Client Runs on Waterfall

[pxue]

I'm secretly loving it.

I got hired to migrate a client's existing excel spreadsheet internal tool over to a customized React SPA.

This is the first contract where I'm not rushing around every sprint trying to piece together half baked features and pushing them out the door. While not strictly waterfall (more kanban) I'm enjoying the heck out of the process either way:

• Everything is rigorously tested and documented.

• Nothing gets released until all the requirements are met. No sprints.

• We celebrate every release.

• Clients give feedback, we spend time talking about it internally, and then do proposal, design and then developers come up with architecture docs and we talk about it some more.

As a 34 year old dev I'm loving this.

Am I just getting old?

[95014_refugee]

At 34, not old. Maturing, perhaps.

It’s worth reading the original Waterfall paper. If you’ve only ever heard of it as the bogeyman, it might surprise you.

From your tone here, I think you are seeing “people over process” play out in your context. Celebrations? Talking? More talking? Time to do work? These all smell like a sensible organisation doing sensible things. Congratulations, you lucky &^%*&^%&^.

[Jtsummers]

The original paper was very obviously not advocating for the initial process it describes. But it does describe Waterfall, and then it amends it into something that could actually work (specifically adding in feedback loops that Waterfall does not include).

Some fools in industry, DOD in particular, looked at the pretty picture and codified it as an option among several for defined development processes. Then contractors and gov't managers picked it because it was straightforward (on paper) to plan and schedule, and it made it into practice. It's still used by DOD contractors today and it still fails. It's not a bogeyman, but Royce himself is not to blame, he was pointing out the stupidity and illiterate fools missed the point.

[bradley13]

Sounds like a client with a clue. Congrats on winning the lottery.

Agile development certainly has its place. Sadly, in most projects, it has become an excuse for clients who don't know what they want, combined with managers who are unwilling to force clarity in the requirements.

[throwaway38375]

> Sadly, in most projects, it has become an excuse for clients who don't know what they want

Thank you for summarising why I hate agile so concisely.

I don't hate agile itself, more the clueless people who think agile replaces requirement gathering and planning.

[croo]

I agreecand I would add that waterfall too has its place as you OP luckily experienced. It forces you to design and plan the whole thing before execution and it also works better on projects that have a rigid unmoveable deadline.

[malfist]

Clients who don't know what they want, but they want it yesterday

[philomath_mn]

A lot of companies heard that big design up front (BDUF) was bad so they decided to do no design or planning up front.

You can't anticipate everything like true waterfall expects, but you also can't just yolo into a sprint without some effort spent on specs / designs / architecture / etc.

Sounds like your client found a balanced approach.

[GianFabien]

Enjoy! Clients like that are the exception.

I'm assuming that the requirements are documented and not altered every other day. Have to wonder whether your client is an engineering (not software) or related firm where they do detailed design before building.

[2rsf]

You are not getting old, you simply met reality. What you are describing is a functioning, mature and high performing company vs the opposite. It has nothing directly related to Agile vs Waterfall.

[runjake]

For others as lost as I am:

https://en.wikipedia.org/wiki/Waterfall_model

https://changelog.com/posts/waterfall-doesnt-mean-what-you-t...

(Links to the 1983 paper within both links)

[badpun]

Waterfall is good for developers, as you get unambiguous requirements that you just need to translate to code. Job is nice and easy.

However, waterfall can be disasterous for the organization. If a single big iteration takes a year, the entire team might be building something worthless for an entire year without getting feedback that what they're doing is not what the users want. I've been on such projects.