Database Theory, Models and Abstractions


Comments are closed.

Keith Davis at 11:44 on 8 Apr 2016

Good job. Your probably know this, but Google's F1 is labeled as "NewSQL" on Wikipedia, to categorize those RDBMS that are going for the performance and scalability of NoSQL.

Brian Hannah at 12:19 on 8 Apr 2016

Outstanding presentation of a typically dry subject.

Mangesh at 12:48 on 8 Apr 2016

Nice topic that most overlook.

I would've liked to have seen some examples of Postgres.

Pretty great talk
hilarious slides, good examples, and a great overview of SQL basics

I only saw 3 things to nitpick - first of all sql and no sql dbs are moving fast, you should do some research on the new features of both (that are blending things into "newsql")

Although Postgresql has the black box json, it ALSO has fully queryable and indexable jsonb. Mysql innodb has key value store support. sql server 2016 has json support, and as others have noted, F1 is really more of RDBMS and NOsql had a baby.

secondly there are some really distinct types of nosql with very different usage - that would have been great to go over a bit (document dbs vs. key value stores vs. column dbs vs. graph dbs vs. multi-mode dbs)

Finally PLEASE do not mix up sql query abstraction with db api abstraction

pdo only abstracts the api calls - so no matter the db you have the same PHP methods that you use to access and query the db it won't translate your postgresql query into something mysql will understand

there are additional dbal (database abstraction libraries) as well - but those are not designed to abstract queries and are never drop in replacements for changing the database

Then there are sql abstraction layers (doctrine, propel, etc) that actually provide sql query writing abstraction

Two different tools, two different uses :)

This talk was very entertaining. I would definitely recommend.

However, it felt like there were two talks here: One about RDBMS best practices and concepts, another about the relative strengths and weaknesses of different DB types.

There's plenty of material to split these topics and turn one good talk into two great ones.

Lots of material in a short time slot. Kind of rushed but a good high-level discussion around both types of databases.

Carlos Copto at 10:33 on 9 Apr 2016

From a general point of view is a excellent talk(good presentation, examples, comparisons, etc), however, I would recommend to extend into something beyond the old-school theory.

David Trower at 17:07 on 21 Apr 2016

Great talk on databases. As a professor, there were some things I was able to take away to add to my classes.