|
Authored by: Kilz on Saturday, April 21 2012 @ 04:13 PM EDT |
but I don't understand the Google bias here.
Google pulled a
schtick on Oracle, and now Oracle fights
back. It's simply two corporations
using the court as
another battleground.
I think that the open
source community as a whole thinks
Google is a good corporate member of its
community. They
give things back, they try to do the right thing most of the
time. They create open source projects and give them away.
Are they perfect?
No, but they are pretty good.
On the other hand Oracle is seen as a "user".
They use and
buy things. They dont give back a lot but they take. An
example
of this is their linux distro. They take red hats
stuff and add very little to
it. Larry Ellison is seen as a
not so likable fellow. He doent do things to
make any
friends but takes, and takes, and takes. Are they evil? No,
but they
give off the perception that they are, at least
IMHO. This useless court case
has made them look even worse
in some peoples eyes. Lord help them if they win.
IMHO it
will be like SCO and how people think of them.[ Reply to This | Parent | # ]
|
|
Authored by: Christian on Saturday, April 21 2012 @ 04:17 PM EDT |
The real bias here is a bias toward ownership rights being limited by law and
the Constitution, not infinitely expansive if the party involved is wealthy
enough. If Oracle is granted copyrights to things that are abstract or
functional, it greatly reduces the rights and freedoms of the individual.[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Saturday, April 21 2012 @ 04:19 PM EDT |
I'm sorry, but just looking at the source code of the different implementations
of the APIs, it's clear that Google's implementation is not an INFERIOR version.
It's quite superior in many ways.
I also disagree with the idea that this was an easier path for them. They likely
could have shaved off at least half a year from their schedules just buying from
Sun, not to mention all the lost-opportunity cost.
To me, the real business sense of what Google did had very much to do with
control. Once they purchased Sun's API, it would be very hard for them to come
out with a better version and to control that. This superiority of
implementation is what really sets Google's products apart, which is why they
are as successful as they are.
[ Reply to This | Parent | # ]
|
|
Authored by: Guil Rarey on Saturday, April 21 2012 @ 04:30 PM EDT |
This really, really, really reads like PR spin from Oracle. "Two fierce
warriors"??????? Please. If you want to post assertions like this, please
don't do it AC.
---
If the only way you can value something is with money, you have no idea what
it's worth. If you try to make money by making money, you won't. You might con
so[ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Saturday, April 21 2012 @ 05:32 PM EDT |
What you are calling a shtick to exploit a loophole is pretty much how the
entire rest of the software engineering community has always assumed things are
supposed to work.
If making your own mostly compatible re-implementation is not allowed, we're all
(probably including Oracle itself) in for a world of hurt.
To put it in another perspective, there was a fair amount of disappointment from
many that Google went to great trouble to give Android a non-GPL userspace. Yet
have we heard one hint of a suggestion from leaders in the GPL community that
what Google did was actually not allowed? No, for the simple reason that what
Google did in re-implementing alternatives to GPL components (libc, java,
busybox, etc) under different "politics" is basically equivalent to
what GNU did in re-implementing alternative to Unix utilities under the
revolutionary "politics" called copyleft.
In a world where what Google set out to do would be illegal, there would be no
GNU.[ Reply to This | Parent | # ]
|
|
Authored by: jbb on Saturday, April 21 2012 @ 07:04 PM EDT |
Anon opined:
they'd [Google] use an inferior, yet passable virtual
machine that could run Java (Dalvik).
This is wrong and this
mistake taints the rest of your analysis. A different anon pointed this out
above but I will elaborate. One way to see this is in
statistics:
Number of smartphones using JVM: 0
Number of
smarphones using Dalvik: 250 million
The reason for this
disparity is that the JVM is not suitable for use in smartphones. It is okay
for featurephones but not smartphones. This is why Google diverged from
the JVM. It did not work. This is probably also why the licensing deal between
Sun and Google fell through. It's clear Google knew the JVM was inadequate. I
imagine Sun was unwilling to make the fundamental changes that would be required
in order to fix the JVM.
I think that is why it made no sense for Google to
make a deal with Sun even if Sun paid them.
As was reported here on
Groklaw over a year ago, Google made two fundamental changes (compared to all
JVMs) when they put the Dalvik VM into Android. Both changes created a deeper
connection between the OS and the VM. I can understand why Sun would refuse to
adopt these changes to their JVMs because the tighter OS integration would make
it harder to port the JVM to other platforms.
