Editorial: webOS App Approval Too Bureaucratic or Just Right?

A really great post has found its way to reddit from Jamie Zawinski, who’s a big name in developer circles (known as jwz), about his rough ride in trying to get a free tip calculator into the webOS App Catalog. Since I’ve been playing with Bell’s Palm Pre for the last week and trying out homebrew applications, it really caught my eye. Here’s The Short Version of this particular programmer’s complaints, but I encourage you to read the whole thing if you’re packing a Palm Pre or developing on webOS.

Inability to natively install applications from outside the market
Requirement of a verified PayPal account and potential fees, even for free apps
Forfeiting rights to post submitted applications anywhere other than the App Catalog

Let me preface my reply by saying that I’m not a developer, and don’t pretend to fully empathize with their wide-ranging plights and challenges when dealing with manufacturers, carriers, and end-users. However, a lot of this post raised some flags.

If you don’t want to play by Palm’s rules to get into the App Catalog (however reasonable/unreasonable they may be), don’t whine when your only option left is homebrew; making apps “off the grid” is not a bad starting spot for someone learning the webOS platform. Palm is in dire financial traits, and they need to harness their app market in order to stay in business – that entails some level of quality control, dealing with PayPal, code tweaks, and all that jazz that might come off as personal affronts to the more quixotic developers out there. How do you think Elevation Partners would feel if, when asking about third-party software strategy, Palm said “Developers can do whatever the hell they want! Woo, open source!” and cracked open a Bud? Instead, Palm saw what a financially successful player, Apple, was doing right (taking control of the application stream) and doing wrong (making draconian app submission decisions) and built their strategy around that. Sure, it might not be the supposed open-source programming Nirvana found in Maemo 5 or even Android, but it’s a balanced approach that helps everyone win at least a little. If you’re writing free apps and you don’t want to jump through the hoops, homebrew is not that complicated for end users.

Enter dev mode
Get Quick Install and your recovery ROM (Sprint, Bell)
Install fileCoaster

There, you’re set. You have a second on-device app store that Palm doesn’t control. Go nuts. Is that really an insurmountable barrier for entry? Will an app really have that many fewer eyeballs on it than if it were in the App Catalog? And if the software is free anyway, why would the creator care?

If developers want to support the webOS ecosystem with a free app, or make a buck with a premium one, they should be ready to make some compromises, including small code changes for the sake of QA. It’s not like Palm is turning into a bunch of Nazis; they aren’t taking shots at the homebrew scene, like Apple does with jailbreakers – in fact, Palm encourages the homebrew scene. The exclusive rights to the code is simply Palm’s way of ensuring the official market remains valuable. Why would anyone use the App Catalog if the homebrew scene had all of the same apps plus others that didn’t make it though the hoops? Sure, end users would get their apps, and devs could do whatever they wanted, but Palm wouldn’t get their cut, which ultimately is bad news for the whole platform and anyone developing on it.

As for the $99 annual developer fee, someone has to handle the submissions; even if your app is free, that someone answering the e-mails has to get paid. If you can’t be bothered to manage a ubiquitous and widely-accepted payment method like PayPal to handle that tiny bit of bureaucracy, then why should they be bothered to look at your submission? The thing that probably set me off most on jwz’s post was the insinuation that Palm was dead because they’re abusing webOS developers like a bunch of naughty puppies. First off, the phone has been out three months, and as I’m sure you can imagine, it takes a lot of manpower to make an app store ready for primetime – calling the time of death at this point is ridiculous. Secondly, this is the harshest amount of discontent I’ve heard from a webOS developer so far, and I’m tempted to think the majority are still willing to stick with Palm through the official App Catalog launch. Let’s hear from other developers who have tried their hand at the submission process, and see who has been satisfied (or at least understanding), and who has been pushed to another mobile platform. Feel free to comment with your own personal experiences.

About The Author

Simon Sage

Simon Sage’s education largely surrounded writing, technology and online community, leading him to begin his blogging career at www.BlackBerryCool.com and to quickly discover a vibrant and active community surrounding BlackBerry and mobile technology. In exploring RIM’s platform, he has learned what enterprises are looking for in mobility as well as what makes the innocuous BlackBerry so appealing to them. Recently Simon’s been covering RIM’s gradual move into an already-crowded consumer market, and the impact of burgeoning challengers, such as the iPhone, as well as long-time leaders, like Nokia, on BlackBerry’s advancement. With plenty of content under his belt, Simon will be branching off a bit to see what other smartphone manufacturers are working on while still using BlackBerry as a barometer. At IntoMobile, you can count on his posts being even-handed, well-informed and thought-out.

Ben Zanin

Simon, I should state my position at the outset: I consider myself to be a user first and a coder second, but I definitely lean in that latter direction. You make a good argument in this editorial, but it’s founded on assumptions with which I strongly disagree, so I must reject your conclusion.

Please allow me to rebut.

The root of this whole kerfuffle is not about who gets to run code on the device, but who gets to make that decision. I think we both agree that mobile devices these days are computers; they’ve got different input/output options and dramatically different interfaces from our desktop machines, but they are computers nonetheless. When I use a computer, I expect that machine to execute the commands to which I point the processor’s instruction pointer. If it is not code that I have personally written, it is at least code that I have chosen to run on my behalf. Crucially, if I discover that the code is not doing what I want it to do, I expect to be able to halt that code, and – depending solely on legal restrictions – either fix it or remove it entirely and replace it with code that *does* represent my commands.

