most of the rest have  a Mac that can run Windows).  You've just said they'd get little or no money directly from people purchasing an Android app.  I continue to assert that it wouldn't drive additional sales.  At a bare minimum of $25/hr and your one-week (40-hour) estimate, that's a $1000 outlay.  If we assume every Ultrasabers buyer has a smartphone (with the average nearly 50/50 split of Android and iOS), and every Android user buying an Ultrasaber also buys this app for $1, they'll have to sell almost 2000 Emerald sabers before they need to touch the code again in order to break even; since both of those are unlikely assumptions, the actual number will be much higher.  They've released a new Obsidian approximately once a year, and each new board has required a new version of the software; it's reasonable to expect something similar from Emerald, so those sales would have to be in a year's time.  I don't know how much business US does, but that seems like an awful lot of high-end sabers.

The flip side of the coin is "how many sales will US lose by not developing the app?"  Their profit margin is harder to estimate, but at a guess it's going to take them over a hundred people per year deciding not to buy a saber in order to cost them what the app development would cost.  (There's also the question of how many sales they'd lose by developing an Android app and not an iOS app; I suspect that most users will not appreciate the much higher technical difficulty of interfacing a USB device to iOS.  In other words, developing an Android app could cost them sales among iOS users.  Bear in mind that, on average, iOS users are younger and more affluent than Android users and thus more likely to buy $400+ lightsabers.)

You also haven't addressed my point about Android as a broad target.  The risk of "I got a Galaxy S5 and the Emerald app doesn't work with it and you need to fix it or I badmouth you on every forum I can find" is high enough to make each new phone and tablet release something that may need support (and thus incur more cost).  While doing this, they need to be sure not to break older versions, of course, for the same reason.  Ongoing support costs can add up surprisingly quickly (every hour of a cheap programmer's time is 25 more copies of the app to sell, and testing takes time).

Or they could consider it a trade secret and keep it proprietary to make it more difficult to reverse-engineer what's currently the most advanced LED driver on the market.  One of those options requires effort on their part; the other requires no effort on their part and could decrease competition and thus increase their sales.

I'm not saying that US will never develop an Android app; they could be doing it right now for all I know.  I would be surprised if it was the case, though, for the reasons given above.

Get a windows 8.1 tablet and problem solved. the new slates will load windows programs same as computer. They have cdrom and usb support etc... Android os is far too fragmented to keep up with, I mean there are hundreds of different devices to keep up with all with different drivers. Like my FIT BIT for example, works great on my galaxy s4 but not on most other phones due to the Bluetooth version.

IOS is also very limited in that aspect. It would be easier I think to integrate wireless capability in the electronics or a dongle internally so that your computer or tablet can modify it from a web interface. This would work via HTML5 or even something as simple as embedded java script. I would be willing to help on that!!!!

Actually was just thinking about it. Imagine a stage show where someone who is a jedi slowly turns to the dark side and his lightsaber changes color as the show goes on. The sound guys sitting on the set or off stage have each person on stage set in their laptops so they can alter color on the fly while play is going on. Imagine the cheers when a hard core sith suddenly changes his point of view and his blade lights up bright blue and he saves the day etc....

Having a wireless access with internal hardware and any wireless access point  you have a lot of options that an app just couldn't do. The rom needed is relatively small and you would just need to program it once for preferred networks. Then when ever they are within that wireless network range you can connect via phone/ tablet/ computer of any sort that supports HTML and be able to adjust color and or sound..........

Think about it.

well, infact I believe you are looking from a wrong point of view...

First of all, "more money" is not always the key to success.

I will not give up buying ultrasabers if they don't develop for Android devices ever. But the "extra customer satisfaction" they would gain by releasing an Android app is priceless.

