[Maria-developers] [GSoC] Student interested in "Optimize mysql-test runs" or adding SQL features
Hello! I'm Pablo, from Mexico. Before going into detail, I wanna give a bit of background: I am a master's student, pursuing a degree on Scientific Computing in Korea. My program is 3 years long, and I'm at the middle of the first one. The reason why I think I have skills that will be useful for MariaDB is that before being a master's student, I was working for Oracle during 2 years, committing into the Oracle executable, so I have experience testing, coding and working inside a database executable. Specifically, I was working on the DBMS_SCHEDULER package. Although I am interested in software development, I am more interested in data analysis. That's why I switched from computer science to scientific computing, and that's why the project that interests me the most is the one called "Statistically optimize mysql-test runs by running less tests". I am very excited by the idea of using data to try to test more smartly. If it was not possible to participate in the latter project, I would also be interested to work on adding support for "OR REPLACE, CREATE IF NOT EXISTS and DROP IF EXISTS". Depending on your feedback, I can get to work on a proposal for either one of the two projects. I'd be very happy to be able to participate in GSoC with MariaDB. Regards Pablo
Hi, Pablo! On Mar 06, Pablo Estrada wrote:
Although I am interested in software development, I am more interested in data analysis. That's why I switched from computer science to scientific computing, and that's why the project that interests me the most is the one called "Statistically optimize mysql-test runs by running less tests". I am very excited by the idea of using data to try to test more smartly.
Yes. You might also briefly review scientific publications in this area. I've found a few (it was a couple of years ago), but their approach was different from what is suggested here. Anyway, my preliminary experiments shows that it's totally possible to find 90% of test failures in only 10% of the time.
If it was not possible to participate in the latter project, I would also be interested to work on adding support for "OR REPLACE, CREATE IF NOT EXISTS and DROP IF EXISTS".
This is mostly a straightforward coding task, not much to explain here.
Depending on your feedback, I can get to work on a proposal for either one of the two projects. I'd be very happy to be able to participate in GSoC with MariaDB.
I'd recommend you to submit proposal for the first one. It seems like a better match for you, and also it seems like you will have less competition there. I'd expect more proposals for the second task. Regards, Sergei
Hello Sergei and all, I started preparing a proposal for the project "Statistically optimize mysql-test runs by running less tests". I have a small draft that seems complete without entering any technical details. I am also taking a look at research on the field. Any suggestions on papers are appreciated. Here is the proposal: draft of the proposal<https://docs.google.com/document/d/1vfwNPO7R6PVdRIgHyn-aaMVZugM7KKf7YVnuqXPWEsM/edit>. All suggestions and comments are welcome, and encouraged. How does it look? : ) Regards Pablo On Thu, Mar 6, 2014 at 5:53 PM, Sergei Golubchik <serg@mariadb.org> wrote:
Hi, Pablo!
On Mar 06, Pablo Estrada wrote:
Although I am interested in software development, I am more interested in data analysis. That's why I switched from computer science to scientific computing, and that's why the project that interests me the most is the one called "Statistically optimize mysql-test runs by running less tests". I am very excited by the idea of using data to try to test more smartly.
Yes. You might also briefly review scientific publications in this area. I've found a few (it was a couple of years ago), but their approach was different from what is suggested here.
Anyway, my preliminary experiments shows that it's totally possible to find 90% of test failures in only 10% of the time.
If it was not possible to participate in the latter project, I would also be interested to work on adding support for "OR REPLACE, CREATE IF NOT EXISTS and DROP IF EXISTS".
This is mostly a straightforward coding task, not much to explain here.
Depending on your feedback, I can get to work on a proposal for either one of the two projects. I'd be very happy to be able to participate in GSoC with MariaDB.
I'd recommend you to submit proposal for the first one. It seems like a better match for you, and also it seems like you will have less competition there. I'd expect more proposals for the second task.
Regards, Sergei
participants (2)
-
Pablo Estrada -
Sergei Golubchik