Lutris Closes Enhydra Source

Ron van Balen writes: "Lutris has retracted the open source Entreprise Enhydra product. The old version will remain open source, but the open source community will not get access to the new J2EE compliant product. The decision was made because Sun J2EE license requirements don't allow an open source release, Lutris says. Lutris also says it wil refocus its efforts to its commercial products and support the open source community at a lower priority. It seems there is one less commercially supported OSS project on the planet." Newsforge has an excellent piece on this as well which gets into the reasoning and details on this move.

Direct complaints to the right place

As indicated above, the reason for the closing of source is the J2EE license from Sun. All complaints should be addressed to Sun Microsystems [sun.com].

Blame Sun

Sun refuses to open Java. No, the JCP is not open. Why at this stage Sun does not open up Java is beyond me. With Microsoft out of the Java community, I don't see ravenous, hostile competitors chomping at the bit to deform Java and destabilize Sun.

They're only hurting themselves and developers with their idiotically stubborn unwillingness to get with the program.

Free Java implementations?

I know about Kaffe, but I just checked www.kaffe.org and it hasn't been updated for over a year. Why has it died? Legal reasons? Lack of interest?

Japhar is another implementation, but it is in a very early stage (current version 0.10).

Do other implementations exist?

Non J2EE App Servers legal?

Are there different licenses for projects like Tomcat? Can you deploy them legally?

What is the future of the JBOSS project now?

If what Lutris is saying is true, will the JBOSS project be able to continue? They are shipping an open source J2EE project now..

XMLC is the core of Enhydra

For lots of projects, XMLC is good enough. Let's keep using the open source part, and avoid J2EE when we don't have to. This is the kind of message we need to give back to Sun.

Uh, this is OLD NEWS- and the title's misleading!

This tidbit ran sometime last week on LinuxToday- and the title they're closing the source to Enhydra's misleading; it's Enterprise Enhydra that they're closing the source to, not Enhydra itself.

sad

It's sad that a developers choice of language/tools can make him lose control over his own licensing policy.

Send your complaints to Sun. We have seen the same with eg. Microsofts Mobile something developers kit, and it is very sad.

open source increasingly under siege ...

The policies governing Java technologies never blended well with open source/free software. Given recent events, we should not expect Java to become more open, on the contrary.

Law enforcement agencies will probably demand closed-source, backdoor-enabled encryption and security subsystems. Open source doesn't lend itself to that.

What is also a problem, is the fact that free software yields incredible consumer surpluses, but very little in terms of company profits. Since the government needs companies to make profits, so that they can levy taxes, they will not encourage free software.

The most dangerous assault will, however, come from the copyright cartels. They will not rest until computers fully implement digital rights management in unalterable binary-only distributions, with source code locked up.

In this light, It will become increasingly difficult to defend the values behind free software.

Why use Enhydra?

I'm not sure why someone looking for a J2EE implementation would go to Enhydra. JBoss [jboss.org] is a much better, robust, mature platform for that sort of thing than Enhydra. None of this is to say that Enhydra is worthless - it's very good at what it does, which is a much more lightweight Java web platform than DB + EJB + Servlets + the kitchen sink which is what full J2EE servers are. In fact, most projects would be better off with the lighter-weight Enhydra, especially published-content type projects.

I guess what I'm trying to get at is Lutris should have kept Enhydra the way it was, and not screwed around with J2EE. We have JBoss for that, and Enhydra filled a much different need. The whole mess could have been avoided.

An Alternative

SourceForge [sf.net] has nice projects: Open Business [sourceforge.net] or Enigma [sourceforge.net] for J2EE business software. It is still far from finish, but at least you can help to make it happen.

Newsforge

Are there supposed to be stories there? All I get is the fluff around the edges and links to the previous and next stories (which are also empty).

This one looks like Sun's fault, but InstantDB?

Does anyone remember the beginnings of InstantDB and the Enhydra project? When I originally picked it up, the buzz (and the statements on their webpages) was that its not Open Source yet, but "it will be Real Soon Now (tm)".

After embedding it in my application, I needed to make a couple of changes and went to look for the source, and there was no more talk of 'open source' but rather 'Low deployment license fees'.

Is this somehow related to the J2EE problem (how? its just a SQL db)? Was there another announcement I missed? Or did Lutris excercise their legally allowable but ethically questionable right to say "Its not open source now, we need $$, this is now a product."?

If its the latter, then it makes one wonder how hard they negotiated with Sun.