The first change had to do
with power consumption. It is explained here.
This change also required deep changes in the Linux kernel. It took years for
those changes to filter back and join the main Linux branch. This was not
because Google was withholding the code. It was because the changes were deep
and it took a while to reconcile them with the rest of the Linux
codebase.
The second change had to do with security. The JVM has a very
primitive security model reminiscent of the DOS single user security model. All
applications that run on a single instance of the JVM run as the same user, the
user who launched the JVM. If you want to have an app store then this primitive
model is unacceptable. It would be prohibitively inefficient to run a different
instance of the JVM for every app. What Google did is they made it so each app
runs as a different Linux user on the same VM. This was a brilliant solution!
It provided the more advanced security model needed for and an app store
without harming the efficiency at all. In addition,
most programmers are
already familiar with Linux/Unix/BSD permissions so very little new learning was
required.
The main difference between a feature-phone and a smartphone
is that there are no 3rd party apps on feature phones. Feature-phones come with
a fixed set of features from the manufacturers, hence the name. This explains
why the crude security model of the JVM is suitable for feature-phones but not
smartphones. In a feature-phone all of the applications can be trusted so
security between applications is not an issue. It is a huge issue for
smartphones.
If you don't care about battery life and you don't care about
security then the JVM is perfectly fine for smartphones. But it you want to
make a smartphone that can actually compete in the marketplace then you need to
use a more advanced VM such as Dalvik.
--- Our job is to remind
ourselves that there are more contexts than the one we’re in now — the one that
we think is reality.
-- Alan Kay [ Reply to This | Parent | # ]
|
|
Authored by: jbb on Saturday, April 21 2012 @ 07:53 PM EDT |
Anon opined:
Sun wanted to specifically restrict implementations on
mobile handsets, and so had the field-of-use restriction on the TCK, which is
required if you want to call your implementation "Java".
So Google went
ahead and found a loophole. They wouldn't call their implementation Java
(eschewing the need for the TCK), [...]
Your attempt to re-write
history in order to make Sun/Oracle look like the victim instead of the villain
is pathetic.
Apache was an active member of the Java community. There
were openly developing their open-source Harmony implementation of Java as per
the Java Specification Participation Agreement (JSPA). That agreement does not
mention any end use restrictions. The only hurdle Harmony had to pass was to
prove their implementation was technically compatible which is perfectly
reasonable.
Your re-write makes no sense. If Sun wanted to require end
use restrictions all along then they would have been in another clause in the
JSPA. That way everyone would know up front what the game was. This is the way
honest and honorable people do business. That is not what Sun did. They were
either scheming and conniving from the start or they changed their minds midway
and kept this change of heart hidden for as long as possible.
The key
thing is that Sun did not inform Apache about their plan to effectively add end
use restrictions to the JSPA. Sun was well aware of Apache's Harmony project.
You can see open letters from Apache to Sun spanning the years when Harmony was
being developed. Sun also knew that end use restrictions were incompatible with
the open-source license used for Harmony. Sun knew that their change in plans
would effectively scuttle the Harmony project and instead of informing Harmony
about this ASAP they strung Apache along for years and gave them the bad news as
late as possible.
The way Sun effective altered the JSPA in order to add
their new field of use restriction was with a nasty dirty trick that was
certainly unethical and possibly illegal as well. Sun couldn't alter the JSPA
directly because Apache would never agree to a change that would flush their
years of hard work down the toilet.
What Sun did is they added the field
of use restriction to their Technology Compatibility Kit (TCK). As defined in
the JSPA and as implied by its name, the sole purpose of the TCK was to ensure
an implementation was technically compatible with the Java standard. There was
no prevision that said it could also be used to alter they fundamental terms of
the JSPA such as by adding a field of use restriction.
You are implying
that Google was bad because they weren't clairvoyant enough to predict that
Sun/Oracle was planning to screw over Harmony in such a devious and underhanded
manner. In fact, before Oracle bought Sun they were urging Sun to let Harmony
have access to a TCK that they could use. How was Google supposed the predict
that Oracle would pull a complete 180 and go back on their word about keeping
Java Free right after they bought Sun?
