While I agree it would make sense to separate the two, note that having a high reputation on Q&A doesn't imply you're a good mathematician either. You may gather reputation by simply asking many good questions. Or you may gather upvotes by giving good pedagogical answers to simple mathematical questions; in that case it's not your ability in mathematics, but your pedagogical ability that gives you the reputation. Indeed, if your goal is to maximize that number, giving good answers in fields that many people understand is likely a better strategy than giving good answers in fields that few poeple understand.

Note also that, unlike Somewhere Else, this number doesn't affect your abilities on the site.

Hahaha, I definitely agree on this. I noticed that some users with high rep on SE are just extremely patient answering basic questions. I though that it would be great to have separate measure that would count the reputation in answering challenging questions in a field. Something in the sense that if someone has high rep on mathoverflow, that person perhaps really understands some field of mathematics, in contrast the rep on MSE indeed does not mean much. I would find it useful to be able to see who are the users that are "strong" in specific field – other than that I notice them providing a great answer to a challenging problem. However, this is rather off-topic.

Well, the reputation does not limit your actions here...isn't it possible just because the community is small, so there is no need to reduce the spamming ability of new users? Or is it just that codidact's philosophy is not to use such a manipulation as a way to make new users work hard? That's noble! But realistic?

On reputation not limiting your actions: It is an intentional design. Codidact does limit your abilities, but it doesn't do so on reputation, but on measures that relate to the abilities. For example, you cannot gain edit ability by answering many questions, but by suggesting many edits that got accepted. Which means you get the edit ability for proving that you are good at editing, not for being good at asking or answering questions.

Thanks a lot for the clarifications. In that case Codidact is more sophisticated than what I thought, I really hope that we can make this community grow, and migrating here from MSE won't involve so much of tradeoff between principles and pragmatism.

Pavel Kocourek‭ See https://meta.codidact.com/categories/30/tags/3917

I think that it's intended to be seen as community reputation rather than mathematics reputation: it measures your contribution to the math.codidact.com community (and that, in turn, is arguably a proxy for activity in the math.codidact.com community, as long as you're not mainly making posts which get downvoted). This is also true on MathOverflow. I have much more rep there than Terry Tao, but I'm not a better mathematician than him! I'm probably not even a better combinatorialist than him, and that's the area that I consider myself semi-competent in.

I completely agree. I was not serious about trying to prove myself being a good mathematician by gaining high rep. I simply meant to illustrate that voting on meta is used to reflect agreement versus disagreement whilst in Q&A it reflects well formulated/written versus poorly formulate/written. Therefore adding up the two measure does not make much sense to me.