@twitter-356555426:disqus LOVE IT! Be the compiler! 

@14f6f3ac6b705ab4883313f779ec01ef:disqus Mentioned it in the first episode actually

I’ll do you one better https://github.com/dominickm

“Mike’s website” and “Chris on Twitter” are both broken links.

Subscribed yesterday for codeschool.com (with the affiliate link, of course!) and just finished up jQuery First Flight.  It was very fun, and had a ton of good tips.  Great tool @twitter-285215642:disqus! Thanks! Next stop, freshen up my CSS!

Ok, so I noticed some stuff that was said in the chatroom (and actually by the hosts as well).
Time to bust some myths:
1. Java is slow
Actually it’s one of the faster popular languages out there today. Java is generally only slightly slower than unmanaged C/C++ apps. It was slow before the java HotSpot VM (until java 1.3 in 2000), but since then IT-compiling made it one of the faster managed code languages.
It is definitely faster than python because python is straight up interpreted (same as java before the HotSpot VM). Java is also generally faster than mono, which uses a similar approach to run code (a VM to run an intermediate compiled language on any architecture supported). The only thing java has is a higher memory footprint and a small startup delay, but these are small enough to ignore. Java also has reliable performance, not too much difference between the fastest and slowest programs benchmarked.
Also important is that the performance of an app is more dependent on the quality of the code than on the used language/framework. 
Here’s some benchmarks to support my case: 
Python 3 vs Java 7:  goo.gl/7bCEh
C# Mono vs Java 7:  goo.gl/bIAaO
All benchmarks in a boxplot:  goo.gl/gWrda

2. Java looks bad/doesn’t look native
This actually depends on the used framework. If you’re using unthemed swing as a GUI framework this statement is true, but it is not necessary to use swing. If you theme swing with the GTK look and feel, it’ll look native on some platforms, although the text will look weird and there’s no globalmenu support for ubuntu. You can use libraries that use a systems native widget toolkits though (Java-Gnome GTK, QT Jambi,  SWT).

I really hope these myths will stop spreading, maybe an idea for a future episode, busting myths about programming languages/frameworks?

About #1: Java is slow when it’s first started. It’s incredibly fast once it’s warmed up good, but it has a horrible start up time making it useless for things that only last short periods of time.

4-7 seconds is a huge time for start up time. Especially for consumers who want their programs to be as responsive as possible[1]. Plus I wouldn’t want my consumers to _have_ to have JRE/Mono installed just to run an app that they’ll most likely have opened for 10-15 minutes.
PyPy has a JIT compiler (; And it’s much much faster than CPython!

Desktop apps shouldn’t have to be written in a huge heavy[2] weight language(Java, Mono, etc.[3]). You just need a language(compiled or interpreted(JIT is interpreted)) that’s reasonably fast and doesn’t take forever to start up and then a good GUI library(GTK, QT, etc.).

My conclusion of good languages to use for desktop apps, are Python, Objective-C, Perl and maybe NodeJS when a good GUI library comes out.

NodeJS is really good for CLI apps because it’s provides Unix style functions so it feels familiar to writing Bash, and it’s incredibly fast if you know how to use it correctly.

1: Most consumers have a really low attention span for waiting for something to load/start. Extremely low, 1-3 seconds tops. 
2: By heavy I mean so many built in features. Java has it’s 3D libraries built in it’s STL, which is a bit much IMO.
3: C/C++ Isn’t included IMO because you choose what to include plus it’s compiled so runtime is basically barebones.

PS: These are of course my opinions, so don’t feel like I’m bashing other languages or putting you down or anything. 

Well, there is a small difference in JIT-compiled and interpreted languages. The JIT-compiled languages compile the most often used code when the program is opened while an interpreted language compiles every bit of code when it is run. The advantage of the JIT method is better performance (since research indicated most programs run the same bits of code 60-70% of the time) due to not having to compile these parts again, but because it has to compile it when you start the program it will have a slow start up. If pypy has a JIT compiler it probably has better performance than plain python, but it probably has more startup lag as well. To the point of them needing a JRE/Mono runtime installed, most people already have JRE/.NET installed, and if not you can package a JRE with it (I’m not sure how this works with mono/.net, but since .net is a part of windows most people have it preinstalled).  Also, python/gtk on windows has been a real pain to run, and certainly to find an easy way to deploy it through installers. The only thing I got to work after about half an hour was an easy database program that just posted messages to the database and could read them later on.

I know you’re not bashing any languages, but I’m just trying to show that java has some advantages other languages don’t have. My favorites: easy to keep clean code (python does the opposite for me, easy coding, but messy to read later on), performance close to C, C++ and not too hard to learn. I don’t see the slower startup as an issue for apps that need the extra performance java gives you, just because it can be up to 60x faster than python so the couple seconds boot time are nothing compared to that. If you are writing an app that does some note taking or copies your pictures from an SD to a library on your hard drive you should use something else since it doesn’t need performance and the slow start up is a clear disadvantage.