To Avoid Confusion: Oracle's Confusing New Java Numbering Scheme

twofishy writes "'To avoid the confusion caused by renumbering releases,' Oracle has announced that it is adopting a new numbering scheme for JDK 5.0, JDK 6 and JDK 7. 'The next Limited Update for JDK 7 will be numbered 7u40, and the next 3 CPUs after that will be numbered 7u45, 7u51, and 7u55." The vendor notes that a more elegant solution would require the changing of the version numbering scheme to accommodate different kinds of changes (for example by using 7u44-2 ). However this cannot be implemented outside of a major release, since doing so might break existing code that parses version strings (possibly including the Java auto-update system)" Here's Oracle's announcement.

Comments

CPU=Critical Patch Update

[fnj]

An absurd TLA overloading.

Re:CPU=Critical Patch Update

[DickBreath]

> An absurd TLA overloading.

To be more concise, simply say: ATO. (ATO = Absurd TLA Overloading.)

Re:CPU=Critical Patch Update

[ArcadeMan]

Oh, a different meaning for an acronym that's been in use for over 50 years. That will sure help with the confusion!

Re:CPU=Critical Patch Update

[Anonymous Coward]

critical update needed today

Re:CPU=Critical Patch Update

[chinton]

That's like overriding toString() to format your hard-drive.

Re:CPU=Critical Patch Update

[gstoddart]

So it's a TLA about a critical CPU that's gonna clear up all this FUD.

LOL, WTF?

Re:CPU=Critical Patch Update

[Tackhead]

Its not like they could have just said Critical update patch...oh no, we need to make things confusing.

What happens when admins get confused and pour the contents of their beverage containers into their servers?

We apologize for the confusion in the Critical Patch Updates. The individuals responsible have been sacked. To avoid further confusion, all CPUs will be processed through CUPS, the Critical Update Patch Server.

And now the goddamn printer doesn't work.

Re:CPU=Critical Patch Update

[Anonymous Coward]

No 3 letter acronym is unique. At best, it's unique in the domain you are currently working in.

Well this one sure as hell fails that test!

Why Does Name Matter?

[Karganeth]

Surely they could just number them from 1 to infinity. Why does name matter? Who knows.

Re:Why Does Name Matter?

[bunratty]

Because programs are used to decode/encode the name. It's the same problem with Y2K, user agent strings, and so on. When programs expect data in a certain format, such as two digits for years or a single number after a u in a version string, they don't react well when the format is changed. RTFA.

Re:Why Does Name Matter?

[LordLimecat]

Because the old numbering system of incrementing by 1 was far too complicated. The new system is much easier to understand:

* Limited Update releases will be numbered in multiples of 20.
  * We intend for Critical Patch Updates to continue to use odd numbers. The numbers will be calculated by adding multiples of five to the prior Limited Update and when needed adding one to keep the resulting number odd.
  * Updates containing Norton instead of McAfee will be prime numbers
  * If an update actually contains intentional regressions, it will end in a 9. We recommend you do not install these updates.
  * If the update number is a prime but also a member of the fibonacci sequence, it is considered a "surprise" update: It will be automatically installed, and will contain 1 or more undesirable features. It may also install the ask toolbar.

Obviously this system is far superior.

Re:Why Does Name Matter?

[unrtst]

Because programs are used to decode/encode the name.

...except that isn't the problem.

They are using an incremental number on the RHS (right hand side) of the "u" seperator, and that number may represent either a critial patch update or it might be a new minor version. The problem arrises when they attempt to associate FUTURE releases with stuff in bugtrackers and their ilk. They could have more CPU releases than normal releases, making the CPU minor release number much higher than the planned normal release, causing a bunch of confusion.

Their solution, IMO, blows. But it's not because of the format of the version string.

IMO, the easy solution would be to treat it as a 3 part number, and make the current minor version part MUCH larger. So, first increase the minor version:
        5u40 becomes 5u00040000

That can be interpretted as: 5.00040.000

Then it's easy. Middle part is the normal incremental releases. Last 3 digit are for CPU releases on that incremental. FWIW, this provides FAR more room for CPU releases than this scheme they just announced, and it's very clear and easy to read. Most languages aslo support some character to allow the user to format the number within the codebase (in Perl, it could be written as 5.00040_000 and perl will just ignore the _), so it would still be crystal clear when written as well.

