Week three of the project brought a change to the group, with the addition of a fourth group member who has extensive programming experience, much to Hannah’s delight (18.03.22 lns 22-31):
R: Okay um and then uh you added a new team member to your group uh you seemed fairly enthusiastic about that, volunteering pretty quickly
Hh: Ah yeah
R: How has that changed the roles within the group?
Hh: Not much he is also quite a good programmer he’s in our Computer Science class … he’s actually brought forward some pretty good ideas for the options panel and um things like that and he’s helped me with the CSS coding cause he understands it well so it’s actually been quite efficient bringing him to the team so I knew he was a good
programmer so I was like ‘yes please’ [laughs] so yeah it’s actually made it a lot easier for us
The addition of this new member, Chris, meant that the responsibility for coding was spread out amongst a greater number of people, and with this in mind the team set for themselves a timeline of tasks. In the earliest iteration of a project timeline, the group created a board featuring several rows relating to the different types of tasks, either programming or design. Tasks were written on self-adhesive note cards and could be claimed by an individual and then moved from left to right in columns such as “to-do”, “doing”, and “done”35. Mentors from Arm came in to offer the groups advice on how to structure their boards, including a suggestion to split tasks into small sub-tasks to make them more manageable.
Hannah and the team took this advice quite seriously, working together to decide how best to approach the vast amount of data they had before them, and the multitude of tasks necessary to present it in a readable form (18.03.22 lns 45-55):
Hh: Uh we’re gonna kind of work with what we’ve got currently um and then try and then once we understand KeyLines more bring in more data and start you know obviously cause then we’ll have different categories and so it will will be much more wider um because obviously once we have the brands some of the brands might make the same sort of products but some of them might not make anything similar so they’ll just be on their own over there so if we could find more brands that make that sort of products then we can somehow link them and if there’s a brand that makes those products AND those products then we can make one big connected network which will look a lot nicer as well than having loads of nodes on their own off in the distance
R: Yeah
Hh: So we’re going to try to minimise what we’ve got and then start look into adding more
In the early stages of product development, the team decides on a deliberate strategy of starting to experiment with the software using small amounts of data from the electronics category36
before adding in additional categories of data to the finished code to create a final product. This moment also provides some insight into Hannah’s concerns about the data they have selected; there is a very real chance that some product nodes may not connect to any others based on the information available. By setting a working plan involving a small amount of trial data and a provision to add more as needed, Hannah and her team seem prepared for multiple
contingencies.
The following week saw coding begin in earnest, with Hannah and Chris working to understand the source code for what they hope will become their signature feature, ‘combos’ which essentially combine multiple nodes into one based on a shared characteristic, such as brand. Hannah was particularly keen to get combos working because she had insider knowledge that combos are a new feature for the company, and so mastering them will be an impressive feat. It is as she prepared for this task that she met her first real check (18.03.29 lns 8-14):
36 Students were provided Amazon review data from the majority of product categories. The team determined for
Hh:…I’ve been familiarising myself me and Chris with the actual coding for making combos and stuff like that uh which we were both kind of struggling with last week because when I did work experience there they didn’t give me data to work with I had to make my own so obviously I wasn’t going to make 4,000 lines worth of data I was only going to do about like 20 so I was only working with a small amount of data so when I was coding them to do things I would just hard code it or instead of finding common links I just wrote all the separate nodes I wanted within a combo
Hannah had come to the project fairly confident that her work experience would give her an advantage, and in many ways it likely had to this point; it did mean she knew what features are possible and what might most appeal to the expert panel of judges later. She could not directly translate her skills from work experience here however, as the nature of the data was quite different. She knew what to do; she needed to “…write um a line of code that will reference to all of the bits of data that have something similar within them and then pull it out instead of writing each separate one like I did” (18.03.29 lns 16-18), but on this first attempt she was unable to find a solution. It is a clear blow to her self-confidence, as she made quite clear that while she had an experiential advantage, it was not enough in this case to help her to work through her current problems (18.03.29 lns 137-140):
Hh:… everyone apart from me hasn’t even used the software before and I’m struggling and I’ve used it and I think there’s only one guy who’s managed to get combos going – ish um but he’s really really intelligent with programming so the rest of us are just kind of like mediocre we really need help
Hannah still clearly believed that she should have had an advantage on this project, due to her experience. However, she here described her programming ability as ‘mediocre’, which is quite at odds with her earlier optimism. Her frustration was perhaps compounded by a concern that the resources available to help her were insufficient, particularly in terms of personnel (18.03.29 lns 127-132):
Hh: Yeah I don’t understand the point of these mentors because they’re not actually from KeyLines … most of us aren’t having troubles with the actual design we’re having troubles with the actual code so what we’ll talk to them and they won’t know kind of
they’ll try to help but they don’t actually know they’ll ask questions about how’s our planning going how’s the report going blah blah blah but they won’t actually know like if cause some of us some of us are genuinely having problems with the KeyLines code and they don’t know that
As Hannah lost faith in her ability to complete the work independently, the freedom and lack of explicit guidance she had been receiving began to feel more like a burden than a benefit. Hannah valued the design team, as well as the need for organisation and long-term vision, but she was feeling unsupported in terms of experts upon whom she can rely. This suggests that Hannah’s feelings of mediocrity mentioned above were impacting her confidence in her ability to work independently as well. She did not minimise the contributions of the mentors to overall group productivity, stating that “…they’ve helped out with kind of minimising tasks how to split it out which has helped um especially with my board and the time plans and stuff…” (18.03.29 lns 162-164) but, as the lead programmer, Hannah was more concerned about her share of the project.
The pause resultant from the half-term holiday provided some opportunity for a
perspective shift, and when she returned from the break, Hannah had a renewed sense of purpose (18.04.17 lns 5-9):
Hh: …I was just kind of organising the Trello board [Appendix B.6] a little bit more and trying to find what problems that we have that we can work on so I’ve identified what we need to do, I just don’t think anyone’s done it uh which is good cause we didn’t really know where to start a few weeks ago so now we know ‘okay we need to do this, we need to do that’ so basically all I’ve done is just set us out some goals and who’s gonna do what
While Hannah had not yet found any solutions to the problems she had been dealing with, she then at least had a sense of what needed doing, and who was the best person to do it. Hannah still had not had any interaction with the experts she felt would be able to guide her in her project, but she was spending time with what resources she did have access to (18.04.17 lns 38- 45, emphasis added):
Hh: Uh well me and Chris have just basically been looking on the KeyLines site and just trying to see how they call from the data file because the files are separate when I did mine I had it all in one file because it was barely any data I could just have it in there um but we have to call* to the other file and drag out the pieces of data that have this
one thing in common and we don’t know the syntax for it, we know the logic perfectly well, we know what we want to do it’s just we don’t know how to write it
so the way we find that out is just go onto the KeyLines on their API references and we we’ve just got to read through it all, it’s just a lot of reading, finding what we need… Hannah’s previous experiences with the company had not prepared her for all the specifics of the task at hand; where previously all of her data was easily identifiable (as she wrote them herself), here she had to find a way to call the data she needed. She and Chris planned to dedicate time the following Challenge to working through the API references, demos37, etc.