I think he’s being a bit myopic in his piece on why developers shouldn’t use, USE in his words, iCloud syncing. While it’s quite possible that developers would develop for more than just the iOS platform those that do concentrate on iOS have customers that will demand syncing via iOS. It’s built in to the platform and easy for customers. That isn’t to say there isn’t a broader issue here that Apple needs to address and that is the sad state of Core Data syncing making the rounds right now. But for developers who only used document-based sync it works quite well.
For developers who are developing for multiple platforms the issue is even worse and his suggestion to not rely on another company and roll your own seems pretty simplistic. If rolling your own sync service were so easy it wouldn’t have taken the Things guys two-plus years to get their service working. The Omni Group is another example of a company who has rolled their own and taken awhile to do so. Sync is not an easy problem to solve. I think I’d rather support a company, like Simperium, who is trying to solve the problem vs. trying to roll my own. Sure there is a risk there but I’m taking a risk just by spending time trying to develop software with no guarantee of success.
As an aside I hope Brett was kidding when he said:
You hope your app will be a hit. (If not, then quit writing it and choose something else.)
The application I am working on will never be a hit by the usual standards that word implies. I’m writing it to help people. Sometimes people do things for reasons other than money. Should I stop working on it because it won’t be a hit?