Thats not an urgent issue. Personally I see it more as a strategic decision and that it would make sense to have such a formatting.
When looking back, i think it was around 1998 companies promised to create applications where databases can easily be exchanged. From what i know it never really worked (that easy).
What happened in the past years, mainly because of license fees, the old established companies started trying to move away from Oracle etc and tried to install PostgreSQL etc. But as there are special functions in the database (e.g. like reporting tables, feeded through materialized view logs updated by triggers on the original operational tables (the exact wording for this is a litte different)) its very difficult to change the db to another vendor. Another issue are user interfaces - talking about Oracle - programs like Oracle Application Express (APEX).
When I saw pyDAL a few years ago with Web2Py and got more used to it, mainly through Jupyter notebooks and the upcoming py4web, I realized the potential. Thats is, what was promised a long time ago (1998 ff).
Having used a 2 page cheat sheet for SQL and seeing the SQL-99 (or whatever newer SQL Standard) now, I tried to use DAL instead of SQL. There are still some areas in pyDAL like views which are missed (but there is also more in pyDAL) compared to the SQL standard.
I see two reasons for having a such a timestamp formatting:
(1) From this point of view I would have said, having a good overlap with pyDAL compared to whats available in SQL-Standard, it would be good. Thats as a general issue. (I someone would have asked me weeks ago I would have said timestamp formatting is there in DAL).
(2) Coming from an application I am currently working, a D3/Vega-Lite real time graph. The sensor data is going into an SQL database. From there - best activated by an insert trigger but currently polling every 120sec - new data should be provided through YATL to an HTML graph page. Initially I wanted to work with the insert timestamp in getting new data when comparing it with the current timestamp. The new data sent based on an insert trigger would be real realtime. Right now i am also using the database record ID instead of the timestamp for selecting new records/measurements.
Re problems or better solutions, on server time and local time, i think it was Luca who suggested always to use UTC.
Sorry that it got long. I am sure there are other ideas about this. It's Massimo who has the best overview and should finally decide. There are workarounds.
~Klaus