I don't agree about the cost of creating such an app. They have the knowledge and they can build it easily. Also, I don't agree about the support cost for newer Android devices.
Like I said before, I have 12 Android devices and all the newer ones support USB OTG in the same way. I have different Android versions from 4.0 to 4.4 ; they all support the same USB disk, keyboard and mouse.
If you keep the app user interface (GUI) simple, it would work on all Android tablets and phones in near future. Of course there is a slight chance that Google might change the USB API on new versions. But developers always supply a list of "supported devices" for complex hardware projects. Ultrasabers can do the same for their app.

On the other hand, as we all know Ultrasabers do not sell cheap sabers. Their sabers are the best quality but they are not cheap. I've spent 420$ (except shipment) for a single Emerald saber. Only the price of the Emerald driver is 110$ (with the cyber monday deal that is -  original price is 140$)
I doubt that its cost to US is not over 50$ (just a board which controls 4 channels to 4 leds and US is purchasing it in large amounts)
If money is very important to them, they can increase the price of driver from 140$ to 150$. I wouldn't hesitate if my saber costed 430$instead of 420$...

In any case, I believe, the best choice for US would be releasing the API for the USB control of the driver if they don't already have knowledge on Android app development.

That's all my idea. I don't say it is all correct. There are lots of assumptions.
But with this thread I wanted to get opinion about demand for such an app. I got my answer; most of the users are not interested in changing the color of their saber on the go without the need for a PC.

FWIW, you could already do that with an RGB saber.  You couldn't do a continuous progression, but a continuous progression from blue or green to red looks a little odd to anybody who doesn't know how colors work.  (A blue-green transition is much more intuitive, but lacks the good-evil change you're going for.)  You probably could do a continuous progression with analog components if you wired them up yourself.  (I think you could do it with a potentiometer, a couple power transistors to minimize loss in the pot, and a two-or-more-color LED.)  I have an upcoming project related to this.

Incidentally, there's also an Anakin/Vader color-changing saber from Hasbro.  IIRC, one push of the button turns it on blue, a second turns it red, a third turns it off.  You could probably splice that board in without difficulty; it would be similar to wiring up an econosound mod, except with two power transistors and two colors on the LED.  (You might even be able to skip the transistors; if it's like the UFX boards, you get close to 1A total by using all of the leads from the blade LEDs.)  It's probably fairly wide, though, like the other Anakin and Vader soundboards, and I'm not sure which hilts could fit it without modification.

As I feel about you.  :)  I think we're going to have to agree to disagree.

I think it's a neat idea, but it's very much a niche application (especially when you bear in mind that there already is an Ultrasabers solution to changing colors on the fly--RGB sabers--although with only 7 different colors).  "Is there an Android app?" and "is there a Mac app?" have been asked repeatedly since Emerald was announced, but so far I think you're the only person to go past "is there an Android app?" to "how can we get an Android app"?  So, yes, I think there are people who would use it if it existed, but there is little actual demand per se.

All this said, I just learned that there are Windows-capable x86 emulators for Android.  They would have to be heinously slow, even on high-end hardware, but perhaps they'd be fast enough?  Cram XP on there, cross your fingers, give USB a thorough test (read-write tests on a good-size flash drive, perhaps), and then give it a go and hope you don't break your Emerald/Obsidian.  (I've read that Obsidian 3.0 is more resistant to breaking due to transfer problems than older versions were; presumably that's true for Emerald as well.)

I don't think an emulator would be a good idea for this particular usage.
Most of the time, emulators do not have all hardware supported.

But I would like to give it a try. I searched for Cram XP but couldn't find it. Can you post a URL ?

Btw, even if emulator works, it would require a tablet with X86 processor, which is not very common among Android devices.

"Cram" there was a verb.  I saw a couple of emulators out there; this one (https://play.google.com/store/apps/details?id=com.motioncoding.emulator&hl=en) seems like the most mature (using that as a relative term; the app is 9 months old and seems to be a side project).  That does appear to be a true emulator, i.e., running x86 apps on ARM hardware.  (The device pictured in their video is an HTC phone; I don't think HTC makes an Android-x86 phone.)  They mention it being built on QEMU, which does have the ability emulate a different processor architecture, so this is plausible.  (I'll also note that the not-very-impressive demo video shows it running Windows 95; considering that Win95 flew on 100MHz Pentiums with 64MB RAM, dynamic binary translation would still give acceptable performance on almost any Android device.  XP would be another matter...)

