On the PHP.cc's site has an article that looks at dependencies in disguise based on a "workshop" one of their members, Stefan Priebsch, gave at the recent Bulgaria PHP Conference.
Yesterday I gave a presentation at the [Bulgaria PHP Conference](https://thephp.cc/dates/2015/09/bulgaria-php-conference) (a great event, by the way). Following an [ad-hoc workshop](https://twitter.com/s_bergmann/status/647732967087939584) that I gave as part of the hallway track and an entertaining hackathon, I decided it was too late to join the party and went back to the hotel with some other speakers. Checking out how the day was reflected in social media, I contributed a few more tweets to a [conversation](https://twitter.com/tim_bezhashvyly/status/647861115721003008) that had started earlier in the day ([here](https://thephp.cc/dates/2015/09/bulgaria-php-conference/solid-mvc) are the slides of my talk that people are referring to). I am writing this to clarify my point, and help everybody to understand better.
He talks about dependency injection as a best practice that's followed in libraries all over the PHP ecosystem, making it easier to work with objects and their needs. Sometimes this means using a dependency injection container and others it's just constructor/method injection. He talks about how these objects are build in factory methods and recommends making one factory but points out that this only really works when all the objects you need are known up front. However, he gives several (code) examples of places where this could be difficult and how some are using service locators to solve the problem. He points out, however, that this then expands the API of the application out way too far, opening it up to objects all across the application when there may be no need. This is where the hidden dependencies can come in, things masked behind the use of a single service locator. He recommends solving the issue with more customized locators, as in his example of routing locator used to handle dependencies for a POST HTTP request.