The fact is that Apache put years of
effort into developing Harmony based on the assumption that Sun would honor both
the spirit and the letter of the JSPA. The first hurdle Sun created for Harmony
was the exorbitant price of the TCK. There was no hint of end use restrictions.
If Apache knew there were going to be end use restrictions then they wouldn't
have wasted any more of their time on Harmony. In fact, shortly after it
became known that an end use restriction loophole had been added to the TCK,
Apache abandoned the Harmony project. Google was already using Harmony in
Android before the underhanded TCK end use restriction loophole was made
public.
In other words, Sun/Oracle's end use restriction dirty trick was
made in response to Google's choice to use Harmony. Google did not choose
Harmony to evade Sun/Oracle's end use restriction. The restriction was added to
hobble Google's use of Harmony.
--- Our job is to remind ourselves
that there are more contexts than the one we’re in now — the one that we think
is reality.
-- Alan Kay [ Reply to This | Parent | # ]
|
|
Authored by: PJ on Saturday, April 21 2012 @ 08:45 PM EDT |
Well, you are overlooking that there is a FOSS
perspective, because the API copyrightabiliy
issue will affect all developers, not just the
two fighting, and all software development, not
just Java.
Also, unless you have information you'd like to
share, what you wrote about Dalvik is not true.
Oracle has, by the way, already stipulated that
it is not suing over Dalvik.
So I question you technical descriptions, since
the one I happen to know something about is not
accurate.[ Reply to This | Parent | # ]
|
|
Authored by: tknarr on Saturday, April 21 2012 @ 09:59 PM EDT |
The thing is, thought, that Dalvik doesn't fragment Java. Where it implements
the Java class library, it implements it exactly as it's specified. There aren't
any missing functions. There aren't any functions that do different things than
what the specification calls for. There aren't any new functions that weren't in
the spec. When you're working with those packages, you work with them exactly as
they are in any other implementation of Java out there.
What Dalvik does
is completely drop those parts of the Java class library that aren't
relevant to smartphones. For instance the very large, very unwieldy and
completely useless java.awt and javax.swing trees that provide
Sun's idea of what a GUI implementation should be. This doesn't bother
smartphone programmers, they never even think about using those parts in the
first place because they just aren't suited to small-area touchscreens (not to
mention the performance is poor even on powerful desktop systems, on much
lower-powered smartphone processors it'd be downright physically
painful).
Sun's goal was of course to make Java unsuitable for mobile
development, forcing people to use the Java ME platform instead (which was never
made openly available and which required special tools Sun could charge for).
But Java ME itself wasn't very suitable, and developers simply didn't want to
use it. Sun deluded itself into thinking it could force developers to develop
applications the way Sun wanted them to instead of the way the developers
wanted, and paid the price. [ Reply to This | Parent | # ]
|
|
Authored by: Anonymous on Sunday, April 22 2012 @ 09:45 AM EDT |
"Oracle's Java implementation is based on the OpenJDK which is
GPLv2,"
Is not true, it is the reverse, if anything at all.
And at that OpenJDK is clean room implementation based on Sun/Oracles Java
Specification.
Which is not the same thing, as Oracles Implementation.
[ Reply to This | Parent | # ]
|
|
Authored by: Gringo_ on Sunday, April 22 2012 @ 10:04 AM EDT |
Google uses Harmony. Harmony is not Java (thanks
to Sun not giving it
the TCK) so how can anybody say Google
fragmented Java?
What Google did
was, in fact, brilliant! Without the TCK,
Harmony was dead in the water. Google
gave it life and a
purpose, and a new name - Android. It is ironic that Sun
itself inadvertently opened the door for Android by refusing
Harmony the TCK.
All in all, what Google did is sort of
analogous to a brilliant remix in the
field of
entertainment.
All the above in this thread has been a
wonderful
discussion - provoked by somebody with a negative view of
Google,
whom I called a contrarian. His arguments
have been thoroughly rebutted,
and to have that collection
of rebuttals here in one thread makes them an
excellent
resource for anybody who has been confused by Oracle's
propaganda.
Nothing remains of the contrarian's arguments. They have
all been torn to shreds. All that is left are his
opinions....
Google has fragmented Java.
The legal
arguments are, of course, very interesting, but
I don't understand the Google
bias here. Google pulled a
schtick on Oracle
Oracle is simply using the
API copyright as a legal tool,
a tactic, to try and nail Google for something
that Oracle
believes (and I do, too) of very questionable ethics.
