Thomas Tuegel

It is not inconsistent to want tools that are very sharp, but only at one end.

@rahaeli > Any person who is running an online service of any kind who lives in the US Does that mean these recommendations would equally apply to someone running, e.g., a personal website? (Getting a DMCA notice doesn't require actual infringement, after all.)

@mattoflambda @ChShersh How does that address the sync exception problem?

@ChShersh Safe exceptions pattern uses an uninterruptible mask in clean up, which solve the last two cases

I was living my life, thinking that 'bracket' in #Haskell automatically solves all the RAII problems. 😨 Recently, I learned that 'bracket' can fail in at least 4 (!!!) ways and your probably don't even expect a half of them! Nowadays, I don't even know what to believe...

The Old God prepares for final battle

@mattoflambda @ChShersh Sounds to me like a worker versus manager question to me! 😉

@ChShersh Feels like a throughput vs latency question

⌛️ If 8 people spend 1 hour in a call to solve a task, do you say that the task took 8 hours or 1 hour in total to be solved?

@ericson2314_ @TechnoEmpress @acid2 @etorreborre @ChShersh In either case, you need an in-language interface type, IMO.

@ttuegel @TechnoEmpress @acid2 @etorreborre @ChShersh I would say one the system is monoglot we can have an in-language interface type, and when the system is polyglot we need a lingua franca interface type, which would be protobuf schema or whatever.

I am building a strong dislike for derived JSON and Pretty instances in #haskell. Because once they are defined there's no way to redefine them. You cannot even filter them out with carefully selected imports because they are imported transitively, possibly everywhere

@typed_hole Then "how" is a proprietary trade secret and it's definitely not by collusion between manufacturers to fix prices 😉

@ttuegel That sounds like who and not how :p

What is the purpose of a list price/MSRP? Like, who came up with it and why, who does it help, and why do market actors take it into account?

@ericson2314_ @TechnoEmpress @acid2 @etorreborre @ChShersh And consequently you have a hard time selling people on (3)-(4).

@ericson2314_ @TechnoEmpress @acid2 @etorreborre @ChShersh (1)-(3) are mechanical, on a "string template" level if you choose. My impression from talking to more experienced enterprise-y programmers is that in a mainstream language with IDE support, those steps are generally automatic. But in Haskell we tend to only automate (2)...

@typed_hole By the manufacturer, who sells to retailers at some percent below the MSRP.

@ttuegel Then how is the MSRP decided? 🤔

@ericson2314_ @TechnoEmpress @acid2 @etorreborre @ChShersh Yes, there should definitely be a second type because there are two things. Did it seem like I was saying otherwise? I agree with the practices in this thread. I'm questioning why we tolerate doing the automatic part by hand in Haskell when almost nobody else does, AFAIK.

@ttuegel @TechnoEmpress @acid2 @etorreborre @ChShersh Hmm? I think this is saying a second data type + deriving might be easier to read than hand-coded serializer/deserializer? (Over time it's many per version data types vs complicated manual parser, but still same basic trade-off)

@TechnoEmpress @acid2 @etorreborre @ChShersh I don't understand.

@acid2 @TechnoEmpress @etorreborre @ChShersh Why should we be doing boring work for the computer? By that I mean, why should we be writing the boring parts of adapters by hand?

@TechnoEmpress @etorreborre @ChShersh Well said - hard was the wrong word. I want boring and obvious.

@NikitaYVolkov @acid2 @ChShersh @etorreborre I whole-heartedly agree about similarity versus equality. Do you have any tools that support this workflow in Haskell?

@ttuegel @acid2 @ChShersh @etorreborre I do! ) It's much easier and safer to map even the awkward generated code in the adapter wrapper, than manually dealing with JSON. Most bugs get caught by the compiler. Also I never use the API types as domain types. Similarity does not mean equality. A very common pitfall

@ChShersh > We will note that they has been in consistent use as a singular pronoun since the late 1300s merriam-webster.com/words-at-play/…

"I learned English in school 20 years ago and I don't like that it has changed since then so you're all wrong" 🤡