Java: IBM vs. Sun
Submitted by sticks on Fri, 07/23/2004 - 07:41
Tagged:
As I was trying to figure out how to use Java's ThreadLocal class, I did my usual routine: Goto the javadocs for the class. Then, when not finding enough information on usage, I google. As is usually the case, the most useful hits on google come back from developerworks.
As I was reading the document I had an epiphany. If it wasn't for IBM and it's work in both actual Java code and the documentation provided from places like developerworks, I wouldn't be able to use Java effectively at all. This led me to the next conclusion.
In my mind, IBM has made java actually usable, and is the only reason that it has not been swallowed and/or forgotten as an interesting exercise.
That's it. Just wanted to share and/or get comments.
Comments
RE: Java: IBM vs. Sun
I would agree that IBM has done a ton to help Java (and both developerWorks and alphaWorks are great resources for java, linux and open source) but I would not say its the only reason Java is viable and or usable.
I would agree 1000% that Suns documentation usually sucks and that their own marketing is even worse. IBM does a ton better job in both those areas concerning java and hence does deserve a lot of credit.
But I think there are other useful Java resources, and other companies using a great deal of Java (they just dont give back nearly as much as IBM).
RE: Java: IBM vs. Sun
See, personally I would give Apache/Exolab/Objectweb more credit for really making Java "usable" than even IBM. The language is great and all, but what gives Java the edge over VB/VC++ is the availability of all those free APIs for doing, well, all the stuff you need to do.
I will grant from a documentation standpoint, however, developerWorks is WAY better than the standard java.sun.com. However, I think Sun has realized this and this (part of the reason) is why they have the Java.net stuff now.
RE: Java: IBM vs. Sun
My argument against Apache is that the code that I use most often from Apache, came from IBM anyway. Besides the "official" things like SOAP, check out the authors of the various common utils, too. A ton of IBMers in there.
Not used the Exolab/Objectweb stuff, therefore (in my purposely narrow-minded view,) they can't be that important, and therefore don't contribute to the livelihood of Java.
RE: Java: IBM vs. Sun
Oh yeah, and Apache's documentation can be worse than Sun's a lot of the time.
RE: Java: IBM vs. Sun
So whats up with ThreadLocal, I havent used this. Why are you using it and whats the difference between it and just synchronized (is it just that the next thread doesnt wait, just gets its own copy?).
RE: Java: IBM vs. Sun
>>My argument against Apache is that the code that I use most often from Apache, came from IBM anyway
True on a lot of stuff, I guess. Frankly though, the whole commons-* project list is just the best thing ever.
I guess I have never really had any problems with documentation on Apache stuff either, barring the Incubator projects.
RE: Java: IBM vs. Sun
To be clear, I'm not dissing Apache at all, I love their stuff. I'm just saying that w/out IBM's contribution to it, it'd me much, much less than what it is today, and I'm wondering if Java would have even been "cool" enough for people to ever have spent time on the non-IBM stuff.
ThreadLocal: Handy little mechanism to make a variable global to a thread and only that thread. The case where we are using it for security. We are having to retrofit a system with hundreds if not thousands of DB connections and be able to get the user's id from one end of the system to the other. From what I can tell, it is pretty is slow < 1.3, not usable in 1.2 and is supposed to be a lot faster in 1.4. People make the same bitches about using it that they'd use on global variables, but as a retrofit mechanism, it seems to be the way to go.