Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Code Review as a Service

pullrequest.com

Post reply on HN
Print view View original

31–40 of 238 posts

Re: Code Review as a Service

#31

Post by tialaramex »

gitfan86 wrote: post #21
hankchinaski wrote: post #7

Having on-demand engineers look at code, without broader context on the project it is in my view the same as something that can be automated either or both via static analysis and custom ci/cd workflow checks. This can probably makes sense on a project with more junior engineers where many basic improvements can be supposedly suggested without needing broader context? I would be keen on hearing the use case

A lot of people use ESLint and Prettier. If you look at the roadmap/issues for those products you'll see that some feature requests cannot be programed, but can be understood without knowing the full context of the business and codebase. At the same time, if you are fixing a critical bug and you use enum instead of boolean, who cares, just push the fix, but it doesn't hurt to have a gentle reminder show up in your gi…

In a decent language enumerated types are strictly better than booleans because they have names. Closed and Open, or Paid and Outstanding, or Safe and Dangerous are better than true and false unless you really meant true and false.

my_bills(Paid) avoids the step of reading the function prototype to find out what my_bills(true) means.

In some languages you can make piecemeal changes to upgrade from booleans because the language is happy to silently coerce between a two state Boolean and a two state enumerated type. This is probably bad news for correctness because it means my_bills(Open) might not even raise a warning but it's convenient.

Re: Code Review as a Service

#33

Post by SomeHacker44 »

cannonpalms wrote: post #27

This entirely misses the point of code review. Say it with me: Code review is a knowledge transfer exercise. Finding bugs, security vulnerabilities, and keeping the code maintainable are merely side effects that we appreciate along the way. The primary purpose of code review is increasing the bus factor of the given piece of code and facilitating organic knowledge transfer. That's it.

Maybe for you. For others it can have all or any of those other things as its primary purpose. One important thing for me is that it prevents or at least mitigates unilateral insider attacks by having two people required to change code.

Re: Code Review as a Service

#34

Post by pelasaco »

cannonpalms wrote: post #27

This entirely misses the point of code review. Say it with me: Code review is a knowledge transfer exercise. Finding bugs, security vulnerabilities, and keeping the code maintainable are merely side effects that we appreciate along the way. The primary purpose of code review is increasing the bus factor of the given piece of code and facilitating organic knowledge transfer. That's it.

"but they have AI".. probably not. Probably the same linters and scanners that we all know + cheap working force reviewing code manually.

Re: Code Review as a Service

#35

Post by lambic »

pelasaco wrote: post #30
Recursing wrote: post #22

Earlier quoted context omitted.

From the FAQ: > Reviewers earn anywhere between $50 and over $3,000/week. Earnings are based largely on the amount of time spent reviewing on the platform and the type of code being reviewed. PullRequest’s payment rates are comparable to those of a senior-level engineer based in the US. > PullRequest issues weekly payments based on review activity during the preceding week. The time that you spend reviewing is tracke…

> Reviewers earn anywhere between $50 and over $3,000/week. They are operating just like a freelancing platform, right? How many hours is a "week" for them? 40 hours? > PullRequest’s payment rates are comparable to those of a senior-level engineer based in the US. Like include life and health insurance, paid vacation, profit-sharing, a generous signing bonus, and more?

Why don't you say what you're thinking instead of disguising your opinions as questions?

Re: Code Review as a Service

#36

Post by noduerme »

adamnemecek wrote: post #16
hankchinaski wrote: post #7

Having on-demand engineers look at code, without broader context on the project it is in my view the same as something that can be automated either or both via static analysis and custom ci/cd workflow checks. This can probably makes sense on a project with more junior engineers where many basic improvements can be supposedly suggested without needing broader context? I would be keen on hearing the use case

I have reviewed for pullrequest and you’d be surprised how many things one can fix. Why do you think that you don’t have context? You can look at the whole project and see what’s up. And no, you can’t write programs to do this.

It's a rare case indeed where I'd be willing to let a third party in on business logic plus the whole source code for something I was developing. I love talking shop but if I'm really having a problem with a structural issue, it's hard to imagine someone outside giving any better advice than people inside who understand the lay of the land. Meanwhile, it's a pretty much totally unacceptable security risk. So what's the upshot to this over an internal code review?

Re: Code Review as a Service

#37

Post by mfashby »

cannonpalms wrote: post #27

This entirely misses the point of code review. Say it with me: Code review is a knowledge transfer exercise. Finding bugs, security vulnerabilities, and keeping the code maintainable are merely side effects that we appreciate along the way. The primary purpose of code review is increasing the bus factor of the given piece of code and facilitating organic knowledge transfer. That's it.

Hmm, what about education? I've learned a bunch through thoughtful reviews of my code by other devs.

Re: Code Review as a Service

#38

Post by loloquwowndueo »

mfashby wrote: post #37
cannonpalms wrote: post #27

This entirely misses the point of code review. Say it with me: Code review is a knowledge transfer exercise. Finding bugs, security vulnerabilities, and keeping the code maintainable are merely side effects that we appreciate along the way. The primary purpose of code review is increasing the bus factor of the given piece of code and facilitating organic knowledge transfer. That's it.

Hmm, what about education? I've learned a bunch through thoughtful reviews of my code by other devs.

That’s them transferring their knowledge to you :)

Re: Code Review as a Service

#40

Post by jph »

Good code reviews are superb for helping teams accelerate into new technology areas, frameworks, languages, and integrations.

As a top of mind example, when a team wants to spike on migrating from C++ to Go or Rust, including using libraries and porting services, then I see very high value in paying for skilled contractors to do code reviews-- because what your team is gaining on-demand upskilling.

Post reply on HN