4 ms·

Knowledge in the software industry tends to get outdated much faster than in other industries. While there are timeless software engineering practices that are

Knowledge in the software industry tends to get outdated much faster than in other industries. While there are timeless software engineering practices that are good and evergreen, about 70% of my knowledge today will be outdated in 5 years. This is not true of other industries. Sure, other professionals need to keep themselves updated as well. But the percentage of knowledge that a bridge architect needs to update is much lesser (say 20%?). The cost of failure is much higher. It is easier to correct shitty software. So the focus is on delivering fast.

busterarm 9y ago
> about 70% of my knowledge today will be outdated in 5 years I really hope this isn't the case for you, and in general I don't really believe it to be true either. Most of what makes me successful as a developer today I knew 20 years ago, in a prior career. You have to differentiate between tacit knowledge and explicit knowledge. The former changes very little and the latter changes a lot. You're also best off retaining almost _none_ of the latter. Those "Teach Yourself <language/framework> in <number> <days/weeks>" books you used to see were written for serious professionals, not newbies. I can often get more done with less using tried, tested tools that change little than my peers can do with 'the new hotness'. It seems every language/stack has its own implementation of cron. Stop that. Use cron. Networking protocol, binutils and SQL don't really change and that's literally 99% of the "make stuff happen" part of my job. A lot of "Full Stack" developers don't know any of this shit. They don't know what tools will tell them if a client's traffic is coming into their network at all or being rejected at their server. Con artists, honestly.
- tracker1 9y ago
  Not to be offensive, but the front end is quite often far more work than the SQL that goes on in the back. I've worked strictly on the back, the front, and in between. I've also had to work in a number of platforms and languages. In the end, keeping things as simple as possible is usually best. If that means using language features that someone might not be familiar with, it should still be more understandable in the end.
  - busterarm 9y ago
    Not offensive at all. I actually would say in the best case scenario, you can't do one without the other. I will agree that frontend is full of churn. I have to do it, but I wouldn't specialize in front end -- most of your role is relying on explicit knowledge rather than tacit knowledge.
humanrebar 9y ago
> ...about 70% of my knowledge today will be outdated in 5 years... I've been around for a while. Some knowledge has become useless, but I learned how to sniff out technology that's going nowhere and avoid it. Since then, I've rarely wasted my learning on things that become obsolete. Besides, the most valuable knowledge is typically how product managers, salespeople, and people that set budgets think. That knowledge evolves much slower than javascript frameworks.
outworlder 9y ago
The kind of knowledge that gets obsolete is replaced by stuff that's very similar. Doesn't matter if it is a new language, framework or whatnot. There have been very few advances in our field, stuff changes incrementally. So you have seen everything before, it just has a different twist. You can pick this crap up in a fraction of time, as the concepts are already internalized. ... Unless, that is, if you used to call yourself "X developer", for whatever value of X. Then you are unidimensional, and screwed.

busterarm 9y ago

> about 70% of my knowledge today will be outdated in 5 years I really hope this isn't the case for you, and in general I don't really believe it to be true either. Most of what makes me successful as a developer today I knew 20 years ago, in a prior career. You have to differentiate between tacit knowledge and explicit knowledge. The former changes very little and the latter changes a lot. You're also best off retaining almost _none_ of the latter. Those "Teach Yourself <language/framework> in <number> <days/weeks>" books you used to see were written for serious professionals, not newbies. I can often get more done with less using tried, tested tools that change little than my peers can do with 'the new hotness'. It seems every language/stack has its own implementation of cron. Stop that. Use cron. Networking protocol, binutils and SQL don't really change and that's literally 99% of the "make stuff happen" part of my job. A lot of "Full Stack" developers don't know any of this shit. They don't know what tools will tell them if a client's traffic is coming into their network at all or being rejected at their server. Con artists, honestly.

tracker1 9y ago
Not to be offensive, but the front end is quite often far more work than the SQL that goes on in the back. I've worked strictly on the back, the front, and in between. I've also had to work in a number of platforms and languages. In the end, keeping things as simple as possible is usually best. If that means using language features that someone might not be familiar with, it should still be more understandable in the end.
- busterarm 9y ago
  Not offensive at all. I actually would say in the best case scenario, you can't do one without the other. I will agree that frontend is full of churn. I have to do it, but I wouldn't specialize in front end -- most of your role is relying on explicit knowledge rather than tacit knowledge.

humanrebar 9y ago

> ...about 70% of my knowledge today will be outdated in 5 years... I've been around for a while. Some knowledge has become useless, but I learned how to sniff out technology that's going nowhere and avoid it. Since then, I've rarely wasted my learning on things that become obsolete. Besides, the most valuable knowledge is typically how product managers, salespeople, and people that set budgets think. That knowledge evolves much slower than javascript frameworks.

outworlder 9y ago

The kind of knowledge that gets obsolete is replaced by stuff that's very similar. Doesn't matter if it is a new language, framework or whatnot. There have been very few advances in our field, stuff changes incrementally. So you have seen everything before, it just has a different twist. You can pick this crap up in a fraction of time, as the concepts are already internalized. ... Unless, that is, if you used to call yourself "X developer", for whatever value of X. Then you are unidimensional, and screwed.