- HTTPS – Perfect forward security
- Six Lessons Learned The Hard Way About Scaling A Million User System
- Troy Hunt: The “Cobra Effect” that is disabling paste on password fields
- uncss – Find unused CSS
- HTTP2, Promises, Web Audio (among other features) in development for IE
When it comes issue of having nullable columns the DB design community is split (you’re either a big fan or normalisation purist). The argument against the overuse of nullable columns is two fold.
1. If the majority of rows in your table contains a columns which are mostly null then it’s wasted space (this is especially bad for large tables)
2. Application’s poor handling of nulls can result in exceptions being thrown.
Arguments for nullable columns:
1. High-levels of normalization results in a increase in 1:1 mappings, which can increase the complexity of joins.
2. I’m of the opinion that the application should handle all inputs gracefully (i.e. pre-condition for a unit of execution). The database is there to serve the application since that’s what the users are interested in.
3. Application flexibility – fields need a default value that’s future proofed and null provides that mechanism
In the end it comes down to your business requirements, there’s no hard and fast rules. Those in the enviable situation of having to work with small databases can go all out and null everything except the primary/foreign keys.