Thesis seminar - Designing for cooks: interactive recipes for everyday use

Designing for cooks: interactive instructions for

everyday useLucy Buykx

n My research investigates how cooks engage and interact with recipe instructions to inform the design of interactive recipe systems

Human-Food Interactionn Helping create the perfect recipe in the kitchen

n CounterActive (Ju, Hurwitz, Judd & Lee 2001)

n Augmented Reality Kitchen(Bonanni, Lee & Selker 2005)

n Kitchen of the Future (Siio, Hamada, Mima 2005)

n PersonalChef (Mennicken, Karrer, Russell, Borchers 2010)

n panavi (Uriu, Namai, Tokuhisa, Kashiwagi, Inami, Okude 2012)

not just cooking...

n Shopping, meal planning

n integrating persuasive & social media

n healthy eating, local food, sustainability

Initial studies with older adults

n focus groups with 15

n food diaries with 12

n meals & shopping for 7 days

n home interviews

Initial studies with younger adults

n questionnaire & interviews

n 80% collected recipes

n 11% used interactive technology

Capturing Family Recipes for digital sharing across generations

n Lucy Buykx, Helen Petrie, Paul Cairns

n include 2011, Helen Hamlyn Centre for Design

n Contribution is innovative technology in the kitchen

n Little theory behind claims

n Low quality evaluations

n Research questions:

n What problems do cooks experience working with recipes?

n What factors cause problems?

n How can interactive design mitigate these?

Procedural instructions

n step-by-step

n little explanation of how the system works

n can be used without prior knowledge

Recipe instruction set

n title

n list of ingredients

n preparation instructions

n method instructions

n meta-data

Guidelines for procedural instructions for initial use (Ganier 2004)

n “Segmented text... numbered steps rather than paragraphs”

n “Chronological linear organisation”

n “Use pictures together with text”

n “Use headings fitting to goals or sub-goals”

Experiments x 2n 24 cooks, 3 recipes

n 2 experimental conditions + control

n Videoed and analysed

n Quantitative measures: time to complete, reading time, cooks’ ratings

n Qualitative analysis: comments and feedback during cooking, problems, errors made

Experiment 1 Experiment 2

Segmented instructions Segmented instructions with pictures

Integrated, chronological instructions

Semantic structured recipe with sub-goals

Control Control

Exp1: Segmented instructions

Exp1: Integrated, chronological instructions

Integrated, chronological

Control

Results mixed

n No timing differences or rating differences

n Cooks disliked long paragraph steps AND segmented steps

n Cooks did not move linearly, needed to understand context

n Cooks did not follow instructions perfectly

Papers at

n Cooking & Eating Activities workshop IEEE MM CEA2011

n Food & Interaction Design workshop CHI2012

n Food for thought workshop DIS2012

Experiment 2

n Compared affect of

n Segmented instructions with goal state pictures

n Semantic structure with sub-goals

n against control

n Add the mince and carrot, and cook for about 8 minutes, stirring all the time, until the meat is nearly cooked.

Segmented instructions with goal-state pictures

n Add the pasta to the meat sauce with a couple of tablespoons of the pasta water.

n Mix well, sauté for a minute.

n Drain the pasta

Semantic structured recipe with sub-goals

Analysis ongoing

n Structured recipe with sub-goals

n rated easier to understand and better organised than control

n fewer errors

n Reliance on pictures - impact on validity

Conclusion

n Instruction design can offer insights recipe design

n But cooking is complex domain

n Results show were existing interactive recipe systems are not helpful ... can inform more sophisticated design

Future research

n Robust prototype to test in-home

n Investigate recipe designs to support deviations

n Work with semantic researchers to create on-the-fly transformations