Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Things that aren't doing the thing

strangestloop.io

Post reply on HN
Print view View original

211–220 of 220 posts

Re: Things that aren't doing the thing

#211

Post by jstummbillig »

mlyle wrote: post #189
jstummbillig wrote: post #161

Earlier quoted context omitted.

The ways in which you can do things around, but never actually producing a single paperclip are infinite.

That's true, but the number of paperclips you'll produce if you only devote effort to things that directly produce paperclips is also very low. It can be really difficult to strike that productive balance.

Somebody once said in light of this dilemma: Do things that don't scale.

This never meant: "It's fantastic if things don't scale". It means: "Trying to start by doing things that scale is probably bad". I mostly believe that.

Re: Things that aren't doing the thing

#212

Post by zkmon »

You know what? There is actually nothing to be called as "doing the thing". It is all layers all the way down. You keep peeling the layers, and there is nothing left.

When you do coding, you are only writing instructions for a computer about how to do something. So it is not actually doing something. When a computer runs your code, it is only telling some hardware to pass current through some circuits. Again it is not doing something. Current passing through circuits is not physically moving anything through the wires or metal. So again it is doing nothing.

Re: Things that aren't doing the thing

#213

Post by p_v_doom »

saulpw wrote: post #31

In the Viable System Model[0], "doing the thing" is System 1. Yes of course you need System 1 or the thing won't get done. But in any viable system, you also have the "meta-systems", Systems 2-5: - System 2: coordination between multiple Systems 1 (which includes prioritization, communication, and exceptional conditions) - System 3: resource allocation and process development - System 4: strategy and risk management…

Never thought Id see Beer references on HN

Re: Things that aren't doing the thing

#214

Post by jimbokun »

lukan wrote: post #146
jimbokun wrote: post #84

Earlier quoted context omitted.

It does add up though. 1. Ticket 2. Feature branch 3. Write unit test 4. Periodically merge from main. 5. Implement to get tests to pass 6. Push changes 7. Create PR 8. Wait for PR feedback 9. Address feedback and repeat 10. Close and merge PR 11. Automated CI deploy 12. Integration testing as needed 13. Close ticket 14. Include in next prod release All of these are good things. But the overhead is significant. And t…

That is all part of doing the thing called quality software develeopment. Quick and dirty does not work in all environments.

But Quick and Dirty works very well in other environments.

A lot of it depends on whether you have a lot of customers and revenue you are putting at risk by making quick and dirty changes (mandates more process) or coming up with an MVP (quick and dirty might be beneficial if you can iterate faster).

Re: Things that aren't doing the thing

#215

Post by verall »

taneq wrote: post #210
verall wrote: post #206

Earlier quoted context omitted.

Yes in all the environments I've been in this is nearly the entire job. But I work in embedded... Need to verify feature X - okay, let me just - - Find the list of CPs needed - Get a recent source tree - Acquire a board - Flash board - Get CPs - Manually assign IP and bring up SSHD to the device so I'm not working over serial - Build and SCP the rebuilt thing to the device - Find whatever byzantine command you need t…

> - Oh did you remember to flub the flabnabber? echo "1" > /dev/i2c5 then restart the flabnabber-daemon Ah dammit I always forget the flabnabber. 10/10 authentic embedded development experience.

> $ echo "1" > /dev/i2c5

Permission denied.

> $ sudo bash -c 'echo "1" > /dev/i2c5'

I forget every time

Re: Things that aren't doing the thing

#216

Post by saulpw »

jstummbillig wrote: post #211
mlyle wrote: post #189

Earlier quoted context omitted.

That's true, but the number of paperclips you'll produce if you only devote effort to things that directly produce paperclips is also very low. It can be really difficult to strike that productive balance.

Somebody once said in light of this dilemma: Do things that don't scale. This never meant: "It's fantastic if things don't scale". It means: "Trying to start by doing things that scale is probably bad". I mostly believe that.

I also believe that in order to bootstrap a viable system, start with system 1. The OP is partially sniping at people who creating all kinds of elaborate structures for systems 2-5, without attending to the first order of business, which is system 1. I've been guilty of this myself.

But it's easy to find dysfunctional organizations that focus almost exclusively on System 1, and it looks like grinding, hustle, "doing the thing", being stuck at CMM Level 1, and in my experience, it leads to burnout. So there has to be a balance.

Re: Things that aren't doing the thing

#217

Post by jibal »

varjag wrote: post #185
jibal wrote: post #131

I was just reading about Ron Jeffries' disastrous attempts to write a Sudoku program because he dogmatically refused to do any preparatory work--including spending any time at all thinking about the nature of the problem and what sorts of data structures would be suitable for solving it.

As I see it he was doing exactly the opposite: prep work in terms of doing the tests.

That's obviously not "prep work", and in any case obviously not the sort I'm obviously referring to, so obviously not "exactly the opposite".

Re: Things that aren't doing the thing

#218

Post by varjag »

jibal wrote: post #217
varjag wrote: post #185

Earlier quoted context omitted.

As I see it he was doing exactly the opposite: prep work in terms of doing the tests.

That's obviously not "prep work", and in any case obviously not the sort I'm obviously referring to, so obviously not "exactly the opposite".

He was coding everything around: checking the validity of moves, auxiliary code, win conditions and on. Anything but the algorithm itself, likely for the lack of experience. Doing the thing here would have been figuring out the algorithm and sitting down to code it which probably felt intimidating. So yeah it was exactly the opposite.

Re: Things that aren't doing the thing

#219

Post by jibal »

jibal wrote: post #217
varjag wrote: post #185

Earlier quoted context omitted.

As I see it he was doing exactly the opposite: prep work in terms of doing the tests.

That's obviously not "prep work", and in any case obviously not the sort I'm obviously referring to, so obviously not "exactly the opposite".

Someone doesn't know what "opposite" means or pretends not to.

Re: Things that aren't doing the thing

#220

Post by jk-pasha »

Oversimplifying doing the thing isn’t doing the thing.

Sometimes it’s even harmful to doing the thing.

But sometimes it’s exactly the right thing to do.

Know when you need to oversimplify and when not and you’ll know Zen.

And when you know Zen you’ll do exactly the thing.

Post reply on HN