Comment by ksec | Hacker News Reader

Comment by ksec | Hacker News Reader

Parent Full thread

This is going to be ranty. I felt this is the same question over and over again. And Tech is worst than Fast Fashion Industry.

C or Death? Lets replace C with Rust.

SQL or Death? Lets replace SQL with NoSQL and now back to SQL and to something else?

HTML or Death? Lets replace everything with JS client side rendering.

PDF? JPEG? AAC? H.264 / AVC? x86 / ARM [1]? Every single thing that is 10 - 20 years old are considered old fashioned and we need to rip them apart without learning about their merits and pitfalls they went through. And sometimes as Alan Kay said only to reinventing the flat tire.

How about Better C? Better SQL? Better JPEG? etc And most if not all the time it isn't the language or spec that is the problem. It is the implementation. I am willing to bet we wont even have an Open Source SQL implementation that is as good as today's Oracle or MS_SQL in 10 years time. There are still insane amount of work on implementation alone.

It seems the Tech industry as a whole is in an eternal philosophy battle of throwing away everything and build something new vs improving what we have today and iterating on it.

[1] A prominent Intel engineers once said by the time you learn everything about Hardware design pitfalls from GPU to CPU and different ISA and be ready to design something truly so much better you will be at the age of retirement. I think this sort of hold true in other areas as well.

Everything you're complaining about being replaced not only still exists but is still widely used and in many cases continues to be updated and improved, often in response to the ideas introduced by these newer technologies. Would relational databases have JSON support without NoSQL’s rise in popularity? I doubt it.

This just isn’t a problem. There are more than enough programmers in the world to explore different approaches to common problems. We all ultimately learn and benefit from those experiments. The cycle of reinvention isn’t some great failure that requires fixing. It's how progress works.

And yes, people will not understand everything about what exists or has existed before trying something new that may or may not work out. We're humans, not robots. We don't have perfect memory and we don't have the bandwidth to ingest every bit of wisdom that has come before us, so inevitably we will retread some of the same mistakes that have been made or rediscover things that have already been discovered. That's called learning.

> It seems the Tech industry as a whole is in an eternal philosophy battle of throwing away everything and build something new vs improving what we have today and iterating on it.

At what point do iterations become so different or just completely ideologically opposed that it stops being compatible with the language? Should we get rid of malloc and free in C? What if the extension I want to add to the language is rejected by the governing board? What if the file format is proprietary and requires a license? What if it's belongs to a company like Oracle or Microsoft and contributing to its development is only possible if you work for them?

The trends in tech come from a demand for better software solutions and innovation in making them. People want modern apps and developers want modern tools. People who want to use brand new technology just for the sake of it exist, but for the most part we're growing and learning and making great things.

We are absolutely learning from the pitfalls of our previous technology to make better things. It's not perfect but we're moving forward. It's not fast fashion, it's more like the car/vehicle industry.

Older developers who cannot tell the difference between a website built 20 years ago vs today or think the developer experience of working with rust vs c++ is just aesthetic are digging their heels in to their own ignorance

Whatever the industry comes up with, it needs time to mature. What’s in question is mature and still useful - we can learn a whole lot from it. So the only take here is to not jump on from bandwagons to bandwagons and stick to some principles to maturity.

One nitpick. We do have an Open Source SQL implementation as good as commercial ones: PostgreSQL.

Reply on news.ycombinator.com