morning!

@Sarov I saw God this weekend... he's disappointing XD

morning all

Hey you :)

i wish i could say it was a good morning but i'm exhausted and i dont have my umbrella but i'm pretty convinced it is going to rain :D

ugh

thats a mildly concerning title...

@AlexRobinson It's funny, but off-topic.

If it is edited to be on-topic, the title can be changed to something more appropriate as well :)

i think it can definitely be edited to be on topic if there's a will

We'll see.

@Tinkeringbell Haha, I dunno, gets better in the 6th season imo.

Meh :P First episode of that one was... sort-of disappointing too :P

Two words: Family Dinner.

Blegh

Hahahaha.

...??? Why was this marked as spam?

new account answering an "old" question using a lot of financial terminology... probably suspected advertising?

I guess.

Hello

Sarov I see you edited my question on PM stackexchange

Would you like me to make some edit to make it more focused and answerable ?

I am not sure what's ideal given that there is existing answers

I made a small clarification anyway @Sarov

@ArthurHv Up to you. Consider that even Tiago's answer notes that "it's really broad", though. I myself wouldn't know how to canonically answer it in its current form.

Hmm. Actually, I may be able to take a stab at it, as a frame challenge.

Yea. I figured it was quite broad despite my attempts to frame it

It really reflects where I am though, I am really in front of a problem I wouldn't know how to tackle as a dev cto or pm

Because I have various indicators something is wrong, but no idea where to go

@ArthurHv My rule of thumb: If your question would end up closed without edits, it shouldn't have been answered and editing to improve it is okay, even if it invalidates the existing answer to an extent. Comments on the answer are good courtesy though :)

If the question itself is okay enough for the site, and the answers you have so far haven't addressed specifics, comment under the answers first. If the answerer clarifies, you can edit your post and it should still match the updated answer ;)

Other option is to ask a second question then, and link to your previous one and point out what part the existing answer doesn't address :)

But that's really hard to do

Ideally I'd say it should be split int two questions since you're asking two separate questions. But they're related so it's a grey area.

(The motivation aspect is also a separate, third question, imo. Though I'm not sure that one can be made narrow enough, aside from just answering 'autonomy, mastery, purpose'.)

So far based on what we could really measure I highlighted bug / feature ratio, as a way to indirectly measure our ability to get it right first try. And oh wow we have 40% bug over total ticket ratio for over a year :D I think this data have raised eyebrows to the new PM and this might be a metric we follow in the next months. One question, related but different from what I asked, could be methods to improve this ratio.

Do you have a QA?

Kinda. The """product owner""" does this.

That can be either acceptable or not, depending on the PO. I've had that both work and not work.

Code reviews?

The problem we have, in our product users input complex data files we have to process. The PO typically tests nominal cases and we crash on all the other ones.

Okay yeah you need a more competent QA then.

@Sarov We have that like i said in the question. They are not finding much. I don't really know

Well, the primary purpose of code reviews aren't, imo, to find bugs.

Then they work correctly I think

Fewer bugs is an indirect consequence of those.

But code quality and architecture quality is distinct

They should both be reviewed for, then.

We have kinda an architecture problem in my developer eyes, but that might be just opinion idk it's where it gets complicated

Why do you think so?

Hmm

I'll be detailed

We are pipe-lining the user files in a program that's inside docker for no real reason, and process the files by creating a temporary database and iterating on it through SQL queries. We have a lot of knowledge as a result in SQL, but my opinion is that our core representation of the data is wrong. We could use basic classes and instances and be half less code and twice faster.

But now we have the data represented on SQL, every feature has to plug in to it.

It's 100.000 lines of code relying on this architecture, mostly sql, mostly non-reusable tools

So you don't actually need data persistence?

No, the data is solely stored for the purpose of processing

it's geo data and we use postgis plugin

Then yeah, adding a round-trip to a DB seems a bit silly.

Seems it'd be non-trivial to fix that, though.

Though, depends on the structure. If you have a repository layer, not actually that bad.

Since all of the logic will be outside the repo and not care whether it's SQL or memory or flat files or fairy magic under the hood.

We doń't really. We have a conceptual model. but we access the db directly

Yeah my first suggestion would be to more more towards a DDD approach first, then.

Separate data persistence (repository layer) from business logic/domain (core layer).

(And also separate presentation layer)

The frontend and pipeline are separate services and even teams.

