Ask HN: app generators, OK or evil? | Hacker News Reader

Ask HN: app generators, OK or evil?

12 pointsTichy12 comments

Ignoring the iPhone niche where they are not allowed to begin with, would you say app generators are acceptable?

The obvious issue is that they could be seen as spamming the app stores. But I tend to think that it just shows that discovery mechanisms in the app stores would have to improve. Even without app generators, there are already hundreds of thousands of apps, so they can't be searched manually anyway.

The web seems to do kind of fine with the issue. For example suppose I would generate an individual app for every free book on project Gutenberg. That would be a lot of apps. But at least on the web, it is possible to find and read individual Gutenberg books without too much of a problem (one book could be considered a page a ka "app on the web"). They are not drowned out because there are so many of them. (I don't plan to do the book thing, it was just an example).

The alternative to the app generator would of course be a "mother app" that bundles all the other apps. The example would be an ebook reader with a special interface to project Gutenberg. Businesses might not be satisfied with that solution, though, they probably want an individual app for their brand.

I'm not sure about the app generator side of things, but I am struggling with trying to work out whether it's best to have one app or many.

In our case this is mapping. Generally I see that we only need 1 map app, and that each slice of data should just be a layer for that map. This is the way I would've traditionally done it, and it's the way that I think is 'right'.

However people seem to search not for layers on an app but for a specific app that fills their very specific needs. An app that shows them locations of ATMs is more discoverable and useful to them than a plugin to a different app that supplies a layer to the generic mapping app.

People seem to prefer to have one app that does one thing well and does nothing more. Even if this means that they end up with 50 such apps on their device and a lot of functionality is duplicated in several of their apps (with mildly inconsistent UI).

So we're seeing in our own apps that multiple apps dealing with one small thing is more successful than fewer apps able to deal with multiple pieces of functionality.

Going back to your example, I'd rather have a Gutenberg app that allowed me to grab the books I wanted, but I suspect you'd see more success having apps that match how people would search for them, which I bet would be by author... a "Philip K. Dick compendium" for example and then a "James Joyce compendium", etc.

I still prefer the fewer richer apps, but I seem to be in a minority.

But your problem comes from the same thing that OP is getting at. You don't necessarily know that it's what users want. It's just that single-purpose apps are much easier to find, because of the under-developed infrastructure of the app marketplace. Search is a particularly bad part of it--for perspective, keyword search is a recent innovation!

And that's the heart of the problem: So much infrastructure has to be re-built around this app ecosystem. Especially with HTML5, apps and the App Store are fast approaching a lame, more unwieldy version of the Internet and web apps. Their only saving grace is the easy payment system.

So there we have the solution: Write an App that makes it easier to find the right Apps! =)

I'm assuming you're hinting at appsaurus there :p

that's an interesting observation, and I think you're right. from an engineering perspective we feel a multi-functional app is the more optimal solution. yet the consumer would rather have a dozen separate apps.

why? I think it's maybe down to the product message. a specific app can have a specific, clear, simple message.

E.g "indian restaurant finder" is definitive. it's extremely easy to understand, it's purpose is clear. and more importantly, when you need to find an indian restaurant - ahh! there's an app for that! it has that mindshare. but an app that is "search for everything" doesn't. when you need an indian restaurant it's much less likely to come to mind - and even if it does, it'll be behind the specialised one.

plus I think there's a trust/quality issue. if an app does just one, precise thing, you will believe it will do that very well. an app that does everything - well for any given search it just wont be as good, will it?

Not to mention that an app for an extremely narrow set of needs will probably have a simpler, more intuitive UI. With a complex application, it's easy for users to not even realize a feature is available.

For example, did you know Word 2007 can generate "Lorem Ipsum" filler paragraphs for you automatically? Most people, myself included, are more likely to go to http://lipsum.lipsum.com/ whenever we need filler text, even when we have Word open and running.

I also wonder about this with AR platforms like Layar, Junaio. They want you to develop augmented reality content for the platform, which can then be shown in the Layar app. The advantage is not immediately clear to me - why not just take some open source framework (perhaps to be implemented yet) for overlaying stuff over the camera image, and create my own app directly?

Could you turn this into a business model?

Write and maintain a generic finder app, as well as a generator for specific finder apps (eg. ATMs, Indian restaurants, etc).

Sell the generic app at a premium, but offer cheaper data for the specific layers. Sell the specific apps for a little more than the cost of the data for the generic apps.

I haven't really thought this through in detail. It may not work. Just a thought.

My opinion is that app generators are alright as long as you know what they're doing. It becomes a problem when you have it magically generating apps for you when you have no idea how they work. When something goes wrong, you'll likely be left scratching your head.

Won't app folders in OS 4 help solve this problem? Granted it's not going to allow you to keep 50 ebooks in one folder, but at least solve the problem where you may have multiple navigation apps for different regions.

I think they are ok, as long as the apps submitted meet some minimum criteria.

i.e. no things like "top 10 _______ images" apps

Clients have found many individual apps to sell better than a single multi topic app that can use multiple pieces of content.

No idea why. But that is the issue and the crux. More individual apps sell better.

Reply on news.ycombinator.com