Anyway, as you've noted, the relevant hardware is the USB port; as long as they have USB support, it would work.

thank you. do you have an idea where I can find a XP image to run with this ?
is it possible to create my own image using an original XP CD ?
thx.

If you look at the reviews, Alboooza Afi mentions a piece of PC software.

well, I've installed Win98 on the emulator to see how it accesses USB port.
Unfortunately Emulator does not bind the USB port to the Win98 OS. When I connect a USB disk, it's accessible through Android but not Windows.

sorry...

Not sure why you would need to port the whole obsidian launcher to android. Aren't we really just needing an app to save and load different config files?

ok. but if you would like to do  it when you are mobile ?
no PC or notebook with you...

I get the attraction of being able to customise on the fly, but for a first version app I'd be happy enough with just being able to save and load config files. Maybe if it gets enough use it would encourage development of more functionality.

Sent from my GT-I9505 using Tapatalk

This is a very interesting thread.  8)

Ultimately, you will eventually find mobile solutions for configuration made by the mobile industry itself. There will be no need for us to create our own app for that with how fast technology is progressing.

Not to go off on a tangent, but I think it would be a very bad PR decision to work on an app for Android while ignoring iOS  :-\

agreed. for an initial app, that would be nice.

ULTRA that is why I suggested wireless connection built in to saber. This way the saber will need to be setup once in the rom on the obsidian soundboard. Attatch it to a wifi spot and give it an IP address. After that point any device can connect to the saber via WEB BROWSER and configure there. NO NEED FOR A STAND ALONE APP. Any device with a browser can connect and modify configuration file in the rom.

This technology is small and with a small increase in the price of the saber could be useful. I also agree ULTRA that the IOS vrs Android would alienate one or the other. By doing the way I suggested will allow even MAC's, Linux, UNIX, BeOS, Windows, and Chrome OS to make changes as long as they can be on the same network and use an internet browser. It is worth thinking about.

that's a good idea, but you'd either have to have an on/off switch for the wireless, or face the prospect of the battery drain that it would cause. I'm more software based, but I think a wireless transmitter opperating constantly with a reasonable signal strength would take a fair bit of your battery life. plus, I don't want someone else chaging my blade colour, so you'd probably need some security too.

While iOS support would be nice, I'm not sure how easy it would be.  The first problem is that right now the Obsidian launcher isn't even on Mac, so I don't know how hard it would be to program for the mobile iOS.  Also, Apple tends to be rather uptight it seems about apps it lets into its market and I don't know what the approval process is or how hard it is to design for iOS due to restrictions in the hardware/software.  Android on the other hand is rather open about apps created on the market and their devices have less restrictions on them.  Android would definitely be the better one to start with I would think, then try and move the idea to iOS and see how it goes.


The Wi-Fi idea is good, but it brings about a whole slew of problems.

Putting Wi-Fi into a saber requires yet another chip and more wires to connect and be an antenna.  I don't know if the metal casing would help or hinder a signal.  Then of course you have to deal with how do you control the on/off nature of the Wi-Fi and also its power draw.  Also, my phone doesn't do Ad-Hoc networks, which means you'd need to connect both to a hot spot like you said.  Problem with that is that most public hot spots I try and get on are either locked require you to authenticate by opening a browser and clicking "I agree".  Neither of these would work with a saber since you have no interface. Even if the network was completely open and you could your saber onto it, how would you get to it?  You wouldn't know the saber's IP.  Unless you had a generic web address that you could go to and it was able to look for saber devices on the same network and IP range as you so you could tap into your saber, you couldn't access it.

Programming would be no more difficult for iOS than for Android.  An app of this type probably wouldn't have difficulty getting into the App Store, either, from what I've seen.  The biggest problem with iOS would be if you wanted to use a wired connection, since iOS doesn't really support being the host for USB connections.  Maybe there's an API there that just isn't widely used, but to be honest, I doubt it.

