Bad, Better, Best in IT

最新推荐文章于 2020-05-20 09:35:46 发布

giga_Zhang

最新推荐文章于 2020-05-20 09:35:46 发布

阅读量302

点赞数

本文链接：https://blog.csdn.net/giga_Zhang/article/details/84387343

版权

Bad: Cubicle

Better: Office

Best: Whatever works best for you.

Bad: Meetings without agendas.

Better: Meetings with agendas.

Best: Meetings whose need is so obvious to everyone that no agenda is needed.

Bad: Specs, waterfall, Systems Development Life Cycle.

Better: Prototyping, agile, scrum.

Best: Developers with enough domain knowledge to just build it.

Bad: No documentation.

Better: Documentation.

Best: No documentation needed.

Bad: No formal process.

Better: Formal process.

Best: People so much bigger than their jobs so that process is rarely

relied upon.

Bad: Theory without experience

Better: Experience without theory

Best: Both

Bad: Help desk without programmers.

Better: Programmers available to customers.

Best: Code that just works.

Bad: Phone calls

Better: Emails

Best: Application software that encapsulates required communication

Bad: Code with early exits.

Better: Code without early exits.

Best: Code so simple because of the underlying data structure that the

early exit debate is moot.

Bad: Bugs

Better: Fixes

Best: Enough 9's to never notice.

Bad: programmer error

Better: user error

Best: What's an error?

Bad: Missing deadlines

Better: Hitting deadlines

Best: A track record so good that deadlines are never given

Bad: Complex org chart

Better: Simple org chart

Best: Technology so sophisticated, less people are needed

Bad: Non-technical boss

Better: Technical boss

Best: No boss

Bad: Management

Better: Leadership

Best: Self-motivation

Bad: Best practices, with a Capital "B" (industry standards)

Better: best practices, with a small "b" (what we figured out)

Best: Just do your fucking job.

确定要放弃本次机会？

福利倒计时

: :

立减 ¥

普通VIP年卡可用

立即使用

giga_Zhang

关注关注

0
点赞
踩
0

收藏

觉得还不错? 一键收藏
0
评论
分享

复制链接

分享到 QQ

分享到新浪微博

扫一扫
举报

举报

数据简化DataSimp社区

08-01

8681

《Better PHP Development》- 2017 英文原版

11-30

It is this ease of access that can, admittedly, sometimes give it a bad name - good resources are few and far in between, and the competition is real, driving people to take shortcuts, skip best ...

参与评论您还未登录，请先登录后发表或查看评论

bad-worse-worst VS good-better-best

mmpire之blog

03-31

1676

Taking too much things into account, will yield a rather confused mess.Devote your life to your country, your career. As a man, its my honor.

Code Review Checklist and Best practices in Java

Stroller

04-28

1373

Code Review Checklist and Best practices in Java What to review while doing Code review Code Review and Unit testing is one of best development practice I always recommend, strive for an

18 Best SEO Tools that SEO Experts Actually Use in 2019

Larry的博客

04-13

2870

18 Best SEO Tools that SEO Experts Actually Use in 2019 Best Search Engine Optimization (SEO) Tools Ahrefs: SEO Keyword Tool Ahrefs is one of the most recommended SEO tools online. It’s only ...

Exception Handling Best Practices in .NET

心如止水--coofucoo的专栏

10-14

3484

ContentsIntroductionPlan for the worstCheck it earlyDont trust external dataThe only reliable devices are: the video, the mouse and keyboard.Writes can fail, tooCode SafelyDont throw new Exception()

Java - JDBC Best Practices

热门推荐

了解➔熟悉➔掌握➔精通

03-18

2万+

分享一个大牛的人工智能教程。零基础！通俗易懂！风趣幽默！希望你也加入到人工智能的队伍中来！请点击http://www.captainbed.net In this guide, I would like to explain JDBC Best Practices. We can follow these best practices in the day to day project work...

