Present: Graeme, Valentin, André, David, Frank, Tommaso, Xingtao, Geri, Weidong
# Introduction (Graeme)
- Google is an issue, so we proceed as outlined (CERN Indico and email, Overleaf for documents)
- proposing to use Overleaf for meeting notes
- everyone can sign up for an account
- with a CERN account one can get a professional account (however, basic accounts also work)
# LCIO (Frank)
- Do need to foresee *schema evolution*, we won't get it right first time
- Persistency is important to exchange data between implementations (languages)
- "aid" is an interface generator
- As good as ROOT for full objects (but no column reading)
- Some thread-safe work in SIO has been done
- casting was ok at the time (Java inspired), but would not redo it this way
- need metadata - if you want to be generic you also need to be abstract
- support users with additional helper methods
- naming conventions get heated, but make sure users can store what they need and they will use it
## Questions
- What is the role of LCEvent? It's a superstructure for holding all of the events (like the EventStore in PODIO).
- Schema evolution? For schema evolution it's done by special code (e.g. adding default values for new attributes as needed).
- Comment that ROOT is well advanced in MT migration
- Would get this for free with PODIO
- Support for detector RAW data? Yes, for the ILC detector studies, fairly generic. Not clear how to proceed there for all `bells and whistles', user extensions without review would be a can of worms.
- pLCIO is independent of LCIO
- pLCIO was a study that guided how to migrate to a new implementation, not 100\% compatible, but effort to migrate would be `reasonable', a lot of code can be kept.
# FCC-EDM (Valentin)
- Noted that being able to attach metadata is essential in LCIO
# AOB
- Graeme will setup a mailing list as agreed. Further discussion and
scheduling of the next meeting on the list.