Every system of this nature is going to be fundamentally divisive, arbitrary, incomplete, and inaccurate. It's not possible to design a "fine-grained credentialism" system without requiring the full dedication of one person's attention to the activities of another, for every waking hour of the observed person.
Divisive: Where today coworkers have no qualms about sending interesting/educational links to their coworkers, like interesting reads in a technology journal or a tutorial on a new feature of some software (for example), if these things will be counted as "credentials" that improve hireability, job security, and/or compensation, then individuals will be motivated NOT to share anything they learn or read with coworkers, since their coworkers could use this to advance their own credentials, and get a leg up on the person who shared it with them. The people who succeed would thus be recipients of well-intentioned coworkers' educational resources and information, without sharing anything back to their coworkers.
Arbitrary: What counts "for (micro)credit", and what doesn't? Where do you draw the line? If you draw the line at some arbitrary place, there are going to be educational resources that people use, which are extremely relevant to someone's job that actually enhanced their suitability to do their work, but don't count for credit. If you don't draw a line at all, or set the bar so low that just about anything can be accepted, then a lot of people could arguably gain "credit" just by watching CNN and claiming credit for the random sound bytes that sound off information that pertains in some general way to the field the worker is in. Microsoft stock went up? Well, I'll claim a credit for Technology! Because Microsoft is Technology! Oy vey...
Incomplete: There are many experiences that can be very educational for someone, but don't have any authenticity, quantifiability or verifiability to them. For example, if you are on a 3-hour bus ride and strike up a random conversation with a passenger who happens to be in the same field as you, and you learn something entirely new from them that opens your eyes and enables you to do your job better, can you claim credit for that? How would the organization know whether you're lying or not? How many of these little nuggets can you squeeze into their system in a day without being flagged for possible forgery? If there's a limit and you can find it, you better believe the min-maxers will find a way to fill up their daily quota, every day, without fail, on their way up the corporate ladder -- walking on the heads of honest people who probably are more competent than they are.
Inaccurate: This is really the biggest problem with the whole idea of "credentialism" from life experience or gaining "micro-credits" for every little thing you do or learn: you cannot implement a system, short of Orwellian 24/7 total surveillance and constant manual, human monitoring, that *fairly* and *accurately* captures exactly what each person has learned every day, and what kind of merit that learning deserves. Those are actually two separate problems: actually capturing all of the distinct learning events, and coming up with some kind of a system to determine how useful, educational, or meritorious those events were with respect to the individual's suitability to fill a certain role in a job.
If the system is too rigid, you miss out on things like open source projects, reading/responding to mailing lists, the aforementioned "bus conversation", etc. If it's too open, people will gamify their careers through lying or taking the easiest course toward getting an advantage over people who are vying for similar jobs, all so they can make more money.
Now granted, the de facto education system is basically an extreme example of a system like this that is simply too rigid and too coarse-grained to be fair, but making it fine-grained doesn't actually solve any problem: you're just shifting the problems to another set of equally severe problems, without making the hiring and retention process any easier for employers, and without making the job market any fairer for employees.
Right now we have a system where someone who, for example, is 1 credit shy of a Bachelor's degree, who possesses a vast wealth of knowledge and experience all across the computing landscape, would be passed up in the hiring process at most companies, in favor of someone who has a 4.0 GPA and a Master's degree, but whose experience in computing is purely academic. When it comes time to write working code, any bets on who would do the job faster and better?
In the fine-grained credentialism model, you end up with the same problem, where the person whose greater knowledge and experience doesn't meet the system's criteria for what counts as "credit-worthy activity", yet the other person games the system to fluff up their credentials and gets hired. It's not any better at all! The people who game the system will go on gaming it, and the people who learn for learning's own sake, because they actually enjoy learning about and working with the subject matter, often get under-valued because the activities they do and the things they learn don't fit the mold.
Wish I had some bright ideas for how to fix this mess, but I don't. I just know that this particular mess is, at best, equally as flawed as the de facto education system.