Bluetooth would be better than Wi-Fi IMO.  The metal casing, however, would be a huge problem; it's basically a Faraday cage, and it's too big to be useful as an antenna in anything close to Bluetooth's or Wi-Fi's frequencies (based on my limited knowledge of antenna theory).

The battery drain could be a big issue.  Security is a relatively small one; a simple passcode system would suffice, since all you're really trying to do is deter what's essentially vandalism.  (As far as my security concerns go, the color of my lightsaber blade would be somewhat below my stock of paper clips in terms of priorities.)

My intention is to provide a alternative to a usb connection via a cell phone. One that has a possibility to work with any OS. Sure there are technical challenges but more so than programming a android and IOS app and then keeping them updated with os upgrades and maintain compatibility with the device?

Bluetooth is an option also but does not allow you to tap a web interface. Not to mention that you can use the hilt itself as an antenna. Sure wifi hotspots are iffy, my phone I can turn on a wifi spot for use, some providers do not allow it unless you pay. I also carry an external wifi spot for my kids to use with their ipods and laptops. It is an idea only, concept for further exploration to do with sabers.

We know that MAC computers don't offer compatibility with the software. In fact average mac owner buys a very expensive macintosh only to load windows on it. Kinda defeats the purpose to go mac....... not to mention Linux, chrome, beos, unix, etc... Most wireless routers have a web interface to program the routers and it is generic and not limited to any one operating system. I was merely trying to suggest we can work towards a common solution that will allow everyone to do what the app alone will do for only android.

I am sure there are electrical and computer engineers that can offer some input on overcoming weakness's. Don't just shoot it down... explore the possibility and strength and weakness's. Security can be hard set in the ROM. A person can carry with their equipment a programmed hotspot for personal use. Some can turn on phone or tablet wifi spot. Some could do ad-hoc with hard set IP address. It is worth a discussion.

Bluetooth would seem like a better idea than Wi-Fi for sure.  Range/signal dampening shouldn't be too bad if your device is right next to your saber.  It would be like any other BT device and would bypass the cable issue.  You'd need to a way to power on the bluetooth connection and tell it to start broadcasting.  Then you'd tell your mobile device to look for it and input some default code(0000, 1234, 9999, etc.) then you could be connected, do what you need, and shut off the BT again.  This would help with the battery issue and also anybody trying to play with your saber.

I can't speak for others, but I personally am not trying to shoot these ideas down. I agree that overall they're good ideas, and I agree they're worth discussion, I just prefer to outline the issue that we'd need to overcome to realize these ideas. after all, identifying a problem is the first step in fixing it.

well, wifi and bluetooth are all good ideas. But we are losing the focus point here. The idea is not to access the device wireless, it is to be able to connect while on the go.

adding a wireless or bluetooth module would increase:
1. hardware cost
2. battery  usage
3. space used
4. weight

we would not want that.
besides, access to IP network from any point is not  that easy. Most suitable way would be using Android's wifi network sharing if wifi is the case.
With that you wouldn'thave the access issue (captive portal etc.)  and you would have your saber on your own network, closed to public.
But in any case, accessing this network would still mean entering ssid and password information tothe board.

well, a wired (usb) connection is costless and easier to implement. (if we talk about Android)

For IOS, I don'thave much to say. I don't know the OS but I'm sure USB host mode is not easily achievable.

Yes, much more so.  See below.

Can you actually use the hilt itself as an antenna?  Perhaps you know something about antennas I don't, but I find it hard to believe a 12" tube with considerable mass could be used effectively to capture a 5"-wavelength signal at anything approaching a usable bit rate (let alone WiFi speeds).