You have no idea how much i'd love to rewrite stuff. But the problem is also the following

Oof. Scrum (and I) say that's bad! I correct myself - my first suggestion is to merge the teams to make them cross-functional.

Our product is aiming at helping collectives to bring fiber to their population. Hence once the job is done, our tools are NLN. Hence, investment in pursuing technical greatness is considered a muda. Except when other countries get interested in our tools

NLN?

@Sarov I agree with this. The PO of the other team thinks otherwise. He's a friend but I believe he's choosing suboptically there

@Sarov No longer needed

@ArthurHv Oh, admittedly that does change things - temporary code doesn't have to be held to the same standard as permanent code - though some things are still overall beneficial. Is all trade-offs.

@ArthurHv Is he familiar with the Scrum Guide?

No, he's by the way a developer that got promoted. I'm not

I'm not sure he knows much about project management theory

...A developer got 'promoted' to a PO? Oof.

xD

I'm laughing irl as you uncover the mess lol

Really what you need is a Scrum Master.

That's the #1 thing you need to do - convince management to hire/train a SM, and empower him/her to enforce Scrum.

If you can accomplish that, things will get better.

If you can't... things might maybe possibly get a little better somehow eventually after you die.

haah

WELL we have work right. I don't know how I can convince the management of something like this. I mean tweak our processes sure, but pay someone to tell them they're badly managing it? That sounds a lot

They don't need to be told - that's the point, the SM needs to be empowered to make changes without direct approval for every change.

If everything is broken, what you need is someone with both the knowledge and authority to fix it.

I can't say I don't like the idea. But I hardly see our CTO handing the key of the building. He's the kind of person to tell you to use that lib he likes because it's more modern and shinier. We burned considerable money following his command and control. But he's learning lean, idk maybe changes in 2 years from now

If there is a leader we expected and isn't here for us, it really is him. I've read a bit of lean theory though, I intend to remind him they do say "orient and support"

Also, regarding the 'pay someone' part - that's why I said 'hire/**train**'. You could just become the SM, yourself.

I AM NOT WORTHY ;_;

Yet.

If we speak long term, then perhaps. I was thinking about lean which I also like and there is a similar role (i imagine ?) that's called chief engineer

although lean's chief engineer has more like product orientation

Not actually that hard to be a SM - assuming sufficient support.

I need to read more about scrum though. I'm not sure I like it, although some of the precepts such as merging teams working on a product make a lot of sense.

See my answer about Shu Ha Ri. One of Scrum's features is how ritualistic and precise parts of it are (though it does still leave much open-ended). This means that you'll almost never find someone in the Ri stage doing Scrum, it's perfect for the Shu stage.

Compare bereavement - we have so many bizarre, byzantine funeral rites and processes, because people who are overwhelmed want to be able to just be told what to do. So they can just follow directions instead of having to deal with everything.

It doesn't matter if a Planning Meeting is the best fit for you, just as it doesn't matter if it's 'correct' that you shouldn't wail too loudly for fear of waking up the guard dogs at the door of the dead who will then bite the souls walking through. You just do it so that you can adjust to the new situation. And only once you've adjusted do you begin to consider changing.

I get the idea. It doesn't hurt (much) to do something you don't understand, may that be unnecessary. So until you know where you go, going by the book is a time saver. I haven't been asked to solve things though and pushing changes in that context might require justifications I don't currently have.

So really, your question is 'How do I convince management that everything is on fire?'.

And that Scrum will save us

Well, I've addressed that second thing.

I mean I'm convinced it will work but convince the management to empower - well me - is going to be a journey on it's own. I did start to point the house is burning though

Yep. Just because it's simple doesn't mean it's easy.

Thank you for your time anyway. I already look at things with a new eye

No problem.

Honestly, the hardest part is gonna be the IPS portion - "How do I convince my grandfather the house is on fire when he doesn't want to admit he started it?"

cooking.stackexchange.com/q/117261 What does "That is our fond" mean?

@Sarov browned stuff at the bottom of a pan

I see. Never came across that term before.

In french "fond" means the bottom of something

It probably comes from there

Makes sense.

Took me an hour of messing with git, but I finally managed to turn 3 commits into 2 commits.

(Important because there are ~20 files with changes that actually need to be reviewed and >1000 files that were just moved to a different folder.)

@Sarov Leant it from binging with babish/BCU

I've watched a couple of those.