symbolk / crowdjigsaw Goto Github PK

问题描述

用户游戏时，因为后台推荐算法过于宽容，所以对于用户曾经多次尝试的难点块，会多次推荐错误。

解决方案

改进后台推荐算法，设置推荐阈值，降低recall提升precision。

涉及代码

graph.js:getHints()

players开始时间不一致

实验中发现，每次round的creator点了开始后，自己会立刻进入游戏界面，但其他player都有3~10秒不等的延迟才能加载出puzzle界面。

可能原因

1，网络延迟
2，通信机制问题
3，客户端图片处理

可能涉及代码

客户端
1，jigsaw.js
2，main.js
服务器端
round.js

描述

出于安全考虑，游戏判断完成应该由服务器完成

改进

在每次checkLinks时，触发服务器端对：
1，当前player是否完成的判断；
2，当前群体是否完成的判断。

问题场景：

1. 片段x, y之间的一条边，具有权值w，表示有w个用户把这两个片段拼在了一起；
2. 用户u，没有反对把x,y拼在一起，目前也没有把两者拼在一起；
3. 如果软件把这条边推荐给了用户u，那么，这条边的权值什么时候变成 w +1 呢？

两种策略

• 策略一：在这条边推荐给u之后，边的权值立刻变成 w +1 。

• 策略二：在这条边推荐给u之后，观察u的行为，直到u做了一次非分离的操作后，如果此时u没有把x,y分开，这时，才把u的1添加到权值w上。

  ”非分离操作“：不是”分离操作“的操作

  “分离操作”：用户把两个原本拼在一起的片段分离开的操作（无论这个被拖开的片段是否被拖动到其它片段的某个边上）。

感觉策略二更好一些。

例如，对于http://pic36.nipic.com/20131126/10174664_152719649117_2.jpg这样一幅狗的图片，因其大片背景是白色，因此，不适合用于拼图。但是，如果把只包含背景颜色的片段删除后，则是可以用于拼图的。

The pictorial jigsaw puzzle is a well-known leisure game for humans that requires the player to recover a picture from a pile of pieces. Traditionally, human players solve the puzzle individually or collaboratively face-to-face. In this paper, we propose an approach to facilitate the human collective intelligence in solving such problems like the jigsaw puzzle. It is inspired by the stigmergy mechanism in biology and works as a continuous loop named EIF, which consists of three activities: exploration, integration, and feedback. As a verification of the approach, we design and implement a web-based system to support physically-distributed human players in solving the jigsaw puzzle more efficiently. Exploratory experiments show that: (1) for groups with 2~10 players, the time to solve jigsaw puzzles decreases 31.36%~64.57% on average when compared with the best
individual player in each group; (2) comparing with the face-to-face or shared-canvas collaboration, our approach provides better scalability with regard to the puzzle size and the number of players.

