Summary
In this paper, the authors aim to study the needs of users for email automation and the resources required to achieve the automation. The authors goal is to design a good email automation system. They led a workshop to group the requirements into different categories, and they also conducted a survey using human computation to help understanding the users needs. After collecting all the requirements, the authors performed another study by reviewing an open source codebase available on GitHub to see which requirements already been met. After building and running the source code, they asked users to interact with the system to find out what is working well and what is not. They find out that there are limitations with the current implementation especially with complex requirements and lots of requirements are not being met. The authors hope that their findings can help future research to focus on the needs that are not met or satisfied yet.
Reflection
I found the method used by the authors to be very interesting. Conducting a survey and leading a workshop to find the users requirements and cross reference them with what is available and what is not with the current implementations is a nice approach to find out what is not implement yet.
I also like the idea of performing a code analysis on an open source project and link the analysis with user requirements. This approach can be used by software companies to search GitHub for current implementations of certain requirement rather than just searching a code implementation for a specific library or a tool.
I like the idea of email automation and I have used rules before to automatically move certain emails to special fodders. Nowadays most systems send automatic notifications and these notifications are necessary but sometimes it make it hard to distinguish between emails that need an immediate response versus emails that need a review at a later time. I also like that Gmail automatically move emails that has advertisements to different folders or different view to let the user focus on the important emails.
I agree with authors that there is a big room of improvements in the current implementation of email automation, but it will be interesting to know what will be the results if email systems like outlook, Gmail, and Yahoo have been deeply investigated to know what have been already implemented in these systems which was missing in the system that they have studied.
Questions
- The authors studied the current implementation using one system and over a week of time, do you think using more than one systems or study the user interactions over multiple weeks or months might lead to a different results?
- Do you think email automation can be used to send critical business emails that might accidentally includes some information that shouldn’t be sent? How can such systems overcome such issues?
- Have you used rules to automate email operations? Were they useful?
To answer your third question, I have used `filters’ and `labels’ to categorize and sort images. They do seem helpful.