Java 1.5.0 Now Officially Java 5.0

Quantum Jim writes "In a move which out-does Netscape's one-version number skip and Winamp's two-numbers skip, Sun has announced that the upcoming Java2 release will be marketed as version 5.0, skipping three-and-a-half numbers. Can version 6.022E23 be far behind? Thanks to David Flanagan for the heads-up."

Other Famous Version Number Skips

[Anonymous Coward]

Slackware's comes to mind. Any others?

Winamp didn't skip version 4

[MikeXpop]

Winamp 5 isn't exactly Winamp version 5. It's more like 3.5. They used the number 5 because they wanted the features of 3 with the speed and ability of 2. 2+3=5. And that's where they got the number.

Strongly Typed Container Classes

[NitsujTPU]

From briefly viewing some literature about Java 1.5.0 (er, 5.0... W0w!) the feature that excites me most about this is the ability to strongly type container classes, such as one can do in Ada or C++.

Joy.

Re:Other Famous Version Number Skips

[Shulai]

Pike programming language (underused but very nice indeed, I prefer Pike to Java) comes from version 0.6 to version 7.0

Re:Where does the 2 come into this?

[barcodez]

When Java SDK went from 1.1.x to 1.2.0 they decided that they had made lots of big changes (IIRC Swing and Collection.. possibly Inner classes *shrug*) so they called it Java 2. However when they went from 1.2 to 1.3 they hadn't made too many major changes so they didn't bother and the same for 1.4. There are lot's of changes in 1.5 so I guess they thought they should give it a new number. However Java 2 version 5 is stupid as is J2SE 5 and J2EE 5 - all very confusing for everyone.

Re:Strongly Typed Container Classes

[MarkWPiper]

While this is a nice feature, it is strictly (at least for now), syntactic. The difference is that the expense of casting is still occurring under the hood; you just no longer have to bother typing it out. I believe people are referring to it as 'autoboxing'. Therefore, these strongly typed container classes are not as powerful as C++'s templates.

I read a pretty good interview w/ Eckel and that guy who has done most of the work on C#. The creator of C# was bashing Java's generics, because they aren't giving the full performance possible. And I agree. There is still such a thing as performance critical code, and Java can make it frustratingly hard to write it. Providing featureful, fast data structures would be a good place to start.

I can't find the article I'm referencing, but this sums up Eckel's view. [mindview.net]

Pizza [sourceforge.net] was an alternative implementation of generics for Java. I wish that Sun had chosen this project as their basis for 1.5's generics, rather than GJ (Generic Java). I believe its implementation is much closer to that of C++'s templates. I'd love to use pizza, but it's just not wide-spread enough to justify it in enterprise code.

Re:95 - 2000

[basics]

actually, since 2000 == nt++, its 4 - 2000.

Re:Other Famous Version Number Skips

[Kourino]

Emacs.

Some time ago, the developers realized that GNU Emacs would probably never change its major version number (which is 1). So, after some point, instead of "GNU Emacs 1.x.y", they started dropping the 1 (since it was constant information and therefore redundant). So the current release of GNU Emacs is actually 1.21.3, but it's called "GNU Emacs 21.3".

This actually appears to be what Sun is doing now. They've done it before with Solaris/SunOS ... twice, in fact.

versions of tomorrow

[MrLint]

Well there has been some speculation for the past several years what will happen when apple set to MacOS 10.9 (X.9)?

Will it be 11? XI?

As a Java developer, I'm all for it

[damm0]

Hey, if it works as a marketing ploy and increases the number of people who want my skills, I'm all for it.

The story behind Java2

[Anonymous Coward]

As Sun was preparing to roll out JDK 1.2 the marketing department thought the name lacked enough flash considering all the new features. They had picked a new name, "Java 2000", and were ready to run with it when Microsoft annouced that NT 5.0 would be known as Windows 2000. I don't know if Microsoft always intended to call NT 5.0 Windows 2000, or if it was a case of them one upping Sun. Either way, the Sun marketing department was caught flat footed and needed to come up with a new name fast. That's why we have Java2.

As far as changing the name of 1.5 to 5.0, I think it makes a lot of sense. The original reason for sticking with 1.X was that a full version jump would indicate incompatibility between versions. That's never going to happen now. Incompatibility would be the death of Java, so the time is right to start using full version numbers for major releases. 1.5->5.0 shouldn't be that hard to wrap you head around.

Re:Strongly Typed Container Classes

[theefer]

If you liked Pizza, maybe you'll like Scala [scala.epfl.ch], a functional, multiparadigm language developped by the same author, Martin Odersky. I have him as a programming teacher, and we learnt functional programming with Scala. It was a great course, and the language is really elegant and powerful.

It has bindings with Java and .Net, but remains functional-oriented.

