Self-disclosure is a vital component in AI-driven mental health support, particularly through empathetic social chatbots. While prior HCI research has examined various strategies to facilitate disclosure, the impact of subtle, scalable interface cues—such as message bubble coloring—remains underexplored. Addressing this gap, the present study proposes a systematic bubble coloring design scheme grounded in color psychology and emotional regulation theory, with particular attention to the rationale for color selection. Two user experience experiments (N = 80 and N = 320) were conducted to evaluate its effectiveness. Results show that bubble coloring can enhance self-disclosure intentions by up to 40.81%, primarily by improving information processing fluency and perceived trust. Moreover, expressive suppression was found to moderate this effect. This research introduces a novel, easily implementable design strategy for social chatbot interfaces, contributing to the application of color theory and emotional regulation within the domain of human-computer interaction.

With the emergence of mixed source software, the existing quality models are not able to better assess the community quality and autonomy controllability of mixed source software. To fill this gap, we propose a hierarchical model in this paper for quality assessment of mixed source software. In our model, the new attributes are proposed to meet the quality requirements of mixed source software based on the ISO/IEC 25010 standards, and a set of metrics are designed for the new attributes. The model evaluates the quality of mixed source software through quality attributes that have been quantified by the metrics. Applying our quality model to some mixed source software and comparing the model results with the actual situation, we verify whether our proposed two quality attributes, community intensity and autonomy controllability, can effectively assess the quality of mixed source software. The results of the experiments show that our model is indeed effective in assessing the quality of mixed source software. An important feature of our model is that the model has good flexibility, and the set of quality attributes and metrics can be adjusted freely, which provides a flexible and feasible way for various software quality assessment requirements.