Fall 2007

• My milestone report can be found here.
• UI team + Search team (largely Jeeyun and Rae) managed to get advanced search working, along with the trust network. This was a great piece of progress.
• I began implementing AJAX interfaces for comments and trust networks. Did not get as much as I wanted done.
• Developed thumbnail viewing for search.
• Displayed content created by users (thumbnails) on a user's profile page.

• Milestone report here
• Additionally, I created tables in the database for permissions and am working on getting basic access control working Please be less vague here. In your milestone, you said that you had left these tasks undone: "I still need to implement methods for adding/removing and querying permissions on an object. I also need to implement methods for updating and retrieving permissions from the database." Are these basic access controls? If not, what are they?
  

• A significant amount of work was put into the relationships model this week. There is now an interface for adding relationships into the system, as well as a way to query for relationships using relationship types, subjects, and objects.
• A tree parser was made to translate abstract search trees of the Search team into relationship model queries.
• Stub methods are in place for creating comments and tags.
• We added methods to the Content and Entity class that allow you to call on an object to give you relationships such as content a person has authored, people who trust a person, people the person trusts, people who rated a content, and other common relationship queries.
• There are stub methods in the backend for people to call on to create comments and tags.

1. Other classes/organizations took up a lot of my time this week.
2. Poor time organization. There is so much to do that I end up not giving enough time/thought to anything. Just an encouraging note: you've been an asset to the team, so don't be too hard on yourself when you have an uncharacteristically poor week. --Chelsea

• We were hoping to have working methods that create comments and tags, which we knew would require extensive coordination with the backend. I was supposed to keep in close contact with the backend, but failed in this regard. I sent the team an email asking whether they supported a comment data structure, but did not receive a reply, and I didn't push them on it. I also tried to get in contact with them on Wednesday, but could only reach Aaron, who told me that Matt was working on the system our team needed. We only received contact from Matt after the milestone was due, and he told us the system would not be in place in time for the customer meeting. Part of the problem was that I saw little reason to bother the Backend team was that we had already put the stub methods we hoped to use. I felt that having the stub would be enough until the Backend supported the appropriate data structures. What I did not understand was that the Backend had a different idea of what the comment data structure would be like than the Relationships team conceived. They had placed their own stubs into the Backend of how they felt the method should be called. The difference between our methods amounted to different parameters, which are relatively minor, and can be resolved fairly easily, but I did not understand this issue when Corey assigned me the task of coordinating with the Backend. A sound case for proactive communication! --Chelsea

People keep thinking in terms of creating their design/implementation based on what another module provides in terms of an interface (= services).  That sort of thinking is backwards.  A module has no idea what the other modules in the system are, only its interfaces out to the rest of the world.  Interfaces are gateways out from a module, not gateways into a different module.   You determine what functionality you need from the outside world.   This determines the interfaces you need.   If those interfaces do not match the service provided by the actual modules used to receive your requests, then it is the job of an adapter to bridge that gap.   That's why the MVC pattern does not utilize direct connections between the model and view.   -- SW

1. I NEED to get the testing project up. There's really no excuse for it not being up by now. This isn't really an obstacle--this section should tell everyone what the holdup is, not just state that there is one. --Chelsea
2. Saturday was ridiculously busy for me - and it's really bad because that makes two weeks like that for me now. NOD and Esperanza two weeks in a row are um...bad for productivity. I know I shouldn't let it be that way - priorities and whatnot, so I definitely need to give 100% this week. Campus-wide events affect everyone, so the question to be asking here is, "how did the rest of the class overcome this problem?" You don't say whether you told your group leader and the PM that you planned on taking this time off from the class; did you? The worst thing that can happen in this class is a failure to plan for busy weeks. This entails letting everyone affected in time to alter your milestones and expectations for those weeks. Unfortunately, exceptions for these kinds of events cannot always be made--but the situation is worsened by lack of honest communication.
3. I feel kind of disconnected from everyone - that's bad since I'm part of the relationships team. If we are as integral as the many diagrams say we are (being smack dab in the middle of all the other groups), then this should not happen.

1. I dropped a class. I'm now down to 13 hours, which is probably the equivalent of about 16 hours if Comp410 is measured at its actual time committment. The solution is not to make time for unfair work distribution. Solve the uneven workload by talking to your PM and the rest of the class--I guarantee that not everyone is being held to this standard.
2. I'm going to make a list of all the things I want to have implemented by Sunday night. This will allow me to focus my mind to the task better. This is something I should have done a while back. Isn't this what the new Sunday/Wednesday milestones are for?

