What did MySQL have during the last decade that made it better for certain applications than Postgres?
You didn't have to deal with schema/tablespaces, or the chance that someone used stored procs, or security (almost everyone just used the equivalent of a root account).
MySQL just worked.
For the most part, with zero monitoring, it stayed working until the server ran out of disk space or died.
MySQL was always available.
No matter which hosting company you looked at, no matter how little you wanted to spend, MySQL was pre-installed and ready for use.
I never thought it was the better choice, but I understood why it dominated. Once things like Wordpress, vBulletin, phpBB and other stuff was written only for MySQL the choice was then made for you.
I've long been a fan of applications that give you a choice of DB, but it's now unrealistic as many applications have designed schemas to cope with the quirks and limitations of MySQL specifically. These would now be harder to migrate than simply changing a database string, and most application devs don't invest in such things when what they have works well enough and is all anyone is asking for.
That it was faster (for certain applications), easier and more reliable does make sense. I understand, for example why PHP got popular even though there were other options available at the time that I believe were better from a software engineering[0] perspective. Things get popular by being better at the things the majority of users care about, even if they're awful at everything else.
[0] Did I really just use that term? I can't think of a better one to describe the issue I'm talking about.
One thing people seem to forget is that Postgres has not been around longer than MySQL, both were released to the public in 1995. Postgres skipped a few version numbers so maybe that makes it seem older.
But you are right, many of the other popular projects did not make this effort.
MySQL made it easier initially by omitting lots of difficult stuff (data integrity, management, transactions) thus making easier to get something working quickly. It also had a reputation for being faster then PostgreSQL, a reputation spread by fanboys (in truth people who were bigging up the only database they had a faint comprehension of).
Then it became ubiquitious. Now, like some nasty rash, we can't get rid of it.
It was indeed much faster when reading (probably still is on MyISAM) due to the lack of everything (especially transactions and any concept of data integrity).
Remind anyone of NoSQL? :)
It also has a single feature that I find incredibly cool: it supports automatically generated row IDs. It's been a while since I last tried Oracle and Postgres, but at the time you had to create a second table just for a sequence number and it was extremely awkward to handle.
Could the auto-increment feature in MySQL be better? Sure, for example it would be nice to have an option where the row is identified by a hash value instead of a continuous sequence. But even in its limited form today it beats having to wrestle with the glued-together equivalent as implemented in "real databases".
Edit: I'm not saying there are, but it feels like there might be some repercussions when you take the underlying implementation into account. It would be interesting to know if someone conducted comparative tests about this.
Must have been a while indeed. Postgres's SERIAL datatype has been available at least since 6.4. Postgres 6.4 was released in 1998.
There are other factors - you could limit its use of disk space (so hosts liked it), but mostly it was a good key-value store with some database functionality.
People who wanted a "real" database used Postgres. People who wanted to write a web app used MySQL.
Then a little bit of support from ISPs and corporations and it's a no-brainer for most folks. All the oss CMS and blogs and everything support MySQL really really well. ALl the ISPs pretty much use MySQL as the "database" solution.
From there forward, any time I was looking into a new host, they ALL had both PHP and MySQL. Any job I was bidding, discussion I was having in IRC or on Forums, anywhere - the two came as a matching set. I can't say that's WHY it became more popular, but I've always assumed their popularity were closely aligned.
Of course, now they've converged a lot, with Postgresql getting significant speed upgrades, and Mysql having InnoDB, with a few 'minor details' like transactions.