SoyViking [he/him]

  • 0 Posts
  • 65 Comments
Joined 4 years ago
cake
Cake day: November 4th, 2020

help-circle


  • I’m the guy on that airplane at the moment.

    I get specs for an external API to use in a new major feature. I begin implementing, the specs doesn’t add up because nobody paid the eastern European gig programmers to document anything. Eventually I derive plausible specs from a frustrating process of emails and trial and error.

    I implement the major feature to the specs provided by the client. The client tests in staging and requests several adjustments. I implement those, client tests again and accepts.

    The feature is pushed to production. The client finds a ton of errors because of course the rudimentary specs I managed to wrestle out of the client and the big-shot corner-cutting third party API developer didn’t describe half of the ideosyncratic data structure they send. Stuff like sending completely empty posts and expecting empty rows to be inserted in the database, sending text comments in fields intended for storing numbers instead of in the dedicated comment field. That sort of bullshit. They want to pour garbage in an have garbage coming out.

    So I had to do a ton of hotfixes directly to production. Everything have to be fixed yesterday because it is a business critical feature. It sucks. It’s a clusterfuck of cherry picking and it becomes impossible to do any sort of quality control.

    A ton of errors got introduced because nobody explained to the Eastern Europeans what the API should do or gave them the time to do it properly. I am the lead engineer on the project and I have to rush to make emergency bug investigations all the time. Most of the things the bug is that the Eastern Europeans didn’t set up their system like they were asked to or that nobody told them how it should work and just assumed they would know. I wrote more emails than code. The client pays a ton of money for all of this and nothing gets done because they rushed into this feature without planning it properly.



  • So breaking things up aggressively into small components you can reason about in isolation tends to be the best way to write reliable code you can maintain over time.

    This is so true. Something that has really improved my coding has been having a linter that whines to me about assignment branch condition size. Compared with learning how to properly stub methods in tests it has helped me break tasks down into simple manageable chunks with little room for error.



  • This has been my entire last week. Our client has a tight deadline for this feature I’m working on, in part due to their own indecisiveness and in part due to an external API developed by a big corporation being late and buggy. This means we’re doing testing and bug fixing simultaneously with doing new development and even with speccing and estimating new subtasks. And with this client, this close to the deadline, every little bug is critical and needs to be fixed right away. Meanwhile, a junior developer is being onboarded to the project and another developer is working on a different feature derived from an architecture I made. There’s always a fire I need to put out, a question I need to answer or a feature I need to describe. I’m writing more emails than code these days.


  • SoyViking [he/him]@hexbear.nettoProgrammer Humor@lemmy.mlJust a quick meeting
    link
    fedilink
    English
    arrow-up
    8
    arrow-down
    1
    ·
    2 months ago

    I normally go “what the fuck did I even do yesterday?” five minutes before daily standup and look at my git commits and calendar for the day before to piece together a plausible version of my workday (I do my timesheets the same way as well btw). Very little serious information gets passed on but somehow it makes my boss happy and he has told me that he likes the way I do standups.

    I work at a small company where most projects only have one or two developers so standup meetings are usually a lot of completely irrelevant information. It’s very boring. “Yesterday I worked on the thing on the project you barely know what does.”





  • Americans don’t have healthcare because their masters don’t want them to have it, not because the money is being spent elsewhere.

    The US has one of the world’s highest per capita public spendings on healthcare and the highest private. There are more money per capita in the American healthcare system than in any other system in human history. The money is there already. What is lacking is political will to make these money go towards actual healthcare instead of grifts, corruption and rent extraction.








  • I don’t know the details of how the US legal system works but isn’t a plea bargain essentially the same as a settlement in civil cases?

    If so, it should (at least in theory) have very little prejudicial value since the courts did not rule on the question if Assange’s culpability.

    I know that in the real world the US regime once again learned that it can get away with murder and journalists all over the world have already learned the lesson that the evil empire will fuck them up if they air their dirty laundry. But from a legal nerd point of view a settlement should be quite useles as a precedent.