Hacker News Reader

Comment by jerf | Hacker News Reader

You're bringing theory to a shootout based on facts. Hypothesizing that Java might be faster with a strong enough headwind doesn't make the actual, factual JGit run faster or C-git run slower.

I mention this because programmers seem prone to this, and letting theory trump fact is a great way to make sure that you never learn anything.

View on HN

It's a pretty specific shootout, of course; the reasons brought up for JGit's relative slowness are not issues present in most applications, they're specific to certain realms of system. Unsigned number support is a major headache about Java when doing low-level work, in particular.

I think it's also important to note that the slowness is relative to an incredibly fast C implementation - taking twice the time sounds painful, but in reality the pauses caused by Git are barely perceptible as-is, and double of that is not likely to be very annoying.

I was countering the headline, which wasn't correct. Java isn't slower than C in general. It may be in specific cases, for specific programmers, etc but that's not what the headline stated.

The title takes one specific application in a particular niche, used in a certain way which a particular programmer can't get to run fast enough, and concludes that Java isn't fast enough. Faulty logic.

The argument about JGit vs C-git for me isn't something I feel you can learn anything from.

> Java isn't slower than C in general.

Really? In general? That statements seams worse than the original headline. Is Java faster than C for anything?

It completely depends. Both are silly generalizations.

Yes Java can be faster than C in some situations.

Can C ever be faster than assembly? Well, in theory no. But in practice, sometimes.

Depends if you're a master assembly programmer who hand optimizes everything.

"Yes Java can be faster than C in some situations."

Unless the person writing the C implementation is grossly incompetent, I would like to see even one example for that.

You can't see a situation where the JVM can optimize things, inline code, etc at runtime and beat a general C implementation? :/

I would say that any C programmer worth his salt (unlike me) knows how to profile the code under a meaningful workload and how to optimize it to be faster.

I remember I used the shortest possible integers and the register keyword in the late 80's far too many times to count.

The point of somewhat higher-level languages like Java (I refuse to say Java is a high-level language - it would be one in the late 80s, but not today) is not to make programs run faster, but to make it easier to make them run correctly.

A Java application will invariably use more much memory than C implementation of almost any non-trivial application. You can inline code all you want, but it'll never change that. The JVM isn't magical -- it's not going to solve all your performance problems. When you're allocating a million tiny little objects with all the necessary accounting you can't optimize that away.

BTW, if it's so obvious why haven't you still provided at least one non-trivial concrete example of Java faster than C?

Reply on news.ycombinator.com