Fall 2007

1. Ensure that the teams were able to integrate and move data into and out of the backend data store.
2. Decide on coding standard for the class.
3. Get everyone in the class moving as far as writing code, producing functionality, and feeling good about what they'll do for the rest of the semester.
4. See a functional first version of the system, enabling us to upload content, view it, and search for it.
5. Settle important design issues.
6. Improve our process of setting up milestones, measuring success, and delegating managment to team leaders.

I believe of these, I've accomplished each of these to some degree. Here are my thoughts:

1. As evidenced by the working demo: http://www.skynet3.cs.rice.edu/comp410f07/atelier/WebApp/Content.aspx?id=72db7eb8-3078-41dd-be02-98c3e20a398d, at least one team has been able to move data from the back-end to watch video. Additionally, by watching the database tonight, I believe the upload team is now successfully creating content.
2. I presented a draft of a coding standards document and we dicussed some points, but I'm not terribly optimistic about really enforcing these without some more work.
3. It appears almost everyone is writing some code, and although we have some issues in our distribution of work, we are making real progress towards full production I believe.
4. It appears we're well on our way to accomplish this goal.
5. With some extremely important caveats, we are moving toward defining the architecture design, and have continued to raise important issues and are moving toward addressing them.
6. The results seem to indicate that the milestones were clear enough this week to get everyone to actually do them. Team leaders are doing a better job of getting their team working, with some exceptions, and the outcomes are extremely promising.

• Figure out design for search once and for all. We decided this and put up a UML. I believe it is here.
• Wrote stub code for indexing classes. These classes will allow search to go through indices first to see if the results are already cached.
• Wrote a temporary backend solution to get search by tags work from the database. Searching only does full string matches so far, would probably want to increase that to substring matching.

1. We haven't settled a tremendous number of issues about how the search team will interface with the backend, and I think they're having some trouble integrating.
2. We still have a long way to go as far as enforcing coding standards, setting up code review process, and a number of other related issues (like code standards for javascript).
3. It's hard to say whether people really feel good, but I think they're producing.
4. I think this will be achieved within a day or so.
5. As referenced in 1, this is far from done. While I'm happy that we're coding, we've got a lot of design decisions to make. I'm fairly sure that we'll need a transaction system in the backend which will require changing the interface as its been defined so far. Additionally we haven't really decided how the Search will interact with the backend yet, and I still need to look into using RDF and SPARQL as an interface.
6. I think we've improved here. The things I'm unsatisfied by are: I needed to get the team milestones out sooner. We need to deal with a couple of teams as far as setting better milestones and working together better I think. But we'll see how this goes.

• Defining the structure. This is still a problem and his heavily affecting the functionality of the search team. The current solution is for the backend to write all the SQL and for the search classes to call it. The problem with this is all the algorithmic logic will be in the backend rather than in the search classes. Perhaps the search class can represent this is in some sort of intermediate language that can be translated into SQL?
• A minor problem is we still haven't decided the range for certain numbers--for example trust and ratings. We need to decide this before we understand the semantics behind our data. For example, it seems explicit trust is a number in the range 1-10, but this somehow gets converted into a float. We need to explicitly decide these conversions soon.