4 ms·

NoSQL rocketed in popularity, because it required zero working knowledge of how databases work. You could get up and running on a project, without having to wor

NoSQL rocketed in popularity, because it required zero working knowledge of how databases work. You could get up and running on a project, without having to worry about what tables, columns, and their relationships to one another meant. You could throw ANY data in, and generally get it back out.

sharadov 6y ago
Exactly! Developers did not want to learn messy databases. I think a lot of folks without experience entered the industry ( mostly from boot camps and such and ruined everything)
dodobirdlord 6y ago
I think that's not the reason. NoSQL rocketed in popularity on the back of adoption by a few large companies with scale problems that had to abandon relational databases due to scale issues. If the requirement is to serve very high low-latency throughput to back something like shopping on Amazon, then relational databases and SQL in particular aren't very helpful. You know your data access patterns up front, and can optimize your database to support exactly your API's access patterns. Ad hoc queries on the production database are prohibited, data analysis work gets done with some kind of ETL pipeline, and the choice to trade off any part of ACID for more throughput and lower latency is a no-brainer.
- overcast 6y ago
  A few large companies helped to get it onto the radar of a lot of inexperienced developers, who found how easy it was just to plug away on it. All of that performance nonsense was second fiddle by a long shot for the vast majority of users.
claytongulick 6y ago
I agree with you partially. I also think that it became popular because a lot of data for many cases is naturally semi-structured documents / complex graphs. Strictly modeling this in many cases is either impossible or prohibitively complex. Like most tech fads, NoSQL is best used as just another tool in the toolbox. It is entirely appropriate for a huge class of problems, and entirely inappropriate for others.
- f6v 6y ago
  I can almost guarantee that for most of the SaaS startups which still go with the React, Node, Mongo stack the data is structured. They have users, orders and whatnot. It just takes some experience to foresee the upcoming product changes. But as someone said here, the nosql stack has been incredibly popular among recent bootcamp graduates.
  - claytongulick 6y ago
    We're going to have to disagree there. Loosely structured graph data is a thing, and there are many uses cases for it. This is why pretty much every rdbms these days has a json datatype.

sharadov 6y ago

Exactly! Developers did not want to learn messy databases. I think a lot of folks without experience entered the industry ( mostly from boot camps and such and ruined everything)

dodobirdlord 6y ago

I think that's not the reason. NoSQL rocketed in popularity on the back of adoption by a few large companies with scale problems that had to abandon relational databases due to scale issues. If the requirement is to serve very high low-latency throughput to back something like shopping on Amazon, then relational databases and SQL in particular aren't very helpful. You know your data access patterns up front, and can optimize your database to support exactly your API's access patterns. Ad hoc queries on the production database are prohibited, data analysis work gets done with some kind of ETL pipeline, and the choice to trade off any part of ACID for more throughput and lower latency is a no-brainer.

overcast 6y ago
A few large companies helped to get it onto the radar of a lot of inexperienced developers, who found how easy it was just to plug away on it. All of that performance nonsense was second fiddle by a long shot for the vast majority of users.

claytongulick 6y ago

I agree with you partially. I also think that it became popular because a lot of data for many cases is naturally semi-structured documents / complex graphs. Strictly modeling this in many cases is either impossible or prohibitively complex. Like most tech fads, NoSQL is best used as just another tool in the toolbox. It is entirely appropriate for a huge class of problems, and entirely inappropriate for others.

f6v 6y ago
I can almost guarantee that for most of the SaaS startups which still go with the React, Node, Mongo stack the data is structured. They have users, orders and whatnot. It just takes some experience to foresee the upcoming product changes. But as someone said here, the nosql stack has been incredibly popular among recent bootcamp graduates.
- claytongulick 6y ago
  We're going to have to disagree there. Loosely structured graph data is a thing, and there are many uses cases for it. This is why pretty much every rdbms these days has a json datatype.