PostgreSQL MVCC and alternative REDO implementation - my thoughts

Hello,

I just read that blog article and I think switching to REDO logic is not 
a good idea.

REDO has the following disadvantages, think of the following scenario: 
Delete a lot of rows having 1 petabyte of data.
1.) PostgreSQL: has just do mark the rows as deleted, afterwards they 
have to be cleaned up by (auto)vacuum. No data has to be written! Will 
be quite fast.
2.) REDO logic: The deleted 1 petabyte of data (!) has to written for 
crash consistency reasons at least into REDO files. And that might take 
time, hours, days, weeks, month.

So please don't give up the implementation of PostgreSQL MVCC.
https://wiki.postgresql.org/wiki/MVCC

What's the plan?

Thank you.

Ciao,
Gerhard