Re:Other Famous Version Number Skips

[aacool]

A picture editing program I use (and like) called Picasa went to version 1.618 from version 1 - no prizes for guessing why

BMW & Canon

[Max Threshold]

Similar to BMW's Z8... it's sort of like a Z3 with a M5 engine. Or at least that's the implication.

OTOH, there is absolutely no logic behind the U.S. market designations of Canon's mid-range SLR bodies. They went from the Elan, to the Elan II, to the Elan 7, to the Elan 7N.

Re:Other Famous Version Number Skips

[Alexis de Torquemada]

Did Warhammer 40,000 have a predecessor? If not, that's a daring version number for a first release.

Re:Where does the 2 come into this?

[still_nfi]

I believe it has to do with the underlying VM. There were changes in the VM going from 1.1 to 1.2. Hence the name change to Java2

1.3, 1.4 & 1.5 have not made any changes to the VM, therefore, they are still being called Java2.

When sun find a compelling reason to make changes to the VM, you can expect to see a Java3 1.0. That's pretty unlikely to happen for quite some time though.

Re:Other Famous Version Number Skips

[jrockway]

Well the first version of TeX was 3.

Then there was 3.1
then 3.14
then 3.141 ...
now it's 3.14159

And don't even ask about Metafont...

Common for Sun.

[jellomizer]

Take a look at solaris versions. Well there were 2 sets of versions Solaris and Sun OS

There was Solaris 2.5 and 2.6 then they made solaris 7 aka 2.7 then solaris 8 and solaris 9 and if you check the version numbers you get Solaris 2.9 and SunOS 5.9

This seems to be common for sun when their product seems to reach maturity and they are not planning on doing a major overhall to their product they will drop the first diget then make the 10s spot the version number. I Think it is more for an advertisement thing because a lot of people dont like getting incremental updates .1 .2 .3 they like getting v2 v3 that way it sounds like they are getting a major version change. But with Java Code being stablized people are still on 1.2 and 1.3 where they really should be at 1.5 for best functionally.

Re:Other Famous Version Number Skips

[1u3hr]

Dont forget there was a word for Unix... I still have the SCO (forgive me for mentioning them) version somewhere around here ;-)

Well, I didn't know that. But SCO Unix was originally MS Xenix. Microsoft Word for UNIX Systems Release 5.1 [opsys.com]:

Microsoft(r) Word for UNIX Systems Release 5.1 from SCO is the most powerful multiuser word processor available today. It brings the best MSDOS word processing features to multiuser UNIX Systems. Ideal for offices of any size, Microsoft Word 5.1 allows you to effortlessly combine text, graphics, and data to create virtually any type of document-with excellent formatting and printer support. Microsoft Word for UNIX Systems is designed for today's demanding professionals who want the best word processing features.

Apparently discontinued in 1996. Obviously, it could be ported to Linux trivially, Bill's dead body being the only obstacle.

Re:*sigh* I hate marketing

[dasmegabyte]

You're joking, right?

You're asking how sun can compete with Microsoft's .NET initiative in terms of confusing people?

Tell me, what version of the .NET framework are you running? What version of studio .net and what version of the project files? Do you know the differences between the syntax of line end points between library version 1.0.3333 and 1.0.5000 (the library version used with Framework 1.1)? Have you checked your global assembly cache lately?

Shit man. Most people still don't even know what .NET is. I've been writing in it for two years and my boss still thinks it has something to do with the internet maybe. The versioning system is very complex and promises to be reminiscent of DLL hell. Each version of the Framework has wierd, subtle bugs that pop up at the strangest times...there's one with visual inheritance and the passing of alt-key mnemonic events in VB variables declared WithEvents that will probably keep us off of Framework 1.1 forever. Luckily for my support department, .NET (unlike Java) is designed to maintain compatibility with previous versions, not by keeping deprecated methods, but by keeping the old CLI and Framework on the machine when a new one's installed (at about 140 meg a pop).

Incidentally, Java 2 is the platform. Java 2 competes with .NET. The other number, 1.x, is the version number of the runtime or the compiler/sdk for the Java language. Java 1.5 competes with C#. How hard is that?

Re:Other Famous Version Number Skips

[dtfinch]

Windows 2000 to XP
XP is a long used symbol for Christ, dating back to some of the earliest christian artwork, with the symbols pronounced Chi-Rho in Roman. We could say the year of XP is 1.

So 1999 version numbers skipped backwards.

Re:Strongly Typed Container Classes

[julesh]

While this is a nice feature, it is strictly (at least for now), syntactic. The difference is that the expense of casting is still occurring under the hood; you just no longer have to bother typing it out.

Are you sure? I understood that it worked by generating implicit subclasses of your generic type that are customized in terms of return type. So if you had say an object of type Iterator, the run-time environment would procede as if the object's next() function was declared as "String next ()" ?

Although I'll admit its a good 6 months since I read it, and am hazy on some of the details.