Tutorial: Bad Geometry – How to find and fix it

wanglang3081的专栏

04-20

2073

In this tutorial I would like to talk about the Cleanup Command and what is considered bad polygonal geometry. Generally there seems to be a whole lot of confusion as how to use the Cleanup command an

PostgreSQL in & = any (values|array)

u014042758的博客

03-19

6667

11 Essential Skills to become Software Developer in 2020

AI天才研究院

05-20

9395

Key skills for programmers and software developers to learn in 2020 In this article, I’ll share 11 skills, which I believe, every programmer should know or learn in 2020. This includes a programming...

Better PHP Development epub

02-13

Refactoring JavaScript: Turning Bad Code Into Good Code [Kindle Edition]

03-15

Refactoring JavaScript: Turning Bad Code Into Good Code by Evan Burchard English | 13 Mar. 2017 | ASIN: B06XK1V629 | ... And making it better doesn’t have to be intimidating or unreasonably expensive.

2013年八年级英语上册 Unit 4 What’s the best movie theater精解与精练（新版）人教新目标

08-19

2. "Blue Moon is better (good), but Miler’s is best (good) in town." 使用比较级和最高级比较两家店。 3. "The Big Screen is more expensive (expensive) than most cinemas, but Cinema City is the most ...

实训商业源码-130个微信小程序源码-论文模板.zip

05-18

实训商业源码-130个微信小程序源码-论文模板.zip

光伏交直流混合微电网系统双下垂控制Matlab Simulink仿真模型：孤岛模式下离网运行与策略分析微电网

05-18

内容概要：本文详细介绍了光伏交直流混合微电网系统的离网（孤岛）模式下双下垂控制的Matlab/Simulink仿真模型。该系统由直流微电网、交流微电网和互联变换器（ILC）组成，分别采用了不同的控制策略来确保系统的稳定性。直流微电网采用电压电流双闭环控制，交流微电网则利用LCL逆变器进行下垂控制，而ILC通过双下垂控制策略协调交流母线频率和直流母线电压。仿真结果显示，在负载突变情况下，系统能够迅速恢复并保持稳定运行，展示了良好的鲁棒性和波形质量。适合人群：对电力电子、微电网技术和Matlab仿真感兴趣的科研人员和技术开发者。使用场景及目标：适用于研究和开发光伏交直流混合微电网系统的离网控制策略，特别是双下垂控制的应用。目标是在不同负载条件下验证系统的稳定性和效率，优化控制参数以提高系统性能。其他说明：该仿真模型需要Matlab 2020b及以上版本，并使用SPS工具箱中的特定模块。文中提供了详细的控制参数设置和仿真结果分析，有助于深入理解各组件的工作原理及其协同效应。

基于SMIC 180nm工艺的高效10bit 20MHz SAR ADC设计与仿真实践指南 Verilog

05-18

内容概要：本文深入解析了基于SMIC 180nm工艺的10bit 20MHz SAR（Successive Approximation Register）ADC设计及其仿真实现。文中详细介绍了各个关键模块的设计原理和技术细节，如栅压自举开关、Vcm-Based开关时序、差分CDAC阵列、两级动态比较器、异步时钟生成以及动态SAR逻辑等。此外，提供了完整的仿真文档、电路图和预设好的仿真参数，使读者能够快速上手并进行实验验证。文章还特别强调了设计中的优化技巧，如通过动态电平转换降低THD、利用Vcm-Based时序减少冗余比较周期、优化比较器时钟脉冲宽度等，确保设计的高性能和稳定性。适合人群：对模拟集成电路设计感兴趣的初学者和有一定经验的研发人员，特别是希望深入了解SAR ADC设计原理和仿真实践的人群。使用场景及目标：① 学习SAR ADC的基本原理和设计方法；② 掌握基于SMIC 180nm工艺的具体实现细节；③ 进行仿真实验，验证设计效果并理解各模块的工作机制；④ 提升对低功耗、高精度ADC设计的理解。其他说明：本文不仅提供理论讲解，还包括大量实用的Verilog和Matlab代码片段，便于读者理解和实践。同时，预设的蒙特卡洛仿真参数有助于读者探索工艺偏差对性能的影响。

