,
30 tweets,
12 min read
Read on Twitter
The first remarkable information about her keynote is the special thanks to her family (husband and kids). Family first!
She is questioning the impact on our research in the software development. Part of our job is to develop an understanding on how developers work. We also create a loot of tools. However, developers already have a lot of tools.
She quoted this tweet from @elonmusk regarding the bad decision of automating a lot
@elonmusk There are different paradigms for research, such as postpositivism (the work that @timmenzies does), constructivism, and advocacy/participatory (the work that Margaret Burnett does). But how about you? What kind of papers do you like? #icse19
She did a very interact poll with the audience. However, it seems that the majority of #icse19 attendees are more likely to use postpositivism.
What kind of contributions do we see? She is now focusing on how design science can be used in software engineering research #icse19
Any software engineering paper that you look at can be seen through the view of design science. #icse19
Take the statement as an example: "To reduce errors, OSS projects use CI". This could be generalized as "To achieve an effect, in a given context use/do intervention." #icse19
Using the design science template in 38 ICSE distinguished papers (with a visual abstract representation: margaretstorey.com/blog/2017/11/0…), they found four clusters of papers: problem solution, solution validation, solution design, and descriptive. #icse19
There are also meta papers, that are not focused on helping developers, but on helping researchers. As an example, she mentioned SLRs as meta papers. #icse19
Only 13 of 25 of the distinguished papers published at ICSE from 2014 to 2018 considered stakeholders in the papers. #icse19
She is now grouping the papers into non-empirical and empirical papers. Focusing on the empirical papers, she is explaining methods such as lab/experiments, field studies, surveys, and in-slico papers (from the medical field). Field experiments are very high on realism. #icse19
Surveys and interviews are good on generalizability, however. Lab/experiments are good to control variables. #icse19
They then categorized 253 papers published at the technical track of ICSE. 195 of papers are focused on the data-strategy (36 focused on respondents, 26 in the field study). #icse19
37 of the 253 papers published at the ICSE main track in the last 5 years used triangulation practices. #icse19
158 of the 253 papers have no human at all as part of the human observation. These papers are too focused on technical artifacts. #icse19
Maybe there is too much effort on paper that focus too much on the technical side of the equation. She quoted Pearl and Mackenzie: "You are smarter than your data. Data do not understand causes and effects, humans do." from this book: amazon.com/Book-Why-Scien…
We may be creating silos in our community. As an example, she is mentioning that the @CHASEWorkshop and the @msrconf communities may need to talk more. #icse19
When we assign reviewers, we tend to assign based on the topics that we bid for. Maybe we need some ways to make sure we have a balance between different perceptions in the review process. #icse19
Replication is definitely important as well as triangulation is important. However, it is very hard to have all aspects covered in a single paper. #icse19
She is saying that one could hack the paper review process, if we know how to write a paper for a particular community. #icse19
The last of point is regarding the things that our community values. Badges (artifacts validated, etc) sends a fantastic message to newcomers #icse19.
She is now suggesting new badges. For instance, a "triangulation" badge, or a "developer tested" badge, or an "industry collaboration" badge. #icse19
She is closing with the message “write less, think more, imagine more”. But she also mentioned that she is a tenured professor ... (whatever it means) #icse19 
Thank you @margaretstorey for this fantastic keynote! #icse19
