java 6 used in large companies?

[elh.maayan]

hi..

we are a large company considering the move from java 1.4 to either 5
or 6.

at the same we are upgrading to was6 (or maybe jboss 4), we are
interested to learn if there are any major projects on enterprise
level involving java 6 (including the use of it's web services
framework).

 

[anderson.ab]

hi..

we are a large company considering the move from java 1.4 to either 5
or 6.

at the same we are upgrading to was6 (or maybe jboss 4), we are
interested to learn if there are any major projects on enterprise
level involving java 6 (including the use of it's web services
framework).

We chose Java 5 and was6. That is all that we need. If you are a
large company, why so worried about web services specifically, why not
look at WESB and/or look open source (Synapse). Any ways, Java6 would
work I guess but we went to Java 5 for now.

 

[Arne Vajhøj]

we are a large company considering the move from java 1.4 to either 5
or 6.

at the same we are upgrading to was6 (or maybe jboss 4), we are
interested to learn if there are any major projects on enterprise
level involving java 6 (including the use of it's web services
framework).

The enterprise world is always behind.

First the app server vendor need to come out with a version
that support Java X+1, then you certify your app for Java X+1
and then you wait for a suitable opportunity to upgrade from
Java X to X+1.

It all takes time.

My guess is that very few EE sites has switched to Java 6 in
production yet.

But many probably has it on the roadmap.

WAS 6.1 is only certified for Java 5 AFAIK.

JBoss 4.2 is intended for Java 5 as well, but they do have
a patch so that it will work with Java 6.

Arne

 

[Roedy Green]

we are a large company considering the move from java 1.4 to either 5
or 6.

I would think the only reason not to go to 6 is to provide code for
people who use older Java. If the project is purely internal (e.g. on
a server) I can't think of any reason to hold back.

 

[Roedy Green]

I would think the only reason not to go to 6 is to provide code for
people who use older Java. If the project is purely internal (e.g. on
a server) I can't think of any reason to hold back.

There is a natural reluctance to switch. You want to avoid the
disruption. You want to avoid the work of adjusting. But when you do
jump, you might as well jump all the way to 6.0.

 

[Lew]

There is a natural reluctance to switch. You want to avoid the
disruption. You want to avoid the work of adjusting. But when you do
jump, you might as well jump all the way to 6.0.

Java 5 and 6 are the same language. The differences are in the features and
APIs.
<http://en.wikipedia.org/wiki/Java_version_history#Major_changes>

Java 6 has been out for more than a year, now.

If I had decision power for an organization I'd go right to Java 6. The
benefits clearly outweigh the risks by a sizable margin.

 

[Arne Vajhøj]

Java 6 has been out for more than a year, now.

If I had decision power for an organization I'd go right to Java 6. The
benefits clearly outweigh the risks by a sizable margin.

Out of curiosity:
- what percentage productivity gain has you set the new features
in Java 6 to provide ?
- applied to how many hours ?
- what have you estimated the retest cost to ?
- what expected additional downtime have you calculated with ?
- and what dollar equivalent have you set each downtime minut to ?

Arne

 

[Arne Vajhøj]

I would think the only reason not to go to 6 is to provide code for
people who use older Java. If the project is purely internal (e.g. on
a server) I can't think of any reason to hold back.

Small facts like that WAS 6.1 is not supported on Java 6 is not
considered relevant ?

Arne

 

[Roedy Green]

- what percentage productivity gain has you set the new features
in Java 6 to provide ?

The apps run faster.
Bugs get fixed.

It is quite a painless upgrade from 5 to 6. It is not like the jump to
5 where you had to generify everything to take advantage.

I don't think I had to change anything. Just recompile. The same ant
scripts worked unchanged.

 

[Roedy Green]

Small facts like that WAS 6.1 is not supported on Java 6 is not
considered relevant ?

WAS = WebSphere Application Server ?

 

[Arne Vajhøj]

WAS = WebSphere Application Server ?

Yes.

Arne

 

[Arne Vajhøj]

The apps run faster.

Actually a bit yes.

Bugs get fixed.

Can you refer to any source for that 6u4 has fewer bugs than
5u14 ?

It is quite a painless upgrade from 5 to 6. It is not like the jump to
5 where you had to generify everything to take advantage.

I don't think I had to change anything. Just recompile. The same ant
scripts worked unchanged.

That does not justify an upgrade to the bean counters.

Arne

 

[Roedy Green]

That does not justify an upgrade to the bean counters.

I am not trying to justify an upgrade, just if you do one, you might
as well go to 6.

 

[Arne Vajhøj]

I am not trying to justify an upgrade, just if you do one, you might
as well go to 6.

So you have tested WAS 6.1 on 6 so you will vouch for that combo and
have some statistics for that 6u4 has no more bugs than 5u14 ?

Or did you just flip a coin to decide on that recommendation ?

Arne

 

[Andreas Leitgeb]

So you have tested WAS 6.1 on 6 so you will vouch for that combo and
have some statistics for that 6u4 has no more bugs than 5u14 ?
Or did you just flip a coin to decide on that recommendation ?

The context was, whether from 1.4 to jump to 1.5 or 6.
Jumping to 1.5 means all the effort of testing, checking
for bugs that bite, that is now done for 1.5 will sooner
or later have to be re-done for 6. So if testing for
those bugs is, what puts the butter on your bread, then
for heaven's sake go to 1.5 now! :-D
Let the bean-counters pay you for their attitude.

 

[Andreas Leitgeb]

Small facts like that WAS 6.1 is not supported on Java 6 is not
considered relevant ?

[/QUOTE][/QUOTE]

If it *is* relevant, then there's probably no choice to begin with.

 

[Arne Vajhøj]

If it *is* relevant, then there's probably no choice to begin with.

The original poster mentioned WAS 6.1 as most likely app server.

Arne

 

[Arne Vajhøj]

The context was, whether from 1.4 to jump to 1.5 or 6.

That was the original context.

And the >> is indeed about whether 1.4->6 is a possible/better than
1.4->5.

The >>>> was a response to:

#It is quite a painless upgrade from 5 to 6. It is not like the jump to
#5 where you had to generify everything to take advantage.
#
#I don't think I had to change anything. Just recompile. The same ant
#scripts worked unchanged.

which is about 5->6 !

It did become a mix of original posters problem and general
upgrade practice. Probably my fault because I get a bit upset
when people start with the "just upgrade to latest" philosophy.

Jumping to 1.5 means all the effort of testing, checking
for bugs that bite, that is now done for 1.5 will sooner
or later have to be re-done for 6. So if testing for
those bugs is, what puts the butter on your bread, then
for heaven's sake go to 1.5 now! :-D
Let the bean-counters pay you for their attitude.

No.

Nobody said that you can not skip versions.

Arne

 

[Andreas Leitgeb]

No.
Nobody said that you can not skip versions.

Obviously then, my rant was about a not-necessarily
existing virtual sub-class of class Beancounters, who
- given the choice - would choose always the smallest
upgrade-step possible.