- Get the problem formulation right, design an experiment where we prove that scheduling in a round robin is worse than by our approach.
- Remember what the bug that I found was and fix it.
- Create a first version ready to be cleaned.
2) Finish the Scheduling cameras (pan/zoom) paper - prepare it for VLDB
- Get the code by Allesio to run on my machine.
- Get the experiments re-run and see the results again.
- Create a paper draft from Alessio's thesis.
3) Collaborate with the Google group to finish the Table Visualization paper - prepare it for VLDB
- Clean the version of the paper, SQL is used for debugging and make sure that it is consistent.
- Create a survey for people to use which will evaluate the results.
4) Collaborate with group for person identification paper.
- See how well the network flow algorithm works.
Seems like the reason you can't have a RR approach is due to 1) networking cost of sending the pocket again. 2) The system cannot scale with a single machine which has to determine per packet what to do with it.
ReplyDelete