基于聚类特征随机凸组合的非对抗图像生成算法

05-18

基于聚类特征随机凸组合的非对抗图像生成算法本研究旨在提供一种从给定数据集生成新图像的方法，而无需训练任何生成对抗网络（GAN）。其想法是使用主成分分析（PCA）算法提取数据集的特征，寻找图像像素空间和特征空间之间的线性变换，执行降维，并对结果集进行聚类。然后，通过随机凸组合生成新的特征向量，随后通过应用相应变换的逆将其映射回图像像素空间。拼音双语对照笔记

实训商业源码-文件快递-论文模板.zip

最新发布

05-18

实训商业源码-文件快递-论文模板.zip

导航猫v2.5.0源码系统全解密+去授权开心版

05-18

更新说明： 1.新增加前台用户自助购买广告功能 2.新增加自动审核功能 3.自动定时提交百度收录！ 4.前台模板1套、置顶在线购买，接口对接易支付） 5.【修复】修复在线支付产生的BUG 环境要求：PHP7.0 安装sg11拓展

AUC-ROC score

01-17

### AUC-ROC Score in Machine Learning #### Definition of AUC-ROC The Area Under the Curve - Receiver Operating Characteristic (AUC-ROC) is a performance measurement for classification problems at various thresholds settings. ROC curves plot True Positive Rate (TPR) against False Positive Rate (FPR). The TPR is synonymous with sensitivity or recall, while FPR can be calculated as one minus specificity. An ideal point on this curve would achieve maximum true positives and minimum false positives. #### Calculation Method To calculate the AUC value, multiple methods exist but generally involve plotting points where each represents different threshold values from which predictions are made by classifiers. For every possible cutoff between classes, compute both rates mentioned above until all unique probability scores have been used up. Then connect these dots smoothly to form an area under such formed curve that ranges theoretically within [0,1]. Higher areas indicate better overall ability across any chosen discrimination boundary without needing prior knowledge about what exact level might work best during deployment phases later down line[^1]. ```python from sklearn import metrics import matplotlib.pyplot as plt # Assuming y_true contains actual labels and y_pred_prob contains predicted probabilities. fpr, tpr, _ = metrics.roc_curve(y_true, y_pred_prob) auc = metrics.auc(fpr, tpr) plt.plot(fpr,tpr,label="data 1, auc="+str(auc)) plt.legend(loc=4) plt.show() ``` #### Application Scenarios In practical applications like financial lending described elsewhere[^2], AUC helps assess how well models distinguish between good borrowers who repay loans versus bad ones likely defaulting over time frames agreed upon contractually beforehand. By evaluating through cross-validation techniques repeatedly split datasets into training/testing sets ensuring robustness generalizability outside seen samples only thus far encountered previously throughout experimentation stages conducted internally before going live externally facing customers directly eventually after thorough testing cycles completed satisfactorily meeting predefined criteria set out initially when project commenced planning phase started originally some period ago now coming closer towards final implementation soon hopefully barring unforeseen complications arise unexpectedly causing delays pushing back timelines further than anticipated currently planned schedule allows ideally speaking optimistically looking forward positively ahead into future prospects lying just beyond immediate horizon awaiting us all together collectively moving forward progressively step-by-step cautiously yet confidently nonetheless steadily marching onward toward achieving ultimate goals objectives aims targets milestones benchmarks whatever terminology preferred most aptly describes desired end results sought after ultimately here today discussing specifically around topic area concerning itself primarily centered upon machine learning evaluation metric known formally recognized widely accepted commonly referred simply put as "AUC". --related questions-- 1. What other evaluation metrics complement AUC-ROC? 2. How does class imbalance affect AUC interpretation? 3. Can you explain the difference between precision-recall curves and ROC curves? 4. In which scenarios should one prefer using AUC over accuracy?