The "average" Mac owner?  Really?  Of the Mac owners I support, I'm the only one who routinely uses VM software (nobody I know uses Boot Camp), and that's because I support both Mac OS and Windows and can that way carry both in one laptop.  At any rate, from a technical standpoint, developing Obsidian Launcher for Mac OS would be easier than doing it for Android or iOS; there's no Mac version of Obsidian Launcher simply because US didn't choose to write it for that platform.  (A decision I wholeheartedly agree with; developing for a single OS that covers 90% of the computing market is by far the wisest choice for a small company which is not primarily focused on writing software.)  Developing for Android or/and iOS runs into the same problem--it only covers part of the market--which is of course why you suggested a Web-based interface.  However...

You're also suggesting needing a full-fledged embedded computer to run an LED driver.  Wireless routers can have a web interface because they already have hardware in place to handle significant processing and data throughput, and enough of an OS to handle a TCP/IP stack.  I doubt either Emerald or Obsidian have hardware of that type, since they don't need it for what they do.  What you'd need would be a small computer like a Raspberry Pi, but then you're talking about something 3.3"x2.2" and probably $50 (once you include the WiFi connection), with considerable power drain.  (Probably around 700mA at 5V; in other words, 3.5W, which is close to the 4W draw of the LED itself.)  You could probably make something similar but small enough to fit in a hilt (since you don't really need all the Pi's interfaces), but the power draw isn't going to change, and the cost would skyrocket.

The only reasonable option would be to have the saber itself act as an access point; I've seen some devices, e.g. projectors, that do that.  Anything else and you run into problems with getting the saber associated to a network.  Devices of that type do function in infrastructure mode, so a tablet or phone would be willing to connect even if it doesn't like ad hoc networks.  You're still talking about a lot of complexity for not a lot of gain IMO.

Ditto.  It's one thing to say "wouldn't it be nice if we could..."  It's another to analyze the problems and devise solutions.  Whether researching or engineering, the idea is to take an idea and try to find out what's wrong with it, not to propose an idea and stick with it against all arguments.  In other words, shooting it down is the whole point; it's when you have an idea that can't be shot down that you're on to something.

Exactly.  WiFi is not the right way to go.  Bluetooth is a neat idea, but really USB gets you 90% of what you want for 50% of the effort.

Unfortunately, this is also correct.  I agree with Ultra that developing for Android while ignoring iOS is, to put it mildly, not ideal; however, actually interfacing the phone to the saber is more of a challenge with iOS.  That's the biggest advantage to Bluetooth versus USB; the wireless aspect is neat, but the ability to support both iOS and Android with one interface is perhaps the most important.  Too bad it spikes the price and complexity.

I continue to maintain that, fundamentally, a mobile Obsidian/Emerald configuration tool is both unnecessary and fraught with complexity.  However, if you were going to make one, by far the easiest is (as ilkeraktuna initially proposed) an Android app communicating over USB.  You already have a USB device interface on the saber, and a USB host interface on the tablet/phone/whatever.

Here's one other crazy idea.  An interface that used to be popular on iDevices was to plug into the headphone jack and use sound modulation to communicate.  You could do that; it would take a little extra capability on the Emerald/Obsidian, but nothing that took real processing power.  (It's basically a modem.  Obsidian's chip might already have this functionality, actually, used for a different purpose.)  US would need to sell cables that were wired differently from each other, because Androids and iDevices have two of the connectors swapped (mic and ground, IIRC), but otherwise there's no hardware interface challenge.  Then you're just developing an app to play and record sounds, and that's an easy target to hit in terms of platform; it probably wouldn't even be hard to port between Android and iOS at that point.  I don't know how much data needs to be transferred, but I would expect it to be small enough that this method's limited transfer rate wouldn't be a deal-killer.

A variation of the last idea would be to build an interface board to convert from audio link to USB, and use that for an iOS interface.  This adds cost and complexity over Android and USB OTG, but could be useful to keep the iOS fanboy mobs off the doorstep.  (Point me at the first person with a $400 iPhone and a $500 saber who complains about spending an extra $50 on the iPhone adaptor, and give me an alibi.)  You could let Android use this interface as well, for devices that don't support USB OTG, although supporting both would make the software more complex.