Hacker News new | ask | show | jobs
by anonymous324 2057 days ago
PostgreSQL isn't exciting. It's boring, with roots in the Berkeley Ingress project from the mid 80s. Which is why 10 years ago, so many HNers went with trendy NoSQL data stores instead, even though they lost data. Data loss is very exciting, and PostgreSQL is very boring. This preference for what's exciting and trendy over what's old and proven is a large reason why so much software today sucks.
2 comments
Someone1234 2057 days ago
That seems unfair.

NoSQL was trying out new ideas in data storage. It was exciting to try out new or re-imagined core concepts, and some of those young projects had teething issues. But several are still around and remain popular, but they're popular for certain niches they excel at (and those niches were largely discovered through trial and error).

In the SQL-sphere a lot of people skipped Postgres because MySQL had, at the time, the momentum in the free/cheap relational database space. Between then and now Progres has grown more elegantly than MySQL, and people are rightfully looking to it.

link
anonymous326 2057 days ago
> NoSQL was trying out new ideas in data storage

Was it? Or was it just rehashing old ideas made obsolete by relational databases?

> https://en.wikipedia.org/wiki/IBM_Information_Management_Sys...

> Initial release 1966; 54 years ago

> https://en.wikipedia.org/wiki/Category:NoSQL

link
treeman79 2057 days ago
MySQL was also a lot faster in the early days.

Plus it “just worked” not sanely mind you. But you had a lot less problems with type mismatches, up/downcast, etc. Well unless you wanted quality data. But back then we were just happy that it didn’t bother us with “minor” details.

link
hoistbypetard 2057 days ago
It was a lot faster, and sometimes the odds that your data might live long enough for those integrity issues to matter seemed very remote.

Around 2001, the conventional wisdom was that it was worth the risk to take the performance victory, keep your hosting bill down, use mysql, and build some other approach to data integrity. (If memory serves, the integrity downsides were bigger then, too...)

The math is pretty different now. I'm not sure it's fair to hang all of it on Oracle, but it feels like that acquisition derailed some MySQL progress that might've made it a close call today. Now, I use Postgres unless I have a very specific, very compelling reason not to.

link
gvenzl 2055 days ago
Boring until it’s not: https://about.gitlab.com/blog/2017/02/10/postmortem-of-datab...
link