Traditional epistemology is based on something called the 'Cartesian Foundationalist Program' (CFP). CFP assumes that natural knowledge (knowledge of the sciences) must be based somehow on sense experience, meaning that knowledge reduces to observation and can be deduced from observation.
However, neither project works. In the case of reduction, as Quine argues, even the most modest of generalisations about
observable traits will cover more cases than its utterance can have occasion to
actually observe. Put simply, observation itself never offers enough content to account perfectly for the generalisations made about them. Similarly, Hume persuaded us all long ago that one can't deduce theory from observation. A scientific theory is not guaranteed by it's observational premises in the way deduction requires it would be.
These problems with CFP are well known and widely accepted, and yet epistemology has not yet been willing to let go of it. To understand why we need to understand what was so attractive about CFP in the first place.
In a nutshall, CFP has been preferred because it offers a clear definition of what a justification is, a.k.a. a justification is a 'certain truth'. It's harder to make the case that we are justified in believing a 'likely truth' or a 'possible truth', but a person who believed in a certain truth would clearly be justified in doing so. Unfortunately, the promise of 'certain truth' is unfulfillable. CFP believed it was possible due to the above two mistakes, but also because of a much larger, often ignored, mistaken belief that sensory data is more credible than it actually is, and the often missed point that observation is theory-laden.
Epistemology, has therefore been forced to find a new standard of justification. It has been suggested that rather that attempting to 'vindicate' theories, then, we should merely attempt to 'validate' them. That is to say, offer some reason why they are good candidates for true beliefs independent of them necessarily being true beliefs. And the validation approach is widely accepted. No one around these days in philosophy seriously believes in vindicating theories. But paradoxically, philosophers still find themselves attracted to CFP, and time and time again, efforts will go into making CFP work.
The problem is, although we're left with validation as our only option, this approach has not been established as a coherent, tight philosophy in the same way that the 'vindication theory of epistemology' had been. That is, it's obvious how, if it were possible, a vindicated approach to justification would act as a candidate for truth, it is not obvious how a validated approach would. I conjecture that this is why CFP, though understood to be false, remains so prominent in epistemology, and I believe it will, until the validation theory is fully thought-through.
For more on the problems of the validation theory of epistemology, see P.2
For my purposed solution see P3.