Because I want them to look pretty as .js

As much as I agree with your recommendations to use the PDO or MySQLi extensions over the old MySQL API, I would still consider it a good idea to learn it. Few PHP developers will be lucky enough to avoid it altogether, at least for the foreseeable future. There is just too much code out there that uses it.

With that said, it should be learned along side PDO or MySQLi, not by itself.

KingCuddles, on 20 September 2012 - 08:10 AM, said:

What are you basing that on? (Besides the claim in that post you linked to, which is not backed up by anything.)

From all I've seen in the last few years, all the benchmarks I've managed to find, and from what I've observed in my own code, the old MySQL extension is actually slightly (emphasis on "slightly") faster than both the MySQLi and PDO extensions. Keep in mind that all three extensions rely on the exact same client library - either MySQL's Client library or, in PHP 5.3+, PHP's own MySQL Native Driver (mysqlnd) - so the only real difference in the performance is in the implementation overhead, and how it's used. As the simplest implementation of all three, the old MySQL API wrapper would (theoretically, at least) incur the least overhead.

Also, the link you posted mentions that the need for a lot of string concatenation when using the old MySQL extension is another performance hit. Interestingly the default behavior for PDO's MySQL driver is to emulate prepared statements, meaning that it will actually construct the SQL string in much the same manner we did when using the old MySQL extension. It's just hidden behind the PDO API. (It's probably done in C, not PHP, but still. The performance difference is practically non-existent.) - Another interesting point is that this emulation tends to be faster than MySQL's own handling of prepared statements. (Although, again, with a trivially small margin.)


Don't get me wrong, though. I'm not suggesting anybody use the old MySQL extension because of this.