I have two points:
1. I wish that developers (at least, our developers anyway) would regard the DB as something less than a place to dump data and as something worth learning. I've had developers tell me that they don't care about the database and they express their contempt for the database with Entity Framework and careless datatype choices. All text-fields are varchar(max), all dates are datetime2(7) and GUIDs are present in every table just in case it may need to be accessed by a webservice. With such an attitude towards DBs, I'm not surprised that DB-modelling is falling out of favour.
2. I would be interested in more articles on DB-architecture.
For example, I'd like to see articles on the suitability of long tables (tables with many rows) versus wide tables (tables with many columns). Which scenarios suit which table type best?
Or articles on archiving: the removal of cold data so that it is still accessible but that it doesn't slow the access of hot data.
Or articles on scaling out databases with the likes of AlwaysOn, replication, log-shipping and clustering — how feasible is a DB-farm with, say, 20 servers that are read from and one server that is written to with the the DBs that are read from being copies (as appropriate to the technology) of the one DB that is written to?
Or something fundamental about the position of data at the various levels in the data hierarchy. Say I have public data like (address and postcode data) and it is used by many, many different databases, what are the pros and cons to having the data within each table (no normalisation), as a separate table within each database (normalised but any change in the data must be made to all databases), as a replicated table from one master (normalisable, one master but the joys of replication), as a table in a separate database accessed by a linked-server (no FKs etc).
Or the pros and cons of replacing a table with a very small number of elements unlikely to change soon (like the sex of a person) with hardcoded values and check constraints (so, 'F' instead of 0 and the removal of the person.sex table and the foreign key that goes with it). Heresy in the eyes of normalisation, one must remember the check constraints and all that.