4 ms·

> programmers repairing shoddy code How is this pointless?

_eze_ 8y ago
How is it avoidable? I feel like this is just pretending that perfect code is possible.
jackpirate 8y ago
The article is pretty clear that the author thinks that the code should have been written correctly to begin with. It's the fact that management prevented this from happening that causes people to have to fix the code. The idea is that if the job was "done right the first time" then the total labor expenses would be much, much lower.
- ganeshkrishnan 8y ago
  Code can never be written "correctly" whatever that means. Even if it was, library changes will break it. And none of the other jobs mentioned are bullshit. They are all required in various degrees. I call it bullshit book
  - kranner 8y ago
    Surely you'd agree that there is a spectrum of shoddiness (and hence required maintenance) of code? The specific example of shoddy code given in the book is that of a programmer gradually replacing a completely ineffectual algorithm written by an incompetent academic who also happened to be his boss. The algorithm was designed to "mimic speech" (this part is not very clear in the book) and the programmer replaces it with Eliza-like heuristics. You should read the book before judging it so harshly.
  - MaulingMonkey 8y ago
    Even when cleaning up preventable messes in code, those messes were often caused or made worse by trying to "do it right the first time" and tacking on a bunch of extra complexity in the process instead of just doing the simple thing that works for now. Start small. Iterate. Do it wrong. Embrace the fact that we're imperfect, not oracles of the future. I'll gladly prefer fixing up some shoddy code over needing to tear down and re-architecture from scratch a larger, even well designed system, that solves the wrong problem. Not all preventable problems are worth preventing. Preventing them causes other problems, if only the problem of wasting a lot of time fixing many non-problems that you mistakenly thought were potential problems in an attempt to do so.
- js8 8y ago
  Another example is maintaining home-grown code that could be replaced by an off-the-shelf solution, but nobody is willing to pay the price for it or risk change.
michaelbuckbee 8y ago
Maybe it would be better stated as a "shoddy approach" to fixing business problems with technology. Some examples might be: - Overengineering. I worked for a startup that committed to a $20k/mo multi-az setup with mass redundancy for what was essentially a lightly trafficked static website. - Not-Invented-Here syndrome. Different place spent months building out a custom shopping cart and checkout solution for sales b/c they didn't really like the built-in options. Full on high priced design team (not a person, a team), UX folks, front end devs, trying to get PCI compliance on their servers (a huge task), etc. or they could have done what Tesla did and just used Shopify - https://shop.tesla.com/us/en.html https://shop.tesla.com/us/en.html In both those examples, there's no need/value to the work being done over similar solutions making it a "bullshit job".
- collyw 8y ago
  These are creating shoddy code not fixing it.

How is it avoidable? I feel like this is just pretending that perfect code is possible.

The article is pretty clear that the author thinks that the code should have been written correctly to begin with. It's the fact that management prevented this from happening that causes people to have to fix the code. The idea is that if the job was "done right the first time" then the total labor expenses would be much, much lower.

ganeshkrishnan 8y ago
Code can never be written "correctly" whatever that means. Even if it was, library changes will break it. And none of the other jobs mentioned are bullshit. They are all required in various degrees. I call it bullshit book
- kranner 8y ago
  Surely you'd agree that there is a spectrum of shoddiness (and hence required maintenance) of code? The specific example of shoddy code given in the book is that of a programmer gradually replacing a completely ineffectual algorithm written by an incompetent academic who also happened to be his boss. The algorithm was designed to "mimic speech" (this part is not very clear in the book) and the programmer replaces it with Eliza-like heuristics. You should read the book before judging it so harshly.
- MaulingMonkey 8y ago
  Even when cleaning up preventable messes in code, those messes were often caused or made worse by trying to "do it right the first time" and tacking on a bunch of extra complexity in the process instead of just doing the simple thing that works for now. Start small. Iterate. Do it wrong. Embrace the fact that we're imperfect, not oracles of the future. I'll gladly prefer fixing up some shoddy code over needing to tear down and re-architecture from scratch a larger, even well designed system, that solves the wrong problem. Not all preventable problems are worth preventing. Preventing them causes other problems, if only the problem of wasting a lot of time fixing many non-problems that you mistakenly thought were potential problems in an attempt to do so.
js8 8y ago
Another example is maintaining home-grown code that could be replaced by an off-the-shelf solution, but nobody is willing to pay the price for it or risk change.

michaelbuckbee 8y ago

Maybe it would be better stated as a "shoddy approach" to fixing business problems with technology. Some examples might be: - Overengineering. I worked for a startup that committed to a $20k/mo multi-az setup with mass redundancy for what was essentially a lightly trafficked static website. - Not-Invented-Here syndrome. Different place spent months building out a custom shopping cart and checkout solution for sales b/c they didn't really like the built-in options. Full on high priced design team (not a person, a team), UX folks, front end devs, trying to get PCI compliance on their servers (a huge task), etc. or they could have done what Tesla did and just used Shopify - https://shop.tesla.com/us/en.html https://shop.tesla.com/us/en.html In both those examples, there's no need/value to the work being done over similar solutions making it a "bullshit job".

collyw 8y ago
These are creating shoddy code not fixing it.