FizzApron logo small.png

Beta ver 2.0

Muchen Feng, Daniel Mar­tinez Villa, Shiya Liang, Kaz Sae­gusa and Kerem Ozcan

Hello FizzApron Fans! We are back again with the second iteration of our prototype we designed, and below you can find our progress blog!

For the ones that were not familiar with our project, FizzApron is a smart apron, with which you can share your valuable recipes with your loved ones as if you are right next to them. They will listen to the recipes from your own voice and you will receive and SMS or e-mail when they finish listening one of your recordings.

You can find the process blog for our first iteration here. The basic idea is the same. We have just made a few design tweaks after our usability tests and developed a much higher fidelity prototype (It’s now a stand-alone device with conductive threads, so you don’t have to be tied to the computer – Yay!).


FizzApron ver. 1 – Geared up and ready to cook now! (Photocredit: User An09 form

We began the spring quarter by evaluating the first version we have and where we want to take the project. After our first discussion, we came up with our goals for the quarter:

1) Conducting a user research to investigate potentials of improvement

2) Developing a high-fidelity prototype of the apron, and

3) Improving the mobile and web interface prototypes.

We also had some considerations about certain features that we weren’t able to put much thinking on. These are:

1) How to separate the cooking steps (Ingredients, overall description of the process and cooking instructions)

2) Audio fast forwarding and/or the possibility of skipping the first sections, and

3) Integrating an RGB LED display for feedback about the status (play, pause, rewind etc…)

4) Possibility of reducing the number of recipes

Based on the new functions here’s the system diagram that we might have:

Figure 1. System Diagram

Based on this diagram we have ordered our new parts:

1 x LilyPad

1 x LilyPad XBee Breakout Board

1 x XBee

1 x LiPo Battery

1 x Micro USB cable

Oh, and not to forget, we have a new member to our team now: Shiya Liang from the MHCI+D program. She will be a great contribution to our team with her tickling designer senses and sewing skills!

First step: User Tests

We decided to start our process with user tests to investigate the potential areas of improvement. We agreed on to use Wizard of Oz prototyping to start as soon as possible. If you are not familiar with this approach, it basically means that the researchers will do whatever the computers are supposed to do. If you want to learn more, you can check the Wikipedia page about it here.

We first decided to make a protocol for the Wizard of Oz prototype. We decided to do immediately. Since the team from the last quarter was all familiar with the project we thought that it was best to use Shiya as our test subject – right then, right there in the design lab!

Shiya: Guys, I am not sure if this is what I signed up for…

Shiya cooked us a Casserole degli board marker, using the material we have in the design lab. We have had her use our interface from the previous semester to record her recipe through the stages. We then had this recipe replicated by another compatriot, Muchen. She gave the rest of the team while she replicated recipe including (but not limited to) stages such as faking to cut the markers, taping them together and changing the caps.

After this session, we discussed some of the issues we can possibly encounter during the real user test and aligned on the protocol. We were initially planning to have one user study per person, but due to the limitations of time we were able to conduct the test with just one couple. Here is our lovely couple: David and Cecilia!

This poor cute couple have no idea what they are about to face.

In our test, David prepared the recipe and Cecilia tested it. We have videotaped the whole process but we promised the couple that we are not making it public so you won’t get to see it (sorry!) But here is a list of what we have learned and our discussions about them:

1) Fast forwarding is absolutely necessary! At times David and Cecilia had to rewind the recording a couple of times to remember something, and they had to wait a lot to get back to where they were. The speed of cooking differs, and we decided that we should definitely have the option to fast forward.

2) Both David and Cecilia said that the strings were positioned too high. They said that they seem to be looking for them around the waist area during the process. They also agreed on that the strings were too long and loose, while they only needed their thumbs to pull them. From the observers’ points of view, we have also observed the same issues, so we decided to address those in our next prototype.

Cecilia trying to replicate the recipe David created. Our researchers caught her murmurring “Will this guy ever learn how to cook?”

3) Cecilia (the user who recreated the recipe) said that she was puzzled about which string was play/pause and which string was rewind. She stated that every time she wanted to use one of them she needed to look at them. We decided that in our next iteration we might find a better solution to differentiate them or at least give the user some kind of feedback.

4) Cecilia also asked whether it would be possible to talk back to the original person who has prepared the recipe. Actually this was an aspect that we considered as well, and we were happy to hear this request from a user without mentioning about it. However, we might skip this for this iteration just to avoid complexity. Although we believe that this is very valuable, we acknowledge that there are other means to communicate with the person who gifted the apron (phone perhaps?), so we decided to outsource this feature until we come up with a better solution.

