Hacker News Reader

Comment by rexpop | Hacker News Reader

This is somewhat off-topic, but I am not going to look into this stuff because, frankly, there's no call for it. Two major problems with advanced anything in IT:

1) business goals basically never require sophisticated software (I am lamenting this; it's most decision-makers' lack of imagination or understanding of the possibilities) and, consequently,

2) engineers love introducing unneeded complexity, maybe to exercise the mastery that decision-makers stifle in us. (I'm lamenting this, too. It's irritating and wasteful; if we right-sized work to fully exploit our human resources, they wouldn't burn out or implement maladaptive coping strategies like overengineering.)

View on HN

I find this incredibly shortsighted and can sometimes hinder one's career. Case in point, Consistent Hashing was "advanced", yet now it is a cornerstone and common knowledge in building systems. Paxos was considered advanced and "never required", yet it became the building block when Google told the world how their Chubby works. Compressed-sensing was a research topic some years ago, yet now it is used in advanced image processing (and yeah, the industry does need "advanced" use cases). If you track SIGGRAPH, many of the so-called advanced theories are adopted in games and CGI systems. You thought engineers could produce photo realistic waves without understanding large-scale Navier-Stokes equations? You thought we could enjoy realistic clouds, fumes, fires in games without engineers studying particle systems? You thought that engineers do not need to study complex optimization to build airline booking systems? Even for the traditional "IT", you thought SAP didn't require advanced data structures and algorithms for their logistics systems? You thought IBM didn't need algorithms and maths to help their customers optimize their warehouse management systems or help the military track their lost weapons including tanks?

Seriously, the list can go on and on for days if not weeks. And a personal lesson: I started to learn about deep learning and large-scale machine learning back in 2010 and they started to pick up steams in the bay area. Yet I thought that level of knowledge was "too advanced" for a systems engineer, so I kept deferring diving deep into the area. Boy was I wrong. Hopefully a new generation of engineers would keep their curious eyes more open than me.

> you thought SAP didn't require advanced data structures and algorithms for their logistics systems? You thought IBM didn't need algorithms and maths

You've read a lot into my words that I didn't say, and you're responding awfully callously to lamentation. I am an engineer. I enjoy advancing my craft. I wish I could have stayed curious, instead of burning out running against business management boundaries.

I would certain never say "IBM didn't need algorithms," and it's disingenuous of you to accuse me of such strawman arguments. All I am saying is that there's a lamentably low threshold beyond which we've too few opportunities to implement sophisticated or novel technology.

My God, what does it benefit you to accuse me of eschewing all of "maths"? Absurd.

My apologies. I just wanted to give examples that those data structures will be useful for ordinary engineers like me. By "you" I meant anyone not you specifically. Certainly didn't mean to put words in your mouth. I guess this is really bad writing, for which I'm sorry.

Content like this is a pure gold for us self-taught developers.

But it's pure lead for your boss.

Reply on news.ycombinator.com