这三种浏览器都是很现代的浏览器，对标准的遵从度也很高。

如果一个软件包不能兼容这三种浏览器，那么就换一个可以兼容的软件包。这应该不是一件困难的事情。

问题描述

客户端游戏时，推荐的块会漂移很远

原因分析

jigsaw.js:showHints()对于推荐块的位置计算问题

描述

现在的图片多次试验后，熟悉度会影响公平性

改进

添加更多的经过选择的图片，比如CISE组的合照。

问题描述

round人数凑不够时，creator想重开round所以退出，但退出后应该解散round，因为其他players也不能start

可能原因

round结束判断的问题

涉及代码

round.js--quitRound

如题。

92-Operation-based Merging
arXiv13-Extracting Instances of Code Change Patterns with AST analysis
arXiv18-An Empirical Study of Android Changes in CyanogenMod
ASE05-UMLDiff An Algorithm for Object-oriented Design Differencing
ASE12-Structured merge with auto-tuning balancing precision and performance (FSTMerge)
ASE14-Fine-grained and Accurate Source Code Differencing (Gumtree)
ASE15-Balancing Precision and Performance in Structured Merging (JDime)
ASE16-Move-Optimized Source Code Tree Differencing
ASE17-Renaming and Shifted Code in Structured Merging
ASE18-ClDiff generating concise linked code differences
CCS12-Dynamically Scoped Change Sets for Controlled Updates in Live Programming
Combining Unstructured, Fully Structured and Semi-structured Information in Semantic Wikis
Comment and Indentation Preserving
CSCW18-An Analysis of Merge Conflicts And Resolutions in Git-Based Open Source Projects
CVSM09-Operation-based conflict detection and resolution
ESE18-Understanding semi-structured merge conflict characteristics in open-source Java projects
ESEM15-Assessing Semistructured Merge in Control Systems A Replicated Experiment
ESEM17-An Empirical Examination of the Relationship between Code Smells and Merge Conflicts
FASE17-Precise Version Control of Trees with Line-Based Version Control Systems
FSE11-Crystal Precise and Unobtrusive Conflict Warnings
FSE11-Semistructured Merge Rethinking Merge in Revision Control Systems
FSE11-Semistructured_Merge_in_Revision_Control_Systems
FSE12-How Do Software Engineers Understand Code Changes An Exploratory Study in Industry
FSE16-Why We Refactor Confessions of GitHub Contributors
ICSE16-SourcererCC Scaling Code Clone Detection to Big Code (SourcererCC )
ICSE18-Assessing the Threat of Untracked Changes in Software Evolution
ICSE18_Accurate and Efficient Refactoring Detection in Commit History (RefactoringMiner)
ICSM04-Dex A Semantic-Graph Differencing Tool for Studying Changes
ICSM04-Supporting Source Code Difference Analysis
ICSME17-Software Practitioner Perspectives on Merge Conflicts and Resolutions
Informatica17-Software Architectures Evolution Based Merging
ISEC16-Delta Refactoring for Merge Conflict Avoidance
IWST15-Assessing Semantic Merge Conflicts with Dependency Analysis (DeltaImpactFinder)
MasterTheis16-A Study of Merge-Conflict Resolutions in Open Source Projects
MSR05-Understanding Source Code Evolution Using AST Matching
MSR18-Analyzing Conflict Predictors in Open-Source Java Projects
MSR18-The Android Update Problem An Empirical Study
OOPSLA16-Purposes, concepts, misfits, and a redesign of git (Gitless)
OOPSLA17-Evaluating and Improving Semistructured Merge (jFSTmerge)
OOPSLA18-Conflict Resolution for Structured Merge via Version Space Algebra
OOPSLA18-Verifying Semantic Conflict-Freedom in Three-Way Program Merges (AutoMerge)
PHDThesis17-srcDiff Syntactic Differencing to Support Software Maintenance and Evolution
PL17-Verifying Semantic Conflict-Freedom in Three-Way Program Merges
PLDI11-Systematic Editing Generating Program Transformations from an Example
SANER16-Supporting Merge Conflict Resolution by Using Fine-Grained Code Change History
SATToSE17-Automated Categorisation of Breaking Merge Commits
SC08-Superimposition A Language-independent Approach to Software Composition
SDE92-Operation-based Merging
SPLASH17-What Merge Tool Should I Use
TOPLAS94-Software Merge-Semantics of Combining Changes to Programs
TSE02-A State-of-art Survey of Software Merging
TSE07-Change distilling Tree differencing for fine-grained source code change extraction (ChangeDistiller)
TSE08-Effective Software Merging in the Presence of Object-Oriented Refactoring
TSE17-Semantic Slicing of Software Version Histories
~$ASE16-Move-Optimized Source Code Tree Differencing
~$ICSM04-Dex A Semantic-Graph Differencing Tool for Studying Changes
~$OOPSLA18-Conflict Resolution for Structured Merge via Version Space Algebra

对于一个图片，在服务器端可以按照事先确定的几种分割方案，对其进行分割。

当一个玩家进入游戏，在等待游戏开始的这段时间内，服务器把分割后的图片（压缩在一个文件内）传递到客户端。

另外，应该无法从碎片的名称看出碎片间的关系（否则，在客户端，如果用户在控制台能够看到这些碎片，就有可能作弊）。

描述

round等待界面维护的用户列表太长，界面不友好

改进

只展示等待中players的名字，round信息放到一行，界面尽量简洁

描述

大家更关注round内的players排行

改进

增加round内排行榜，综合考虑贡献、步数、时间等。

描述

用触摸板游戏时，发现块选取必须按到中心附近，边缘不能很好识别。

改进

增加灵敏度，块边缘点击则选取最近的块。

在分割图片时，可以给定一个整型参数x，表示需要把分割后的边缘像素抹除掉多少行/列。

实现时，等价于：先在原始图片的分割线上画上宽度为2x的纯色线条；然后再对图片进行分割。