Purpose of Java in Bioinformatics

[rskr]

How is someone supposed to debunk "cross-platform" or java is faster, to complete novices?

The problem with Java is that there is no way to port it. So there are features which are not cross platform especially GUI features, which are not part of the cross-platform contract with Microsoft(which let its contract expire). Since java is cross-platform the language has no porting constructs, so if you need a feature which isn't implemented exactly the same on more than one system(anti-aliasing is an example) there is no way to maintain the code effectively.

C and C++ are faster, by up to 20times in neutral language comparison bench marks. Besides that if you know how indirection works, you know java can't be very fast, unless you write everything in a giant static main with local non-object variables, which well no one really does that do they?

Anyway, untrained Java programmers appear to work for cheap, but the question is do they bring anything to the table. Cross platform has been debunked.

[NicoBxl]

and it's OOP compared to basic scripting languages

[damiankao]

It's easy to use and fast to develop. Bioinformaticians usually are not hardcore programmers.

[mudshark]

I use Java a lot, coded a lot of bioinformatics and NGS tools. I prefer it to C and scripting languages because of
1) easy memory management (overhead is not much of a problem)
2) easy development process (very good IDEs)
3) and it's considerably fast (if you e.g. re-code some R-workflows in Java you will reach speed of light..)
4) easy setup of threads/multithreading which is a major plus in NGS analyses
5) cross platform (Windows, LINUX, OSX without any problem)

[ETHANol]

This guy is clearly a troll.

From the perspective of a biologist, I've found some pretty useful tools written in Java so Java has clearly contributed a lot to bioinformatics. My thanks to the people that made them. Perhaps rskr is jealous he didn't write them. Maybe he should read this book:

http://www.amazon.com/Java-Bioinformatics-Biomedical-Applications-Harshawardhan/dp/0387372350

and get back to us with a full report.

[ECO]

rskr: Was this a serious question? Seems pretty clearly a troll, similar to your last thread on 36 base reads.

I love a good snarky post now and again, but trolling to insult java in general is non-productive.

[rskr]

Disbelief. I know lets everybody program in a super slow language, with poorly designed code(see the first post), then buy a cluster to run it. When it could have been done on a single box, but hey at least if you run your java bork so no one else can show you up with a real language, because you have installed hadoop on it, and well nothing uses hadoop.

[BetterPrimate]

Last I checked Java couldn't run under SGE.

Then it's about time you checked again.

[krobison]

In addition to the assembly viewers, there are also tools written to use Hadoop that are in Java. The SeqWiki has over 60 programs currently listed for Java covering a wide range of topics.

I haven't seen any Android bioinformatic apps yet, but that is also a Java world.

[rskr]

cross platform

Last I checked Java couldn't run under SGE. So no not really.

[maubp]

Most assembly viewers these days seem to be written in Java using Picard for SAM/BAM support.

As a bonus, they are cross platform and so can be used on Windows as well as Linux or Mac OS X. Unlike most Bioinformatics applications I would say, where Windows is poorly supported.

[rskr]

Does java contribute anything to Bioinformatics, other than a cheap complicit labor pool? I know in certain business applications the cheapness and complicity is the penultimate bottom line, but in bioinformatics, I don't see what java brings to the party.