Anyway... this is NOT the same as Y2K, nor as user agent strings, and doesn't have anything to do with programs expecting the version string to be a specific format, except for the fact that they can't change the current format. How they plan to work within that limitation and the issues they have encounted is the problem.

Just like line numbers in BASIC

[Errol backfiring]

You know, to be able to insert something between two existing lines.

Why not copy MS and have 2 ver numbers

[Joe_Dragon]

Why not copy MS and have 2 ver numbers like how windows 7 is also 6.1

Re:Why not copy MS and have 2 ver numbers

[ArcadeMan]

Who would be stupid enough to check for the OS version that way? It should have been something like:

if ( ( MajorVersion * 10 ) + MinorVersion >= 51 )
{
}

Re:Why not copy MS and have 2 ver numbers

[X0563511]

Wrong. Ubuntu uses YY.MM date-based numbering.

Re:Why not copy MS and have 2 ver numbers

[jellomizer]

There was a lot of terrible code that happened back then.
From around 1996-2006 was a real low point in Quality Code.
Before then Most Applications were DOS based, or mainframe based. Confined to forms that fit in 80x24/25 and did rather simple things. Then with Windows 95 release developers started a mad Rush to to make their Apps Windows Compatible. For many developers at this time, they didn't have much experience with Windows development and many developers were not really prepared for the complexities of win development. Tools like Visual Basic and Visual Fox Pro made such development easier, They tend to create code that had a lot of interesting compatibility problems, especially with people still using Windows 3.1 or earlier versions of windows, Windows at the time was in a lot of changes fixing problems (which broke hacked work around to other problems) 95, 98, ME, 2000 XP. All really had their own development issues.
Post 2006 is when .NET started to kick and be popular in and fixing a lot of compatibility issues, with Windows. As well as more advanced development better made for the newer OS's that is less version dependent.

IQ

[rossdee]

Mensa can use this as a 'What is the next number in the sequence' for their IQ test.

Re:IQ

[SJHillman]

I hear Mensa is planning on using the Microsoft Copy dialog instead.

"You have 4 minutes left, then 3 minutes, then 29 minutes, then 1800 years, then 32 seconds, then potato. What is the next number in the sequence?"

Re:IQ

[riT-k0MA]

"Insufficient Disk Space"

Re:IQ

[Barefoot Monkey]

That's easy. The next number in the sequence is (not responding).

Re:IQ

[ArcadeMan]

"You have 4 minutes left, then 3 minutes, then 29 minutes, then 1800 years, then 32 seconds, then potato. What is the next number in the sequence?"

A blue screen.

Oracle has always had confusing version numbers

[jfdavis668]

Every time they try to standardize version numbers, they make it more confusing. 11G database release 1 was 11.0, but release 2 is 11.2. Where was 11.1? App server 9i was actually an 8.0 base. Most of the time I can't even figure out which product I am actually buying.

Re:Oracle has always had confusing version numbers

[stevencbrown]

No it wasn't...
11gR1 is 11.1.x, and 11gR2 is 11.2.x.

11.0 doesn't exist.

I agree 11g badging has been a bit confused for some products (11g WebLogic Server is 10.3.6), but the database versioning is pretty clear.

Simpler solution..

[xtal]

I resolved the issue by removing it.. perhaps, that is the intended effect. Apologies to those with no choice.

Ask Toolbar

[Anonymous Coward]

I don't care what they call it so long as I don't have to install any toolbars.

Like seriously what is this 2003? It's crowding out my "Bonzi Buddy"

Apache

[slash.jit]

Oracle... we don't care. Please just donate Java to Apache.

Obviously attempting to confuse the Malware

[sl4shd0rk]

It was easier to change the version numbering scheme to break all the current malware rather than fix the myriad of holes in the code.

Re:All eggs in 1 basket. It may have been nice...

[devman]

Java 6 and 7 are both GPL'd with OpenJDK. Nothing says you have to use Oracle's JRE/JDK.

Wow. Just wow.

[QilessQi]

After reading the Oracle announcement through three times, I actually checked my calendar to see if April 1st had somehow come around again.

Which is a distinct possibility if Oracle takes over the Gregorian Calendar next.