Chem info futures
Changes in our field
- The representation problem is (mostly) historical
... but what a legacy
- Storage/retrieval/display problems are still with us
Data sources and producers as "network services"
User interfaces as "web browsers"
Security and data integrity issues
- These problems are getting easier!
Doubling time of chemical information is ~20 years
Doubling time of computational capability is ~2 years
- Chemistry is not the first science to go down this road
Art becomes commodity
Priesthood becomes proletariat
Things work better, standards go up, other problems arise.
- Unification with neighboring sciences is possible
Bioinformatics and genomics
Traditional polymer chemistry
Future Daylight software
- Major simplification of information model
Objects have properties
Properties are objects
That is all.
Simple and powerful, but much work needed to make it happen.
- Unification of Thor and Merlin
Archival/integrity handled by O/S
Searching done via mmap/pthreads/FSMs
Modern O/Ss, multiprocessors make this possible
- Encapsulation of cheminfo processing
Back to RDBs: ORDBMS datablade (Informix) cartridge (Oracle)
Up to "pure" ODBs and OOP interfaces (JDBC)
Over to object brokers (CORBA)
Down to small processes (e.g., transient databases)
Producer/consumer identities get blurred
- Toolkit functionality remains intact
Oopish programming interface was a good move.
We'll keep it.
- Version 5
Formal languages (e.g., SMILES)
Unified thesaurus database servers
Intimate with other types of databases
Relationships as primary objects
"EZ" API will be superset of "DT" API
Daylight Chemical Information Systems, Inc.