Python & PostgreSQL, a Wonderful Wedding
This was more or less summary of various ways how python and PostgreSQL may interact. There were couple of python libraries introduced (psycopg2, SQLAlchemy, Alembic), then some PostgreSQL python-related extensions (like PL/Python, MultiCorn).Especially MultiCorn is something which gives quite wide range of options to python developers above PostgreSQL, it basically allows to write foreign data wrapper in python, so whatever data you can get with python (be it a text file, embedded DB, XML document or another database), you can then access it within database using SQL.
Schedule entry here.
Nix, NixOS, NixOps
This seemed to be a bit chaotic demo about Nix* tools and OS. I expected to get some ideas about solving packaging-more-versions issue, but the talk was more focused on vagrant-like provisioning of virtual environments.It seemed also a bit interesting that the speaker didn't know about environment modules existence, so obviously he was not able to tell what is the difference.
But hearing some guys from audience that environment modules are great to solve many issues for them was quite interesting for me, especially related to Software Collections.
Schedule entry here.
Upstream Downstream: The relationship between developer and package maintainer
Talk given by Oracle guy who is btw. responsible to maintain relationships with mysql package maintainers in other distros. He tried to show Oracle is not that evil and that they want to fix bugs if they know about them (which is not happening in some cases).It was a bit less technical talk, where also some interesting numbers were mentioned -- that devel count on MySQL project has been doubled and QE count tripled recently.
Even after personal straight question what happens with MySQL in the future I got quite straight answer that I shouldn't be afraid, it won't get closed more.
We also talked together with MySQL/MariaDB debian maintainer Otto and MariaDB architect Colin about need to bring packages in various distros closer, so we unify users' experience.
Schedule entry here.
Modern SQL in PostgreSQL
This was half technical and half motivation talk which was supposed to convince developers to use latest features from SQL databases and also SQL standard. On few examples we saw like one keyword that is not known by majority of DB developers may help in readability or even performance.It was interesting seeing when some of the latest SQL standard features are implemented in various databases. Usually IBM's DB2 or Oracle database supported the features even before or shortly after introducing it in the SQL standard, while PostgreSQL with MS Server usually implemented those features few years after.
And MySQL did not support those mentioned features almost ever, so it proofed that MySQL is better to simple use cases and less advanced (traditional) usage. On the other hand, PostgreSQL with MS Server follow Oracle DB's and DB2's development being few years behind.
Schedule entry here.
ProxySQL : High Availability and High Performance Proxy for MySQL
ProxySQL is a promising tool that may help setting up even complicated deployments with various replication schemes, since the proxy server communicates either with MySQL server or another ProxySQL daemon with pure MySQL protocol. Thanks that feature it allows to do various things according to regular expressions defined to behave differently on various queries.It was almost the same talk as the last year, except there are probably some bugs fixed. Anyway, the tool seems worth packaging to Fedora and playing with it at some point.
Schedule entry here.
User-land and developer-land chat
Interesting talk about communication within open source project -- designers vs. devels vs. users. It was nice to see the opposite side of the relation, for example that users often don't understand why bug reports get closed and why some changes in design happens, which often leads to their frustration.Some simple things like indicating briefly, when closing a bug, that devels still want to collaborate, they just don't have enough data. It may be that simple, so users don't feel like they are ignored.
Another example with Gnome 3 and it's firstly problematic and later quite large adoption showed that even if devels do rapid changes in GUI, users may feel good, but only after some needed plugins have been introduced, which brought some of their favorite features back.
Another issue was that even if upstream is open for collaboration, people still need some borders, so setting up some guidelines similar to Tango Icon Theme Guidelines may help producing consistent output and the collaboration works much better with then.
The last thought was dedicated to need non-technical people, since devels are often reluctant to write necessary documentation, maintain proper social networks, etc.
These all are few topics everyone involved in OSS may think of.
Schedule entry here.