Hacker News Reader

Comment by systemvoltage | Hacker News Reader

It would be great if we stop adding features now. Are there any pieces of software that has started doing this? Why are we perpetually in this loop of feature creep?

Like a painter, you must know when to stop painting. This is a real dilemma all painters face, especially when dealing with watercolor. If anyone has tried painting here, you know exactly the problem - how do you know you're done? It applies to writing and music as well, but it is more evident in visual arts than any other field.

I wish there was a Postgres branch that took previous version and then just applied optimizations and bugfixes. No more.

View on HN

> It would be great if we stop adding features now... I wish there was a Postgres branch that took previous version and then just applied optimizations and bugfixes. No more.

To be fair, quoting from the article:

> There are no big new features in Postgres 13, but there are a lot of small but important incremental improvements. Let's take a look.

But also, in general, yes there are pieces of software that do this - most recently Moment.js[1]. There was some discussion earlier this week[2].

[1] https://momentjs.com/docs/#/-project-status/

[2] https://news.ycombinator.com/item?id=24477941

It's unfortunate that performance is so rarely considered to be a feature.

There's plenty performance work going on in postgres. Most major releases had significant improvements in some aspect of that.

Postgres usually adds plenty of performance from one version to another.

Not much on .0 versions, those are more feature packed, but all the other time.

Actually, Postgres does not add any performance features to point releases at all. Point releases are only for bug fixes. You may be thinking of MySQL.

Occasionally, there are cases that could be argued to be exceptions to the general rule. But that's a hard argument to make -- everything committed to a back branch is officially a bug fix. Things like optimizer regression fixes are "performance enhancements" in a certain sense, but are nevertheless justified as bug fixes.

“.0” versions are the only versions that would have a major performance change. Anything else would be a fix version.

It's not as if Postgres is compromising on quality to add features, they're still very good at keeping Postgres stable even at the impressive rate that new features are added.

And while there are cases where I might agree with the general sentiment, I strongly disagree with this for Postgres. The new features are important and useful. Postgres is not a very specialized tool, it is a general purpose database that is used in many different ways. It isn't just done and feature-complete, there's still a lot of potential to improve it for various use cases.

> Like a painter, you must know when to stop painting.

You presuppose that Postgres (or all software) is comparable to art, without making or explaining the comparison.

You say that "you must know when to stop painting", but what if the developers behind Postgres know when to stop, and it just isn't finished?

What makes you reason that Postgres is finished?

Seriously, any analogy would have been better than an artistic painting one. They could have said "like a goatherd, you need to know when to stop feeding your goats" and been closer to the mark.

Suppose they had done that 10 years ago, would we be better of? Don't you want to know wher Postgres will be with 20 more years of innovation?

I'm afraid if postgres stopped adding features now, in 10 years it would no longer be relevant in software development. Most developers would probably jump ship to a more innovative database.

The world changes, which means requirements change. What was once good enough, will be insufficient in the future if you don't update it.

Well, there are branches that just apply bug fixes. We have five years of support for our (yearly) releases.

Optimizations kind of are features (sometimes kinda bugs too). Why should those be singled out vs other features?

The primary "bottlenecks" using postgres are different for different people. For you it may be performance. For others it's easier administration. For others it's SQL level query capabilities. Etc.

Postgres 9.5 is still getting updates and bug fixes, each major version has 5 years of support.

If you want to move very slowly and deliberately, you can do, depending on your requirements.

https://www.postgresql.org/docs/9.5/release-9-5-23.html

How does the introduction of new features negatively impact your use of Postgres?

You need to RTFA. Also, database development is absolutely nothing like painting. "Postgres 13 features and enhancements :: Making a watercolor painting" is Top Ten Worst HN Analogies material.

Why? There's lots of features that Postgres doesn't have, but could, which would fit perfectly well into the current offering. Right off the top of my head, built-in bitemporal support recently added to MariaDB would be welcome in Postgres, and would have no negative impact on any existing functionality. "Stop adding features" is sensible when adding features degrades existing features or the experience thereof, it makes no sense at all when it has no negative consequence.

Some people need or benefit from things that don't interest you. If that's the case with a RDBMS, why not just... not upgrade?

If you go through the changelog, v13 as a release is mostly about improving existing features but I don't think pg is at that point yet where it could "stop" adding features. It almost stopped 8-10 years back and that is when mysql had overtaken it completely. I believe it still is the best open source SQL engine but features like native partitioning, JSON support etc that have been added in the last few years were absolutely needed in this century to be competitive. And now that they're there, they need to be improved and made stable.

There is this special biologist word we use for 'stable'. It is 'dead'.

-- Jack Cohen

Fork source and organize like minded developers (if any) around that concept.

Of course, that is a lot more work than posting comment...

I'd be game if the project was called "Postpostgres", but only for that reason.

If postgres stopped adding new features, we'd be forced to adopt Oracle

Good news! It’s an open source project so you are free to do this yourself!

Be the change you wish to see in the world.

Be the change you wish to find in the couch.

I feel like a million bucks!

[Insert default "its OSS, just create a fork" here]

Reply on news.ycombinator.com