Hacker News Reader

Comment by xorcist | Hacker News Reader

> Support shiny thing -> it's not difficult, everyone can do it

Sometimes people say those things, but I can't understand for whom it rings true.

What's especially hard about new things? Maintenance on the other hand, can be much more tricky. Especially when taking backwards compatibility in mind you really have to think about every step going forward.

It's much easier to build a brand new highway than repairing an existing one which is in use without disturbing traffic, if the car analogy can be forgiven.

I think the situation here has more to do with Google culture where the are too scared to stop throwing stuff at the wall and see what sticks.

View on HN

You're right, support is difficult. But there's no "glory" in support. there's glory in making something new, everyone being amazed, world-is-changed-forever.

Of course that is pretty much 1% of the journey. Turning that light-bulb moment into an actual product, getting mass market adoption, finding the killer app, that's hard. And anonymous. Keeping customers happy is hard. But again, anonymous.

Are you more impressed with a CV that says "invented iphone" or one that says "worked on iphone 2 to make it better than iphone 1"? Or worse yet "accumulated customer feedback to improve iphone 1, worked with team developing iphone 2"?

So it's not just google (although magnified there.) It's everywhere. You remember who _invented_ the light-bulb, but you've no idea of the people who developed that into a product, with manufacturing, distribution, sales and developed bulb-connector-standards and so on.

In our business it's the same. Here's a new wizz-bang feature. 100% glory. Maintaining feature for the next 20 years, adapting to ever-changing world requirements, keeping customers happy, and using it - anyone can do that. (and nobody wants to - it's too much like hard work.)

Mass adaptation is not even that hard, profitable mass adaptation is.

> But there's no "glory" in support. there's glory in making something new, everyone being amazed, world-is-changed-forever.

Glory is where you place it. And Google, like most of those Companies nowadays, are putting glory on the simple path of new creations, not the hard path of everything else.

> Are you more impressed with a CV that says "invented iphone" or one that says "worked on iphone 2 to make it better than iphone 1"?

Neither. If you have a speck of competence, you will evaluate actual skill, not the outcome of a whole team's work, of which you were just one screw.

But if you know history, or care for your companies future, I would be more interested in the second, because the first iPhone wasn't really that impressive or groundbreaking. The later iterations had more substance for the company.

Taking this opporuntity to link to Charles Proteus Steinmetz[1]. Among the many, many things Steinmetz did, he invented the Metal Halide Lamp which was later refined for mass production decades later. This took the innovation of others less well-known to history[2]. So great example, bringing up the light bulb!

[1] https://en.wikipedia.org/wiki/Charles_Proteus_Steinmetz

[2] https://americanhistory.si.edu/lighting/20thcent/invent20.ht...

> Are you more impressed with a CV that says "invented iphone" or one that says "worked on iphone 2 to make it better than iphone 1"? Or worse yet "accumulated customer feedback to improve iphone 1, worked with team developing iphone 2"?

None of those are impressive in isolation, honestly. What's impressive isn't what product you worked on, but how well you worked on the team, how well you technically performed, and whether the skills you were using are relevant to the position you're applying for.

> What's especially hard about new things?

Because it doesn't exist yet, you need to investigate previous approaches/research, come up with architecture/design, evaluate pros/cons, hardware requirements, staffing and so on.

When things become clear, it is easy to build it.

Support is harder in the sense that it takes a lot of time and digging into the platform to understand how can you solve it, but for a small gain (0.5% users will not see this issue again, unless you introduced new bug). Imagine solving 100 different support tickets, you need to solve at least 10 of them to start seeing the pattern, if there is any, to come up with generic solution. But at the time you solve 100 tickets, someone with experience can build solution which can make your product obsolete.

Support vs new has different trade-offs.

- Support -> takes more time and persistence, very small incremental improvements (88% -> 88.4%,..). But happy customers

- New thing -> takes more time on how to approach the problem, after problem/solution is clear, much faster to build for 85% cases

I think there's a personality aspect to it as well.

I've noticed that some programmers live for the front-end of a project. When you have a clean slate, are doing design, developing the data structures, approaches, all of that creative stuff. By the time the project is mature and the big problems have been resolved, they tire of it and want to move on to another new thing.

But there are others who hate that front-end stuff. They love taking a mature product and maintaining it. Debugging, adding new features, etc.

A successful product absolutely requires both kinds of people. A company who devalues one or the other is a company that is shooting itself in the foot.

Maybe support isn't a great word here. Maintaining software goes way beyond just supporting someone.

Staffing and hardware requirements aren't terribly important for greenfield development without any users yet. New products can be brittle and constantly in flux and that's probably ok.

Actually scaling the product once it exists and have users, building out new features without disturbing existing workflows, that's the hard work.

Which do you think is harder, inventing twitter as the fail whale product, or going from that to a hundred million users without disturbing the customers you got? I can't imagine the difficulties can even compare.

> It's much easier to build a brand new highway

In software world, new highway is not a new thing, it is just a copy/paste. In software new thing is mostly relates to inventing a solution for specific use case, solution itself could be invented long ago, but using in your area could be novel.

Even in construction industry, building new highway is not 100% new thing, processes, structures, materials, rules are already defined. You spend time for different type of thinking (where to build, which road to buy,...)

Reply on news.ycombinator.com