还有一点建议 我觉得如果能有一个对于此次地震主要数据介绍的poster可能更能给有兴趣的人一个更 直观的概念 比如那个中国地图上标明震中和多少地区有震感的图片 比如death confirmed和estimated death toll;number of people saved from rubbles in 72 hours,和after 72 hours我们自己救出来的和国外救援队就出来的人 数,我觉得这几个数听到的老美不少会赞叹不已,回去也会情不自禁的传播;10 million people affected;4.8 million people dislocated 可以拿面积和人数和美国人知道的一些数据作比较,这样他们可以更容易想象这次地震对普通人造成的伤害有多么巨大 again今天的事情让我知道,很多美国人真得不care世界其他地方和国家的任何事情。 但是也有一些人是关心的,有一个美国老头提到不能怪建筑水平,就算符合标准7级也 扛不了7.9级。如果我们能向这些关心地震的美国人澄清一些信息,他们可能会帮助我 们传播。我跟好几个人推荐了NPR all things cons
w*e
9 楼
pulp fiction印象最深刻的就是这段舞还有Samuel Jackson 在餐馆那段和抢匪说的话。
The problem is, it wont be that easy. I see a lot of cases, in which the PhD or graduate students ask for recommendation letters from supervisors but get refused, simply because they get graduated and no longer work there, regardless they have been contributing for years. You really need to be in great luck to meet a good advisor before starting anything. The risk is very high because you might end up gain nothing.
w*g
17 楼
这一段可以说是对John Travolta在77年的大热片【Saturday Night Fever】中的表演 的parody。 录像厅混大的昆汀大概对这段舞很喜欢。
【在 g*****g 的大作中提到】 : Of course it's not my problem, but these are the best practices in the : industry, and it's a fundamental one.
m*2
37 楼
我觉得是。
【在 w***z 的大作中提到】 : "Zed's dead baby, Zed's dead." ?
p*2
38 楼
对了 大牛你家跟google一个路子?
【在 g*****g 的大作中提到】 : Of course it's not my problem, but these are the best practices in the : industry, and it's a fundamental one.
e*9
39 楼
K,这片子印象最深的情节是大老黑被JJ和Travolta在马桶上被打死。
g*g
40 楼
It's not a G recommendation, it's a well known practice in the industry. It' s up to individual team to do what they want to do, but I always recommend best practices with reasons.
【在 g*****g 的大作中提到】 : It's not a G recommendation, it's a well known practice in the industry. It' : s up to individual team to do what they want to do, but I always recommend : best practices with reasons. : : 的。
Here is what I do, I write integration tests for the rest services, using Jersey test framework, https://jersey.java.net/documentation/latest/test- framework.html There are few reasons I do that 1) The service I am writing normally is micro service, so running a full round of integration tests won't take more than a few minutes. 2) I can run them locally before I push my commits, so bugs can be identified quickly and thus fixed quickly 3) Jenkins job also runs those tests, so again it's not likely the major bugs will be there when we release the service to QA 4) For our team, there is no dedicated QA. I simply don't have time to write both unit tests and integration tests.
【在 g*****g 的大作中提到】 : I am not gonna lecture the importance of unit test here. Quality and : quantity is always a tradeoff.
The pitfalls here are that 1. It may not be easy to identify the bug when integration test fails, especially if you didn't write the code in the first place. You are coupling data, other services dependencies and your code in your test. That makes the debugging process longer than it can be with unit tests. Unit tests won' t catch all your bugs, but it can catch most of your bugs. 2. And integration tests can typically cover happy path, not much on corner cases. And if you try to cover corner cases, it starts getting too long to run your build. Yet corner cases always hit in production. Again, it's a tradeoff, we all cut corners where we have to, I am just talking about industry best practice which typically favors quality over quantity.
【在 w**z 的大作中提到】 : Here is what I do, I write integration tests for the rest services, using : Jersey test framework, https://jersey.java.net/documentation/latest/test- : framework.html : There are few reasons I do that : 1) The service I am writing normally is micro service, so running a full : round of integration tests won't take more than a few minutes. : 2) I can run them locally before I push my commits, so bugs can be : identified quickly and thus fixed quickly : 3) Jenkins job also runs those tests, so again it's not likely the major : bugs will be there when we release the service to QA
c*o
55 楼
哈哈,对john的一贯评价就是,虽然美国人民很喜欢咸鱼翻生 come back story, 但是 曲沃塔到底要翻几次阿。。。
【在 g*****g 的大作中提到】 : It's not a G recommendation, it's a well known practice in the industry. It' : s up to individual team to do what they want to do, but I always recommend : best practices with reasons. : : 的。
In most of the cases, it's a wishful thinking. I an simply not given enough time to mock all the outside dependencies. As goodbug said, it's all about tradeoff, that is the choice as an organization you need to make.
A bug cost 1 to be found and fixed by dev, 10 to be found by QA, and at least 100 to be found by customer, just think about the communication costs, let alone implicit churn. So while unit test is against intuition for agile development, it's cost- effective for services that will be maintained for years. I found front end engineers are less used to unit tests. For one it's harder to do, for two many of them are used to manual testing.
costs, end unit test只是保证quality的其中一种办法,而且是最不重要的一种办法。
【在 g*****g 的大作中提到】 : A bug cost 1 to be found and fixed by dev, 10 to be found by QA, and at : least 100 to be found by customer, just think about the communication costs, : let alone implicit churn. : So while unit test is against intuition for agile development, it's cost- : effective for services that will be maintained for years. I found front end : engineers are less used to unit tests. For one it's harder to do, for two : many of them are used to manual testing.
n*n
76 楼
有哪些办法?重要性排序依次是?
【在 p*****2 的大作中提到】 : : costs, : end : unit test只是保证quality的其中一种办法,而且是最不重要的一种办法。