Is
this lawful? Maybe. But it's certainly a nasty
business trick. So now Oracle is
trying to do a courtroom
trick. That's all.
It was I who
jokingly made that remark above about
"stinkin' contrarians" and "heretics"
late last night when I
wasn't logged on. I think it is important to own up to
that,
lest someone think it was a sock puppet of the contrarian
himself who
injected that. The reason I said that was
because I felt the contrarian was
sincere but misguided. The
wonderful rebuttals he provoked were a net gain, and
neutralized his disparagements about Google. Just before my
comment, PJ had
jumped in, basically accusing him of being a
troll. At that moment, I felt a
little sympathy for the guy,
even though I soundly disagree with him. Now, I am
inclined
to agree with PJ. Of course, PJ has a wealth of experience
after
running Groklaw all these years, and by now she can
probably spot a troll from
a mile away.
Having been thoroughly rebutted, all that remains of the
the contrarian are his opinions, which he has every right
to. We cannot take
them away from him. But now when I look
at the opinions expressed by the
contrarian, I wonder what
kind of person could hold such opinions, in light of
the
facts? He seems an intelligent person. I am suspicious that
these
statements do not represent honest opinions at all.
They look more like they
represent an agenda. Maybe
not from an Oracle sympathizer - more like
from one of the
"Google is Evil" shills. I think PJ's interjection was right
on the money. [ Reply to This | Parent | # ]
|
|
Authored by: Dave on Sunday, April 22 2012 @ 11:51 AM EDT |
Well, that's certainly *one* way of looking at it. Others have a different
view.
As I see it, Sun made some very public decisions over the last 15 years that
affect a lot of people, and Oracle just wants to undo them now to get a slice of
the mobile pie that they never earned. That kind of double-cross never goes
over well with the tech community.
Open or closed? It's always a tough decision for a company to make. Over the
past 15 years, Sun made Java more and more open, but decided to use the
name Java as the point of control for the language, platform, and community.
They were clear: third parties were allowed to implement the language spec,
build compilers and VMs, rewrite the platform libraries, but they couldn't use
Sun's code without a license (by the end, GPL was an option), and they
couldn't call it Java without Sun's permission.
The idea was simple: try to build an open community of interest around Java,
but restrict commercial use via the name. After all, it was Sun's Java empire
that was strong and credible in the business world. Java runs the most secure
Web sites, they told us (on IBM and WebLogic's application servers). Millions
of devices ran Java (though did any of them ever do anything useful or
important with it?).
But Google had the audacity to take Sun at their word and succeed. Android
has found far more success on devices in the last 4 years than Java ever did
in the same time and the decade before. The Java brand has been completely
devalued in this space.
I'm sure Sun wasn't happy about it, but they recognized that Google was
playing by the rules they had set. Oracle, by contrast, decided to lawyer up and
try to rewrite history. They deserve to lose, and it's increasingly looking like
they will.[ Reply to This | Parent | # ]
|
|
Authored by: SpaceLifeForm on Sunday, April 22 2012 @ 12:58 PM EDT |
You wrote:
Google didn't like the GPL licensing because it would
require handset
manufacturers to open-source their modification to the
code.
That is a false statement. The apps, written in
Java,
do not have to be licensed under GPL at all. Some are in fact.
The
handset manufacturers in fact, have no need to worry
about any mods, unless they
are to the Linux kernel or
to the userland apps that are GPL. Yes, there is GPL
userland code in Android, but there is likely no need
for a manufacturer to
modify that code anyway.
The handset manufacturers would normally be doing
their
branding customization to the apps that are visible to
the user on the
screen. Apps written in Java.
What is interesting is the timeline of
events, that
may tell us some things.
2003 - Android project started (this
is not Google)
2005 - Apache Harmony project started
2005 - Google buys
Android
2007 - OpenJDK started
So, between 2003 and 2007, a lot of
technical decisions
had to happen to get us to where we are today.
But, at
this point, I believe it should be feasible
to replace Harmony with OpenJDK on
Android.
And that would have no impact on the manufacturers.
But, it
would impact the legal arguments from Oracle, no?
---
You are being MICROattacked, from various angles, in a SOFT manner. [ Reply to This | Parent | # ]
|
|
|
|
|