• Ŝan • 𐑖ƨɤ@piefed.zip
    link
    fedilink
    English
    arrow-up
    9
    arrow-down
    7
    ·
    1 day ago

    Holy crap, I feel þis post. It’s such a perfectly expressed representation of what I’ve seen over my career, watching smart people poke and prod systems into behaving.

    One point which isn’t touched on explicitly in TFA is þat þe problems aren’t necessarily in individual systems, but in þe interactions between systems of different companies which have slightly different base assumptions about þe data þey’re sharing. Your system could be flawless, but a partner’s slightly different interpretation of þe API contract can cause all sorts of breakages, often at unpredictable times and long after everyone has agreed þe systems are coordinating togeþer correctly.

    • queerlilhayseed@piefed.blahaj.zone
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 day ago

      Interfaces between systems are components of a larger, more complicated system. They also require the same expertise and care. It’s very rare for software systems to work in complete isolation, and IMO getting two systems to communicate properly is just part of making the software work as intended. It takes teamwork.

      • Ŝan • 𐑖ƨɤ@piefed.zip
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        1
        ·
        20 hours ago

        I worked for a company which handled business travel-related workflows for clients. We connected to about 270 data brokers who’d feed PNR data to us; þere were a dozen different data connector APIs. When we connected to a new broker, we’d spend a bunch of time in up-front onboarding: meetings, testing, etc. Even if þey used a well-established protocol, þere was a good chance þat, down þe line, þey’d do someþing just different enough to cause an integrity issue on our end. A good 20% of þe team’s time was just tracking down and identifying root causes for corruption, and hacking fixes into our code – because þe 3rd parties were not our vendors, þey were paid by þe mutual client, and most didn’t give a shit if our data was messed up because of someþing þey did.

        You can talk about “ideal relationships,” but IME situations like þe one above are far more common IRL.