• I feel the issue might've been better solved without assigning a single liason between teams. What the problem required was agreement between both teams about what the data structures would encompass, and more specifically how comments differ from tags. Perhaps a specs document? Having a single person communicate this back and forth is not effective as having the two teams as a whole in discussion. Corey definitely had a more complete understanding of the issue than I did, and when I was in contact with Aaron, we felt we couldn't decide between just the two of us on making a decision that would change how the entire system would work. I think our teams are small and flexible enough that a group meeting to flesh out critical components like tags and comments would be very possible.

1. Okay, okay it's getting done now. Again, not a solution (to a...not-obstacle?).
2. Prioritize! Got my planner ready and everything. I did spend some time today to organize all the work for the rest of the week, so this week should run much smoother. For all that google calendar, ical, etc. do - I still find that nothing beats out some pen and paper...well a physical planner at least. I've also discovered the miracles of e-mail tags - too bad thunderbird sucks at utilizing them :( Gmail is my savior (and see above for other solutions).
3. This week I'm going to have to maintain connectivity with Hubert, which should be pretty easy since he's only a floor down from me. But beyond this I'm going to try to make sure that I keep in contact with Corey, Matt, and Rae, since everyone can use a liaison from the relationships team and Corey's not always available (neither Kevin to boot).

• As during last week, it seems the teams are much more focused and efficient. In particular, it seems Corey's take-over of the Relationships team and hence Rae being allowed to concentrate primarily on Searching has allowed both teams to make quite a bit of progress.
• Teams are communicating much, much better. When important things need to be coded, people from the affected teams are there in the same physical location which helps productivity a bunch.
• I was able to delegate my work this week better. Delegating advanced search to Jeeyun allowed me to study, understand and implement AJAX interfaces for a couple of our features.
• We finally have a bug tracking system up (Trac), thanks to Corey, but still need to get around to using it extensively. This will help us keep track of tasks which we have to do, and not forget them. Yes, this will be useful, but why not just use Sharepoint and keep it all in one place?
• Working on a single branch for the most part is working. I think we were using branches incorrectly earlier. This time, Felipe used a branch to develop image tagging and that was a correct use for it. He was also able to merge it in correctly when the time cam to put it into the main branch. In fact, no one else even realized (from a build error point of view) when he had merged the new functionality in. This is good because it showed the importance of identifying an interface early from a development point of view. This is a really important nuance.

• The impending deadline for testing with users has forced us to be more serious about what needs to get out the door to see something usable, and I think in the end it will produce a much better looking product.
• The new team arrangements are working much better as expected for a number of reasons: Corey has brought the desired gusto to dealing with relationship issues, Backend has well defined limits on what it's responsibilties are, Rae is no longer burdened by two tasks and is thus dealing with her stuff better. The UI team is still doing a lot but I'm worried about what is coming down the line for them. I think the Content team is doing well, but they've fallen from the limelight, and I'm worried about how lots of little issues in their section will irk users. Perhaps you should take some time to talk with these teams that you're worried about and see if they have any insight into the situation.
• Communication has improved greatly. I actually think a lot of this just has to do with time passing and people getting to know one another better, but team building things probably help a little. In the interest of helping you figure out what makes the class run smoothly, I would suggest you consider how milestone assignment plays into "encouraging" (aka forcing) communication.
• We are managing the source tree a lot better by working from the same branch. Even though this causes lots of annoyances, it also fixes headaches before they become migraines as everyones going from the same picture of the tree. Indeed.
• We are doing a better job testing stuff I believe as its gotten easier to continually deploy to skynet and see what's going on, what works and doesn't and assign bugs.
• Related to the last point, I have high hopes for the bug database giving us a bump in productivity as it facilitates communication. I mentioned this is Corey's journal, but was there not even a list of bugs before? We used Sharepoint--thus keeping the information in the same place as everything else--and created a list of bugs there...and it seemed to work just as well as a separate bugs database.

• It was remarked during the customer meeting that we have aligned our milestones more closely with the customer meeting than ever before. There was not a last minute scramble for fixes. Well, there was that one try-catch block Sohum had to throw in the code during the meeting, but what is important is that there wasn't a mass scramble of people from different teams pounding away at the system to get the parts to integrate together. The trend will continue as long as we maintain what velocity we've gained this week. I'm glad things are going well, but I would be hesitant to call one week a trend. Continued effort will be needed to get the ball rolling on this one. --Chelsea
• We had, perhaps the best customer meeting yet this week, having informative and engaging presentations, having a relevant demo of our progress (despite losing Dr.Wong a dollar), and ending on schedule.

