Krebs and Myers, from UBC, have a blog post that makes some good points: https://www.zoology.ubc.ca/~krebs/ecological_rants/?p=396
In it, they discuss the 10 most pressing things that are holding ecology back as a discipline. A quick overview: Lack of money, lack of commitment, lack of connections between people, and lack of connections between people and their own history. Many of these have been discussed before (not that they don't bear repeating), especially the simple lack of people and money.
The missing piece of natural history has been discussed before too, and it's extremely important. We have considerable information on a few species, and that's about it. Even working in northern forests, with relatively few species, there is a dearth of data on all but a few of our exceptionally prevalent friends.
One that I wanted to highlight, since it's been a pet peeve of mine, but one that is extremely hard to get away from, is their #7, the focus on very abstract concepts instead of testable mechanisms. There are all sorts of papers which basically use language so vague as to be meaningless, and certainly not testable, especially in the field of resilience. I about cringe every time I have to write that word, knowing it's going to bring on "the definition wolves," as a colleague described them. But the conundrum - I'm working with other smart colleagues from a variety of disciplines, trying to craft some cool, interdisciplinary projects. We're spending considerable amounts of time just trying to understand what we mean when we say "resilience" or "stability," among other terms. So these squishy conceptual blobs are causing all sorts of issues... BUT, they provide a bridge between the disciplines nonetheless. There is a commonality, even if ill-defined. On one hand, the confusion is a pain. On the other hand, squishy definitions allow for evolution in definitions, which may lead to progress - as long as that progress is testable. It's tempting to categorize everything into conceptual frameworks, and without data that gets really amorphous and nearly useless real fast.
Krebs and Myers have it right on - but it's worth noting that interdisciplinary work (which they mention as extremely important in a later point) can benefit from these squishy, abstract concepts which don't hold much mechanistic water. That may be because, to be relevant to multiple disciplines and serve as bridges across the ecological, social, and economic sciences, they need to be ill-defined.