摘要 :
Dogfooding refers to the idea of "eating your own dog food". In one of the largest players in the mobile industry, dogfooders use prototypes of unreleased products as their primary mobile device. This entails using on a daily basi...
展开
Dogfooding refers to the idea of "eating your own dog food". In one of the largest players in the mobile industry, dogfooders use prototypes of unreleased products as their primary mobile device. This entails using on a daily basis both hardware and software that is still under development and constant changes. For the past 4 years, Dogfooding has been a fundamental piece of the product development process for a number of reasons. Firstly, it allows us to cover unpredicted test scenarios that traditional software quality assurance methodologies do not. Moreover, the usage of prototypes in more consumer-like environments provides valuable data from which we can assess key aspects of the product such as battery life and connectivity. Lastly, dogfooders serve as a user base for collecting feedback on more subjective topics such as design and usability also allowing to detect regional idiosyncrasies that regular test cases would not. Despite all of these benefits, running the operations of Dogfooding is far from trivial. We have a total of over 4,000 dogfooders spread across 4 different sites in United States, Brazil, India and China. The Dogfooding programs need to run in sync with the development of new products as well as over-the-air upgrades. This paper aims to present the processes, tools and people responsible for running Dogfooding operations on a global scale. We'll also cover the lessons learned throughout the past 4 years and challenges that still lie ahead of us.
收起
摘要 :
Dogfooding refers to the idea of "eating your own dog food". In one of the largest players in the mobile industry, dogfooders use prototypes of unreleased products as their primary mobile device. This entails using on a daily basi...
展开
Dogfooding refers to the idea of "eating your own dog food". In one of the largest players in the mobile industry, dogfooders use prototypes of unreleased products as their primary mobile device. This entails using on a daily basis both hardware and software that is still under development and constant changes. For the past 4 years, Dogfooding has been a fundamental piece of the product development process for a number of reasons. Firstly, it allows us to cover unpredicted test scenarios that traditional software quality assurance methodologies do not. Moreover, the usage of prototypes in more consumer-like environments provides valuable data from which we can assess key aspects of the product such as battery life and connectivity. Lastly, dogfooders serve as a user base for collecting feedback on more subjective topics such as design and usability also allowing to detect regional idiosyncrasies that regular test cases would not. Despite all of these benefits, running the operations of Dogfooding is far from trivial. We have a total of over 4,000 dogfooders spread across 4 different sites in United States, Brazil, India and China. The Dogfooding programs need to run in sync with the development of new products as well as over-the-air upgrades. This paper aims to present the processes, tools and people responsible for running Dogfooding operations on a global scale. We'll also cover the lessons learned throughout the past 4 years and challenges that still lie ahead of us.
收起
摘要 :
Over the years, several problems regarding the analysis of face images have been addressed, including face detection, recognition, identification, and verification. The advent of Convolutional Neural Networks (CNNs) gave rise to a...
展开
Over the years, several problems regarding the analysis of face images have been addressed, including face detection, recognition, identification, and verification. The advent of Convolutional Neural Networks (CNNs) gave rise to a drastic improvement on state-of-the-art performances for these problems. With the increasing popularity of 360° cameras, the demand for models to extract relevant information from spherical images has also emerged. However, traditional CNNs, originally designed for planar images, are typically not suitable for spherical images, as it is necessary to project these spherical images onto a plane, leading to severe distortions. This work presents a method for face verification on spherical images that relies upon CNNs to extract features for training a binary classifier, as well as two new face datasets with spherical images. The effectiveness of our method is assessed through a comparative analysis with relevant planar and spherical CNNs.
收起
摘要 :
This paper proposes an automated test framework for eNodeB's physical layer development, comprising procedures for checking data integrity, stability and performance. The proposed framework is based on a simplified LTE MAC layer, ...
展开
This paper proposes an automated test framework for eNodeB's physical layer development, comprising procedures for checking data integrity, stability and performance. The proposed framework is based on a simplified LTE MAC layer, which operates as a software element that communicates directly with the physical layer and performs mapping procedures between logical and physical channels, reception and transmission of physical layer data, user data scheduling and data exchange with mobile terminals. All above procedures are performed with no further dependency on other LTE network element, thus providing a stand-alone test framework.
收起
原文获取