Yes, right. What I am saying is that any non-trivial project will be certain to use custom allocations, which means that freezing the allocation at the standard level is not expected to add much value beyond the most trivial applications.
Let us shelve this discussion though because I think it is now clear that one way or another we will find an agreeable solution to this problem by utilizing these Vadimfiles. We have been stalling the progress long enough so with the fragile near-consensus we have reached in this discussion we should move on to the other parts of the standard. Unless there are objections, I would suggest we move on to define the first UDRAL network service starting with something relatively basic, like, perhaps, ESC or servo. @coder_kalyan and @bbworld1 have already done some work on the design document so perhaps we could leverage that. Those who reside in compatible timezones are welcome to join the dev call tomorrow to discuss this; those who can’t join shouldn’t be worried because all of the relevant context will be also provided here on this forum.
I suggest we avoid discussing UDRAL network services in this thread, that would be off-topic. Let us start a new one.