1. ... Motivation is the only thing I can think of. I think we are communicating and developing as effectively as we can with our current levels of motivation. I, in particular, have become less excited about this project than I was at the beginning. However, I think this is contrary to the direction of change in motivation of other members of the class. This is interesting, since I thought you would be more excited now that you are on the end of the project that you are interested in.
2. While we have been treating integration quite lightly over the last few weeks, this week will be a test of our system's integration level. I think we need to focus our energies towards the deliverables for this week, seeing that we have a hard deadline for some sort of usable product. Yes, this is very true.

• It is really hard to find relevant information about what every team is working on (which is important when you're trying to work around dependencies), their current progress, and what obstacles and problems they're running into (which, in a dependency-heavy system can easily be your obstacle/problem as well). Again--documentation? The milestone tasks page still has old and outdated items that make it hard to find current work items. Yes, updating this list is essential, both for the team's use and for helping the staff grade! Thank you so much for bringing this up!! Please get it on the agenda for class this week for everyone to read the comments on their last milestone reports--I've made some suggestions. This issue is extremely important, and you've very eloquently stated why. We tried to solve this problem in a way, a while back, using an IRC channel, but asking someone a question on there did not guarantee an immediate answer.

1. For me, the coding parties are way too early. And I admit that it does sound ridiculous since the coding parties are at 11:00 AM - but I'm definitely not a morning person, and have a penchant for sleeping in until 2-ish or so on the weekends. I don't think I'm the only one like this, so really it shouldn't be such a problem for me.
2. I also didn't get much of anything done at the coding party to be honest. The other times have been rather productive, but this time the atmosphere didn't help me concentrate at all. I guess it's hard to predict effectiveness for a given week since I don't know what my mood will be like on any given day, but maybe I should have left earlier to work somewhere else. Legitimate concern. Coding parties should be productive times for those who need to be productive.
3. For whatever reason, despite the e-mails from Dr. Wong and Felipe (although I was working on the journal by the time that Felipe sent me it), I have had absolutely no enthusiasm to write this journal. But I don't speak of journals in general, just this particular one. I don't know why - perhaps because there really isn't all that much to write from my end. It was a lot easier to write something down when I was more central to the project at least - perhaps I was not as involved as I would have liked. It's hard to tell, but I will definitely push myself far harder this week with the 8th coming up so quickly. Dig a little deeper. What's the underlying issue, and how does it relate to your productivity on the project itself?

1. You cannot force someone to like something. I think each person individually just has to think of something really cool that they can add to the project, and that is what will make them motivated about it. This is the reason personal projects are most exciting--because you basically do whatever you want. In this project, we have to channel that energy towards the project. A fine example is the authoring team's development of image tagging. While this wasn't a high-priority deliverable on the customer end, Felipe and his team thought it would be a cool thing to add and hence put in a lot of creative energy into it. (Think of ways to sustain that excitement in the project. One way is to get people to come up with their own personal ideas for what they want the project to be. Keep a list of these and if you have time at the end of the semester (or as you go if you think you have the time) start implementing some.)
2. Start early. I think we are shooting for a Tuesday deadline to have everything ready. We are also going to be user testing this weekend with friends and such, so there is an added motivation to getting things working. We will have to work extra hard on Saturday and Sunday to get this thing in good shape. This may be the case for this week, but how can you keep yourself out of the "work really hard this week" cycle?

• I'm a pretty big fan of the new Trac system Corey put up on the Skynet server. It is really to see all the bugs/tasks that are assigned, by priority/team/individual. You can easily create a new task/bug that others can pass on or take on. I think if everyone used it we would all be much better informed about what others are up to.

1. This point really is more of a personal thing. (Yes and no. I'm sure you're not alone in this, and if you have an important role to play for a particular integration, you deserve the chance to be at your best. How can you take steps to make that happen?) Much as I don't like it, I'm going to have to start going to sleep earlier on Saturday nights and get up early enough to be productive at the coding parties. I make exception for myself the last two weeks since they were the weekends of NOD followed by Esperanza in succession.
2. Hm. I don't want to criticize the spirit of coding parties because of the unique dichotomy involved. On the one hand, there's coding - and on the other hand there's partying. It seems like an either/or kind of thing, where there's no real way to tell what it's going to be like on any particular day (or at whatever time). I'm aware of how I work best, and I need to be more aware of my surroundings and more active in terms of changing my surroundings to get things done. I like the coding parties - I like the opportunity for group coding as well as group partying (though not so much at the same time), so I cannot criticize the institution, rather I must make ammends to my own approaches to work.
3. It's relaxing being less central to the project, but we can't afford relaxation. And if we can't afford it, then I can't afford it. This week should be far more productive for me - I really feel like I can do a much better job of motivating myself tomorrow - get out of this funk/daze. But how? You're an intelligent person; if it were as simple as "snap out of it!", I'm sure you would have already done it.