Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I was a development manager for years, for a Japanese company.

The Japanese are type A++ (at least, in Tokyo). They are masters at applying stress across the Pacific.

As a manager, it was my job to insulate my team from the stress, and I often took the hit for telling my bosses that I wasn't going to push my people harder than they already were working.

It seemed to work out in the end. When they finally rolled up my team, I had been managing it for 25 years, and the person with the least seniority in my team had a decade with the company.

It's a whole different world, out there, now. I had a manager at a startup tell me that conventional management assumes that engineers will only stay at a company for 18 months, so they really pile on the stress.

I can't even imagine that. There were a lot of downsides to working with the corporation that I worked for, but they treated us all with a great deal of respect, and made it possible for me to keep valuable, senior-level C++ developers for decades; despite rather sub-optimal pay, and a not-so-thrilling work environment.



Game development has a degree of notoriety for its working conditions and I’ve noticed a similar management style to yours appear subtly in certain teams here and there. The turnover is much lower, although the work tends to be less glamorized. I think it works well for the older programmers that have started families.


How does the pay compare in game development? I have heard such mixed things


I'm sadly not particularly knowledgeable on US game development salaries. I can imagine it varies a lot by company and region.

I work in Vancouver, Canada, and a large portion of its industry is due to Canadians commanding lower wages. Combined with a weaker dollar, higher education availability, and single-payer healthcare, outsourcing to a Canadian studio can save a lot of money. FIFA, for example, is made in Vancouver and the franchise commands a large revenue stream. Provinces such as British Columbia and Quebec also provide tax incentives too, since it's high-skill work.


> I had a manager at a startup tell me that conventional management assumes that engineers will only stay at a company for 18 months, so they really pile on the stress

which is ironic, because isn’t that why people stay less and less longer at a company (can’t stand the slog)?


I think economically it may make sense to hire fresh people for less money, burn them out and then have them leave soon. This may be one reason for ageism .


That may be why there is so much emphasis placed on dev frameworks like React, because you can get JS programmers right out of BootCamp onboarded fairly quickly (and cheaply).

Also, there is a great deal of emphasis placed on process. If you will have new people taking over the code all the time, you should have a fairly strict requirement for patterns like MVVM and VIPER, and require new hires to be able to adapt to these quickly.

The Japanese had a similar structure, because they rotate engineers around every couple of years. The engineers stay with the company forever, but they change jobs frequently.

That means that it's very important to have a strict policy of heavily-documented, inflexible process, and you can have vast legacy codebases that people are afraid to change. If you can farm a lot of that out to a dependency framework, then it becomes S.E.P. (Somebody Else's Problem).

I'm not a huge fan of this approach, but many companies seem to feel it works.

There was a company called Taligent that took this to extremes. I read their style guide (I spent a great deal of time researching how other people did development -it made me quite the cynic). They had the strictest, most convoluted structure I've ever encountered.

Taligent eventually collapsed under its own weight. The idea was to only have Architects be creative, making line programmers little more than data entry. Programmers couldn't do stuff like create new classes without getting permission from an Architect.


Don’t forget that large companies internally are very similar to a communist planned economy. They have five year plans and central planning. Everything flows from the top. They value predictability over everything else. So it makes sense to run everything like an assembly line where every position knows what they are supposed to do and can easily be trained. Only a few elite people get to work on the big picture and be creative. Personally I don’t like this but I’ll admit that it works.


It does work, but is not a particularly good approach for consumer (or even pro) software. You keep delivering yesterday's technology, tomorrow.

It's the kind of approach that you might use for space launch software, but even that can be iterated (as SpaceX is showing).


As long as the money keeps coming in, I don't think the business guys will see any reason for change.


It's funny that if you invert the sentences, they even kinda slightly make some sense.

Small companies are similar to capitalistic free market. They have no plan, but everything flows from the engineering. They value innovation over anything else. So it makes sense to run everything like a lab where nobody knows what they are supposed to do but cannot be easily replaced. Only a few elite people get to work on their own pace without turning on their brain. Personally I like it but I'll admit that it never works. As a result, they always strive to deliver tomorrow's technology yesterday.


The people I've seen add the most value and be the most productive are always the ones with the longest tenure. Because they know the software backwards and they know the domain too.

18 months is barely a blip on the radar.


Absolutely agree. At my job, 2 years is still a baby, many people are at 5, 10, or even 15 years; and the company is only about 25 years old!


Absolute bullshit, but the metrics being gathered are all askew. When you've already fucked your productivity with bad management, what difference does a high turnover rate have?


> I think economically it may make sense to hire fresh people for less money, burn them out and then have them leave soon.

If the job is focused on smashing rocks with a hammer to make smaller rocks, that's probably true.


Provided it's easy to get new people up to speed and experience isn't overly important


Long term, no. Short term maybe, or it's not "economically" but "politically" desirable.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: