While that can often be the case, the author here pointed out that the critical parts (image munging) were written from scratch and that things like analytics/crash-reporting were done using the platform tooling.
The justification for the former was that they only needed a small subset of available functionality (as opposed to 100+ filters they weren’t going to use), and for the latter that they were a small team (two people) that would’ve just been overwhelmed by the data.
It’s not really “easier said than done”, I venture, provided you take the time to understand your tools and your product needs.
Easier said than done…
While that can often be the case, the author here pointed out that the critical parts (image munging) were written from scratch and that things like analytics/crash-reporting were done using the platform tooling.
The justification for the former was that they only needed a small subset of available functionality (as opposed to 100+ filters they weren’t going to use), and for the latter that they were a small team (two people) that would’ve just been overwhelmed by the data.
It’s not really “easier said than done”, I venture, provided you take the time to understand your tools and your product needs.