Hacker News new | ask | show | jobs
by _jal 2601 days ago
If you build on someone else's platform, the best case is you get to be a sharecropper and can make money as long as you don't make so much that your platform-betters get jealous.

The more typical case is this, where you get to spend your time and money doing real-world R&D and discovery of what works for them, for free.

If you're going to dance with a vampire, don't be surprised when it bites you.

2 comments

One way or the other aren't most of us building for someone else's platform anyway?. Mobile, desktop, browser...
You can build alongside someone else's long-term demonstrated strategy (developing Microsoft desktop software) much more securely than developing a feature in someone else's closed garden.

You can also spread your footprint. As just an easy-to-discuss example, Facebook and Twitter develop for many different someone else's platforms, by supporting multiple browsers [who in turn support multiple OSs], multiple mobile platforms, etc.

Developing Microsoft desktop software you're still building on someone else's platform and hoping they don't decide to alter the terms of the deal. Microsoft may be more forward thinking than Google, but companies change.
Absolutely. That risk is much lower than developing for the Alexa, Nest, Twitter, etc platforms. At some point, you’re forced to build on someone else’s platform(s), even if that platform is “Intel” or “AWS” as no one is doing the entire end-to-end value chain.
Or build on an open an open source stack. Don't like Intel? Switch to AMD. Don't like AWS? Switch to another cloud.

Proprietary platforms lure developers to their stack by making development easy. Learning an open source is typically more difficult, but the reward is greater freedom. Believing you're forced to build on proprietary technology is a fallacy.

This is the kind of thinking that leads startups to build their own autoscaling for their pre revenue CRUD app. Trying to roll your own cloud infrastructure will kill you far quicker than AWS shutting down your service.

And you will always be relying on someone else's proprietary tech - whether it's laptops or power stations or cloud infra. The trick is deciding then to outsource and when to build your own.

Open source only works as a business, when one can get money with some kind of subscription service, or selling hardware.
Agreed. What's funny is that people now think Microsoft is this warm and fuzzy thing. They used to be far worse than they are today. For example, Microsoft Excel is nothing less than Microsoft's successful attempt to destroy Lotus 1-2-3. If the owner of the platform thinks you're getting too big for your britches, you can bet they'll try to take your revenue.
A difference is that Microsoft (currently) can't render your install base useless. If Microsoft drops Windows (or core Windows APIs) your existing users can stil use your software. With Google or Twitter shutting down "cloud"/web APIs all is gone.
People seem to have forgotten that Microsoft's unofficial motto is "Where do we want you to go today?" They've made an empire out of cutting off competition by changing their ecosystem.
Microsoft is all about cloud and web now. Clients are supposed to be web browser (asp.net) or mobile (xamarin), and they plan to add java, objc and swift interop to better target android and ios.
Websites should not be tied to a specific browser. But, if you mean plugins’s then sure that’s a huge risk.

Mobile and desktop both have more risks than the web, but they also have vastly more dependency on their ecosystems.

One could say that while some have a platform strategy, others merely have an aggregator strategy

https://stratechery.com/2018/techs-two-philosophies/

Which is why we have organizations that fight hard to try to keep those as open as possible.
> If you build on someone else's platform,

But isn't that the height of specialization? What does a platform mean anyways if its only used for one thing and not by others?

It it somewhat unfortunate that things like this make it abundantly clear that such a reality is not possible.

Platform is nothing but a sales&marketing word. It means whatever the owner wants it to.
Google positions angular as a platform.
It is possible. We just have to prioritize open standards much more than we have been doing in the last decade.

History really repeats itself in this regard. First we get ourselves in a tight situation with lots of closed platforms, which is bad for everyone. Then someone comes along, spouting a new philosophy of openness (Stallman comes to mind). The philosophy takes hold and open technology flourishes for a while.

But then a huge corporation appears, offering to contribute to this new abundant ecosystem with great new things. By now, people are too relaxed and optimistic, so they readily accept this. Yet, little by little, the corporation exploits this, seizing more and more control, until we get right back where we started.