Hacker News Reader

Comment by _skel | Hacker News Reader

Being culturally Japanese didn't prevent Japanese companies from achieving dominant positions in electronics, automobile manufacturing, video games, shipbuilding, steel production...

Why is software so special?

I would argue that the kind of risk aversion that holds back software companies is just as evident in most of Europe as it is in Japan. It is also the norm in a bunch of American companies in the tech industry, especially the older ones. And China, which is also often brought up as a society where individuality is not valued in business (due to "Confucianism" and similar tropes), is not held back by similar risk aversion and has a massive software industry that can compete internationally.

View on HN

Software is different from those other industries cause it's not really a science. No one's figured out a "right way" to make software, and it seems being well-made is a very small factor in the grand scheme of whether a piece of software becomes ubiquitous/successful. So throwing things at a wall and seeing what sticks works better than iterating and perfecting a technique.

Video games are a strange exception I suppose, because they are software. But it seems like above-described culture obsession with not making mistakes doesn't apply so much to the arts.

> Software is different from those other industries...

...because it is the intersection of art and engineering.

I think video games are the exception that proves the rule. They are software, yes, but due to their primary use being casual enjoyment (as opposed to productivity), being well-made is important for them -- video games have to run well, look pretty, have few bugs, etc. precisely because there's no other reason to play them other than to enjoy them. Meanwhile productivity software offers something else into addition to being merely well-made.

Have you played Japanese pc ports on steam? I… recommend against it. Most of the ones I own only work under specific conditions if at all. Several are literally unplayable, or require you to downgrade your video drivers to something from 3-5 years ago. The best of them often still drop frames, requiring fan patches to maintain 60fps. Sometimes you fall through floors. They are never patched

The primary quality of what people consider "well-made" for a video game is the artistic part. Sure, nerds will compare graphical fidelity, resolutions, framerates, etc. but most people just care about whether they're enjoying themselves.

Shadow of the Colossus infamously ran pretty badly on its original release platform (PS2). A good amount of the N64's library ran at sub-20fps in certain areas. Bloodborne launched with 60-second load screens. Hell, remember GTA5's multi-minute loading screens that people put up with until someone reverse engineered the code and found it was mostly pointless JSON parsing? Dark Souls was 30fps on PCs for years (without a mod that caused bugs of its own) until the remaster came out.

There's lots of examples of badly-running and/or ugly and/or buggy games that are still regarded as great.

There was a big video game market crash in the USA in the 80's before Nintendo arrived on the scene. The issue was precisely quality (see: E.T. the game)

Because move fast and break stuff doesn’t really work in non-software business I’d say.

But again, how is that a Japanese problem specifically, when the same risk aversion exists in corporate culture in other countries (as I mentioned)?

The reality is that only a few countries are dominant in software, and they dominate to such an extent that they draw in software engineers from the rest of the world. That didn't happen because all the other countries have a Japanese business culture... A convincing explanation of why the world is that way needs to be based on traits that are not unique to Japanese culture.

Fair point, I don’t really have oppinions on that.

My point was the inverse, that the archetypical Japanese approach to engineering suits the iterative careful development that hardware requires.

Right when the basic product can not change much it makes sense to improve it and the process to manufacture it gradually. Really Japanese quality movement was much about process quality. You produce the same thing but with a gradually improving process that brings in better quality and better efficiency of production.

But because software is "virtual" it can change-shape in whatever way as long as it does something useful. The software products change rapidly whereas we are still in the beginning of trying to understand what are the best processes for producing it, why because process depends a lot on the product. So if the product changes fast process must change too. And when it changes fast it is difficult to improve it gradually.

Move fast and break things gets you established in the marketplace early and can bring some advantages from that. I have yet to see evidence that it leads to companies that survive for the long haul. At some point, people want products that actually work and don't completely change with a given update.

You can still move fast and break things in development.

What allowed Japan to modernize in the 1980's was a huge full society effort (with a lot of government backing) to remove the hierarchism from their work culture. All the great management innovations they started at that time are of the form of bottom-up knowledge transfer and distributed control.

If the software development is done under their "office culture", instead of the "factory culture" that was changed, it would be a very good explanation for why it's not very innovative. But I have no idea if that happens.

Software is special, because in software the blueprint is the building, the process of design and development is one, or rather can be one, but that requires a cultural shift.

Reply on news.ycombinator.com