Cecilia “Oops, just chopped my finger… Nevermind, it will probably taste like chicken”

5)  Cecilia also asked about the possibilities of taking pictures. One of her concerns were that she was not sure about the size of a tomato that David described. We actually thought about this issue even during the first prototype last quarter. However, we decided not to address it in this iteration. The thing is, we are not looking for precision. We prefer to use the metaphor of a family cookbook, rather than a recipe recording device. We understand the users’ desire for clarity, but we believe that the ambiguity is part of the process! We could indeed easily add a photo button to our app by which the recreating user could check throughout the process but we unanimously decided that we prefer not to.

6) Cecilia also asked if it would be possible to include a measurement tool with the apron, because she was never sure of the amount she should put (David contributed to that by giving ambiguous definitions such as “add some salt”). We assumed that the person recording the recipe would tailor it according the abilities of the person recreating it, but we realized this is not always the case. Well, once again we think that this is an issue of ambiguity, which we see as an undocumented feature rather than a shortcoming. We know that our users will understand us as they use the product more and more. We believe that further research about this issue would turn the odds for our favor, but for now we will just have faith in our intuitions.

We have also made our own observations:

1) We noticed that Cecilia (the person recreating the recipe) would make some remarks such as “this smells good” or “great!” We believe that those are really valuable and we wished that we could record those as a separate layer on top of the original recording. This could also potentially turn the apron an heirloom you can pass through the generations. Unfortunately, the time and effort needed to make this real would take longer than what we had in our hands. But at least we know what we can spend our time for the next iteration!

2) Cecilia also kept asking questions like “should I cover the pot?” or “do I need to put this ingredient now?” But of course, there were no answers from the apron, so she just made her own decisions. This can actually be seen as more of a concern compared to the similar ambiguities before, because it can affect the taste of the meal in a sense that is greater than a simple precision ambiguity. But again, it is not possible to ask questions to your family cookbook either, so we couldn’t really address this issue without letting some other qualities of FizzApron to go.

After our user research, we returned back to drawing board:

Do you mean drawing table?

Following the user test, we decided to make these changes in our final prototype:

1) Make the strings shorter

2) Locate them just above the waist area

3) Add an audio feedback

4) Locate the speaker on the shoulder (we later decided to put it in the center of the apron because that suited better for different length adjustments).

5) Use as less cables as possible and make the wiring with conductive threads, using them as a part of the aesthetics of the apron.

6) Add fast forward (by pulling the left string for 2 seconds)

7) Add fast rewind (by pulling the right string for 2 seconds)

Here is a Dadaist-o-naïve sketch of what we are aiming for:

Sorry, the Louvre asked to borrow this drawing so we don’t have it anymore…

After the decisions it was time to begin prototyping. We needed to learn how to sew and reprogram everything using LilyPad and XBee. This has taken a lot more time than we anticipated because of two reasons (1) we decided to include all functions to run as a standalone system only with Arduino, while the previous version of FizzApron was comprised of Processing and Arduino. Since Hardware programming generally offers less feedback upon debugging, it took much longer time to finish coding for this time. (2) LilyPad MP3 requires FTDI cable to program it, which we didn’t have. Our efforts to make a DIY USB-FTDI conversion cable using a USB extension cable didn’t work, so finally after losing more than a week trying to get it right, we finally gave up and ordered an USB to FTDI cable with hardware drivers.

A special thanks to Andy Davidson for keeping his temper when we asked him to make another order just to get this part!

We also needed to learn how to sew using the machine we had in TATLab. Shiya had some sewing experience before so after watching some YouTube videos, we were finally able to get some fancy looking stuff on our test material. The sewing machine worked absolutely fantastic for us!

 That is the kind of pattern we are looking for!

Shiya also tried to teach Kerem how to do this, albeit with limited success…

Shiya’s inner voice: Do all Turkish men lack talent or is it just this one that I have to deal with?

After getting our sewing skills straight and receiving all the necessary hardware we decided to make a final test session before we start making our high-fidelity prototype!

First we began by putting everything but the SMS function together in the LilyPad using alligator clips.

Doesn’t things look sweet when they just work?

Before the process began, Kaz asked the group whether anybody had any MP3’s to test. Kerem had some, and he picked a Turkish folk Music out of them: Değmen Benim (A.k.a Ben Bir Selvi Boylu Yardan Ayrıldım). Boy, would Kaz regret it…

Kaz testing the LilyPad for the 967th time using the Turkish Folk Music.

Who wouldn’t want to spend a beautiful sunny day in Seattle working in a lab?

We also added the XBee on top of our LilyPad. It was a bit more complicated so we weren’t able to get it work the first time. We have done the soldering, but we decided to handle the software part in a later session.

Form: Check. Function: Next Time!

We also found our perfect Apron in Amazon. It also had three pockets, which made us think about the number of recipes we had. We knew that we wanted it to be limited, but why 5? We decided that this could be customizable during the time of order. For this iteration, we decided that our apron would have three, same amount of the pockets the apron has. This, however, could be customized during the order. If the user wants to record, say, two recipes we will ship it with two pockets.

This is your average guy using an apron.

At this point Daniela came up with a great suggestion of using conductive paint changing colors when used. The paint would eventually get worn out as used and the color would stick. We agreed that this would be a sign that shows that how much the apron is used and add another layer of value on top.

At this point it hit us that we are only two weeks away from the deadline. We decided to meet on the Saturday before to put everything together and get them working.

And when the day cometh, thou shall prepare thyself with sufficient coffee, bagel and hummus spread!

Ellen style selfie with our Saturday morning breakfast!

We started the day with a divide and conquer approach. Kaz is making sure that the LilyPad MP3 works seamlessly, Daniel is programming the XBee, Shiya is sewing the conductive thread, Muchen is working on the mobile interface and Kerem is designing and 3D printing a case for the speaker, documenting the process and writing the blog!

Kerem’s view around 10am Saturday morning. The crazy Japanese guy already started his day with two cups of coffee and a can of Red Bull.

Kaz, after a 4 hours of debugging and 7 half bagels

Shiya has permanent eye damage after non-stop sewing. She will have cross eyes for the rest of her life.

Muchen (aka the Queen of Bananas) is trying to decide whether she should finesse the interface a bit more or get some more chips before the guys finish it all.

Few mortals experienced the joy Daniel has had after he discovered that sending 120 SMS online costs merely 60 cents.

We are not sure which one Kerem is enjoying more: His cheese bagel or his case design for the speaker.

After a few hours of work the parts started to come together:

Here’s the 3D printed part for our speaker. We used Rhino 3D to model it and Makerbot Replicator 2 to print.

Fits perfect! All of our group members have their own wizardy!

Oh! Look what I just got on my phone!

Seems like we finally got the XBee working right!

Meanwhile, Kaz helping Shiya for the final stitching details.

The process also gave us the opportunity to make an empirical study about the adverse effects of caffeine on humans. Well, seems like you don’t need IRB approval if you are testing on yourself. I mean, Alexandre Fleming also did it, so why shouldn’t we?

And I just got another message, and another one. Wait… What’s happening here?

OK, now it’s starting to get weird. When did I have children and when did my daughter learn how to cook Arepas?

We also tried to get the reactive paint a try, but unfortunately it was all dried up. It is also getting a bit late, so we decided to say our farewells to it and add it to our “to do list” for the future iterations.

Reactive paint, you are perfect. It’s not you. It’s us… We don’t deserve you. There are other makers out in the world that will make you much happier…

Kerem is done with his share of work and he will be live blogging from this point onwards. It’s just like twitter, but no followers.

Reporting from the scene: It’s almost 9pm and we were very close to get everything together. That is if the conductive thread conducted the electricity. Just a minor problem right? Well not, really if that’s what all your prototype depends on! The team is working on alternative solutions such as conductive paint and conductive tape. If it doesn’t work, our plan B is old-school wiring.

Shiya talking to herself at the corner of the room: Stupid, stupid… Next time test it while you are sewing… Waste of time…

Update 9:20pm We are now trying another conductive thread with less resistance and it seems to be working. Cross your fingers!

Update 9:28pm The new thread we use breaks much easier but we are not here to give up!

Update 9:32pm Muchen is done with the new mobile interface! Below are some screen shots and you can reach the full thing here:

The interface. I guess this is pretty self explanatory. Why not try it yourself?

Update 10:24pm We finally gave up with the conductive thread and decided to use conductive tape, which practically has 0 resistance. We also have some conductive paint but it is very limited, so we might cover the main areas using the tape we have (which is also limited) and finish the detailed areas using the conductive paint:

Sleeping beauty or raging beast? We say raging beast.