3 ms·

First reaction: Has this person been my coworker? I feel like someone’s been taking notes on places I’ve worked. Second: But OK, those ones about shifting prio

First reaction: Has this person been my coworker? I feel like someone’s been taking notes on places I’ve worked. Second: But OK, those ones about shifting priorities? That’s startup life. Sometimes you throw something out the door because a customer wants to pay you $12M if you can salve over that one pain point. You know darn well that hack’s going to be there until The Rewrite (which, if you’re lucky, will never come). It’s ugly, you’re ashamed to have your name on the git blame, but your stock options went up 8% the moment you deployed to prod. The customer rarely pays you to be an artisan. They just want that feature before they write the check. I’m still highly sympathetic to the sense of pride and ownership we all want to feel over our work. That’s the best! But getting paid’s pretty cool, too. Get a hobby project you can polish into a masterpiece after work.

dcminter 2y ago
> OK, those ones about shifting priorities? That’s startup life. I've seen it at pretty much all scales. I think it's usually worse in bigger corps actually. Great point though - it behoves us to remember where the money for our technical airs and graces comes from.
rswail 2y ago
If you're following the bullshit that is SAFe/"Agile" at a startup, you're going to fail. Capital-A Agile is project management for dummies that won't learn about things like GANTT charts, what "slack" means in links between dependent tasks, what a "critical path" is or how to analyze it, let alone things like S-curves and earned value.
- inetknght 2y ago
  > GANTT charts Can you give me an estimate for these sixteen different things, and also verify which ones depend on each other?
  - hinkley 2y ago
    The Gantt is a lie to get the project approved and then beg forgiveness later. Nobody ever figures bus numbers into the chart, because it would make the timelines so much worse that management wouldn’t approve it. But the blame for going over the estimates rolls downhill. Those stupid devs can’t get anything done on time. Thankfully I haven’t had to look at one of those stupid things in about 20 years. They used to make me see red.
    - rswail 2y ago
      When techniques are misused they are misused. The fact that there are bad project managers that don't know how to use a schedule are just as bad as the crappy "scrum masters" and "product managers" that don't know what "agile" actually means. A good project manager will be continuously revising their schedule, based on updated estimates on the tasks. They will know how to compare it to, and update, the baseline. Not sure what you mean specifically by "bus number" but happy to learn. The GANTT chart is just a format of displaying tasks, their resources and dependencies. The actual DAG is a pure data structure. If you put a confidence level indication on the estimates, a good project management tool will also be able to do analysis of the critical path (ie the length in time of the longest path) with a level of confidence on earliest/latest predicted end dates. People think that nonsense like "t-shirt sizing" actually means anything other than "wild-assed guess" or that burndown charts are useful when adjusting future estimates. All techniques to avoid doing the hard work of scheduling and project management.
  - rswail 2y ago
    Tell me you don't know how to do a WBS rollup and dependency linking without telling me. And if you don't know how they are dependent on each other, then how the hell are you estimating each of the 16 different things anyway?
- kstrauser 2y ago
  Kids these days barely know how to make a proper waterfall model before they start writing. It’s disgraceful.
hinkley 2y ago
I have worked at places that prioritized the wrong customer and ended up with a product that was tuned to work only for the ones who won’t pay more than what we originally offered the product for. Even inflation adjusted I’m sure it was less than $12M. You have to watch what your costs per year end up being for that check you got three years ago. The VCs will notice when they look at your books.
- kstrauser 2y ago
  That’s all true, but it works both ways. I’ve also been at a place that stuck to its plans rather than incorporating customer feedback and requests. You haven’t heard of it, and that’s why. I guess my main point is that you have to always be making the right thing. Sometimes that changes, and you have to be ready to pivot quickly.

dcminter 2y ago

> OK, those ones about shifting priorities? That’s startup life. I've seen it at pretty much all scales. I think it's usually worse in bigger corps actually. Great point though - it behoves us to remember where the money for our technical airs and graces comes from.

rswail 2y ago

If you're following the bullshit that is SAFe/"Agile" at a startup, you're going to fail. Capital-A Agile is project management for dummies that won't learn about things like GANTT charts, what "slack" means in links between dependent tasks, what a "critical path" is or how to analyze it, let alone things like S-curves and earned value.

inetknght 2y ago
> GANTT charts Can you give me an estimate for these sixteen different things, and also verify which ones depend on each other?
- hinkley 2y ago
  The Gantt is a lie to get the project approved and then beg forgiveness later. Nobody ever figures bus numbers into the chart, because it would make the timelines so much worse that management wouldn’t approve it. But the blame for going over the estimates rolls downhill. Those stupid devs can’t get anything done on time. Thankfully I haven’t had to look at one of those stupid things in about 20 years. They used to make me see red.
  - rswail 2y ago
    When techniques are misused they are misused. The fact that there are bad project managers that don't know how to use a schedule are just as bad as the crappy "scrum masters" and "product managers" that don't know what "agile" actually means. A good project manager will be continuously revising their schedule, based on updated estimates on the tasks. They will know how to compare it to, and update, the baseline. Not sure what you mean specifically by "bus number" but happy to learn. The GANTT chart is just a format of displaying tasks, their resources and dependencies. The actual DAG is a pure data structure. If you put a confidence level indication on the estimates, a good project management tool will also be able to do analysis of the critical path (ie the length in time of the longest path) with a level of confidence on earliest/latest predicted end dates. People think that nonsense like "t-shirt sizing" actually means anything other than "wild-assed guess" or that burndown charts are useful when adjusting future estimates. All techniques to avoid doing the hard work of scheduling and project management.
- rswail 2y ago
  Tell me you don't know how to do a WBS rollup and dependency linking without telling me. And if you don't know how they are dependent on each other, then how the hell are you estimating each of the 16 different things anyway?
kstrauser 2y ago
Kids these days barely know how to make a proper waterfall model before they start writing. It’s disgraceful.

hinkley 2y ago

I have worked at places that prioritized the wrong customer and ended up with a product that was tuned to work only for the ones who won’t pay more than what we originally offered the product for. Even inflation adjusted I’m sure it was less than $12M. You have to watch what your costs per year end up being for that check you got three years ago. The VCs will notice when they look at your books.

kstrauser 2y ago
That’s all true, but it works both ways. I’ve also been at a place that stuck to its plans rather than incorporating customer feedback and requests. You haven’t heard of it, and that’s why. I guess my main point is that you have to always be making the right thing. Sometimes that changes, and you have to be ready to pivot quickly.