I would expect a servant to respond to a command to break a window, slash a tire, or set a house alight with disgust, outrage and refusal; but if I personally attempted to perform the same actions with a crowbar, a knife or a match, I would be beyond livid to be thwarted by a faulty tool.

A computer is not a servant. A computer is a tool, and the responsibility for the actions performed with that tool lie solely with the wielder, not with the toolmaker. An App Store (or any other software distribution system!) is not problematic except when it becomes the sole legitimate source of software for a device.

The problem is not that the toolsmiths at Palm made a mistake, or that they’re learning, or that they need the fee they charge (however small!) in order to stay afloat; it’s that they are selling a tool but behaving like it’s a loan. That’s carrier behaviour.

Palm knew this once, and I was one of the thousands upon thousands of device owners who loved them for it. For them or anyone else to claim that they have forgotten what being a computer operator means, or that they don’t know what they have taken away, does them and us a subtle but crucially important disservice.

They’re better than that. So are we.
Stefan Constantinescu

People also don’t remember that Apple didn’t even have apps for the first year of their device, so Palm is off to a nice start. Where will they be in a year? Anyone’s guess.
James M

Ben, I think you missed one fact that makes the Pre a bit different from pure PDAs: it’s also a phone which is a service. More importantly, several fundamental aspects of the Pre require network connectivity of some sort (Cell or WiFi)

Given the fact that it the device’s main purpose is to be a network connection, the Pre, and all Palm-approved software, has to be reliable. Many Treo users complained about how unstable their phone was when the issue was one or more 3rd party utilities that didn’t work quite right. That has a distinct negative impact on Palm.

The sideloading homebrew mechanism is, IMO, a rational middleground between the walled iPhone garden and the PalmOS free for all. Anyone who wants unapproved apps has to step through one very easy hoop (dev mode) and download the QuickInstall software package. That’s only one more step than the old Palm Desktop. But it means anyone who does it knows that they are loading unofficial software that hasn’t been vetted for stability and gives Palm a clear line between “official” and “unofficial” apps. I suspect that if non-Palm coders had not developed the Quick Installer package that some kind of IPKG installer would have “leaked” from Palm.

I think Palm is also reaping the benefit of an unofficial cadre of beta testers who use the quick installer to enable features in the code that haven’t been fully tested. I won’t say they were cunning enough to create the situation at the start but I honestly think that with the patch/feature enabling utilities out in the wild Palm is now intentionally adding disabled beta features to production updates to get feedback.
Ben Zanin

James M, I’ll thank you to extend me the courtesy of assuming I know the Palm Pre is both a device and a service delivery vehicle, and I promise to cheerfully and respectfully do no less for you.

Mobile telephones are starting to offer capabilities once available only on desktop microcomputers, and later on the pocket computers that PDAs became. For a variety of reasons, North American mobile phone users have become accustomed to a poor and underperforming selection of handsets – this is dismaying and deleterious to progress. Now that handsets are becoming portable computers, we must, we MUST expect at least as much of our new computers as the models they replace, or they are no replacements at all.

Palm-approved software must be reliable, you say, or it will reflect poorly on the device itself? I don’t want to dive too deeply into technical minutiae, but if Palm cannot build a reliable device on the foundation of a preemptive multitasking operating system running a modern task scheduler on a processor with a hardware memory management unit, then I have no reason to believe that their application reviewers are any more competent than their own system coders. Palm has a long history and a roster of terrific programmers, and we’re two generations past an operating system with a single shared memory space and extension authors who think nothing of hot-patching privileged interrupt handlers. The reliability argument is, not to put too fine a point on it, an ill-informed crock; and even if it weren’t, it is unreasonable and insulting to forbid users from exercising our own judgment. If an expertly staffed quality assurance process mated with an integrated catalog and payment system isn’t enough to make officially blessed software competitive with the products of hobbyists, one has to wonder just how much more stacking the deck requires.

Palm sure is reaping the benefits of an enthusiastic developer community, and they’re doing so while relegating them to second-class status with that “unofficial! unsupported!” disclaimer. If you serve me a hot dog and call it a steak, I must conclude either that you’re an idiot or that you think I am one. In either case, you have botched my meal, and I would be a fool to hand over my money or grant my trust. Palm is running short on both.

I care because I want Palm to succeed. I’ve loved Palm machines for six years or more. The restaurant went dark for years, but now they’ve returned! The decor is beautiful, the lighting is brilliant, the art is daring; rich scents waft from the kitchen; the silver dome is lifted… and THAT is no rare porterhouse.
James M

Ben, I think most people aren’t aware how different WebOS is from PalmOS. You can disagree with the need for cloud computing but any device that is a cloud device is going to have some form of execution controls to ensure the safety of the network connection. That’s why a lot of other phones use java for their applications; it has an innate sandbox that protects the underlying OS.

You’ll note that the Pre isn’t unlike the Android where non-market applications have to use a program derived from the SDK to install the package. If you want to publish for Android you have to pay a registration fee using a credit card, have a profile, and agree to their Ts&Cs, which includes various non-competes, no-jailbreaking, and guarantees against attacking the network.

I don’t hear a great hue and cry about the Android being a locked environment.