I work as a network and computers systems administrator in one of the departments of public university (this is a throwaway account, as I’m not authorized to represent my employer) - we use Rocket.Chat CE for some non-classroom projects. As we receive public funds, in most cases we are not allowed to “go and buy whatever you want”. Depending on cost, we may be required to go through tender calling procedure. As you may guess, this takes time. Even buying a cheap wireless router took two months… With less than month advance notice, there’s not much I can do (short of leaving Rocket.Chat out-of-date, probably firewalled from the Internet…).
Buying software is difficult. I am not allowed to say “I want to buy Rocket.Chat EE for X users”, I have to describe features I need (and this means that the best bid may be for a competitor product, not Rocket.Chat). Many times there were no bids for our tender call, because nobody in our country was a reseller of software we needed and our order was too small for any other intermediary to bother. In the end, after several failed tender calls, we are allowed to buy software in a “usual” way - but this takes a looooong time…
And one more thing - each time I have to explain why my purchase is necessary. It may be difficult to justify buying Rocket.Chat licenses when Microsoft gave us MS Teams for free… (“I don’t like Teams” is not an acceptable explanation.)
With regard to changes, I find it strange that you consider LDAP filters an advanced feature. Different LDAP servers (especially home-made ones) may use completely different LDAP schemas and object classes. I’ve seen setup where users, groups and computers were all mixed in single OU and LDAP structure was flat (“filtering” using base DN was impossible). How will Rocket.Chat distinguish between user and non-user objects in such cases?
To be honest - I don’t really like Rocket.Chat, there were just no feasible alternatives back when we deployed it… Now - even if we decide to switch to some commercial solution (but that’s not my call) - we will reconsider other alternatives first (some of them were mentioned in this thread and they look nice).