Zipwow

How did the Blackdown project do it?

I am interested in porting the Web Start product to FreeBSD. However the legal requirements are unclear to me.

I would really love to roll a distribution, like the Blackdown group did with Java3D for Linux, but I don't know how to get the OK from Sun.

What is required? The present solution I see is leaving the user to sign the Sun's Communite Source License [sun.com] himself, and just offering a source patch set. For application a blessed binary release would be much nicer.

Again, is anyone from the Blackdown guys here, how could explain what is needed?

Regards,
Marc

Red Herring

Lutris has been leaning commercially for a while. I think the licensing 'issue' is a red herring they've thrown out as an excuse to transition into a closed-source only product. A few months back when I was trying to get the open source version of enhydra, I couldn't find it. They buried it in their enhydra.com website and redirected enhydra.org to it. Everywhere I looked was a 'purchase' button.

Anyhow, how can JBoss have an open source J2EE implementation ?(which is lightyears better, in my opinion) Maybe becasue they don't have so many suits trying to put a spin on the product in order to get it to sell.

It really seems like Lutris is just trying to transition back to the closed source model because they can't sell an inferior, late J2EE application server when you can see what is 'really under the hood' - an almost J2EE 1.1 compliant application server. They are chasing JBoss' and others' tails on a prior standard even.

I used enhydra 3.01 for a major project and it was/is quite good: scalable, robust and fault tolerant, but it seems to have been poisoned by commercial interest and delays in implementing J2EE.

Lutris is pulling a MySQL

Remember MySQL's old documentation, where it poo-pooh'ed the whole concept of transactions, which they didn't implement at the time? That which they couldn't or wouldn't implement, they trashed with FUD. I read something similar from Lutris [enhydra.org] in their "making waves" column that basically trashes J2EE for being, from what I can decipher from the article, an overall platform name and version for several technologies. The gist seems to be that the name J2EE has so much marketing power that you can no longer use the single pieces of it you need in your application and discard the rest, simply because in order to be branded J2EE, you have to (gasp) comply with the spec. And of course, since the spec is versioned, then well, the little guys can't keep up, so this is Sun's ploy to squeeze them out.

I suppose this confusion is normal when his application happens to be a J2EE app server, but it's utterly absurd and wrong to say that an application running on a J2EE app server is somehow forced into a monolithic API. It sounds like Lutris is just facing the fact that they started with an app server that was not J2EE then went on a crash program to make it so, and are running into a shortage of manpower. So to compensate, they are including the code from Sun's own J2EE reference implementation.

No, I'm not a fan of Sun's closed and expensive testing process, but Lutris's argument isn't about that, and it simply doesn't hold any water. Lutris is using Sun's code, not just their specs, and they are griping that they can't sublicense it however they wish. They might have been better off pulling a Zope instead, and just building on their existing app server and damn the J*-acronyms from Sun. Enhydra was damn functional, but as far as front-ends go, they have a lot of catching up to do with Zope.

Sun man speak with forked tongue

I love this quote from Sun's PR flak:

...we are far closer to the Open Source community than someone like Microsoft and, dare I say, IBM.

1: Sun is closer to Open Source than Microsoft is: It is equally true that Seattle, WA is closer to Mexico than Vancouver, BC is. That doesn't mean they're actually close.

2: IBM: What a load of Lamborghini exhaust! With IBM, you know exactly where you stand. This is Open, that is Closed. Period. There's no lawyer-speak, snake-in-the-grass, hidden-gotcha licence like Sun Community Source License to worry about.

License doesn't allow Open Source

When MS' licence agreement doesn't allow Open Source programs to be developed with it, the world is coming to an end. When Sun's License agreement doesn't allow it, it's just matter of fact. Why the difference in views? Is Sun magically immune from all of the flames that MS gets?

FEAR NOT! The problem is not with J2EE!

The problem here is not with the J2EE license, as some have claimed.

The problem is that Enhydra is based on Sun's reference implementation of J2EE, not on a clean-room implementation like JBOSS. Sun's license for the reference implementation is the problem, not the J2EE license.

OSS and J2EE work together.

Ok, sort of a dumb question...

but since I'm not a Java developer it's sort of an "on the outside looking in" thing.

Sun developed the J2EE SDK, and released it to developers with the licensing requirements (and whatnot) fully disclosed. Lutris then comes along later and is upset that Sun won't rewrite their licensing procedures and open source their language interface just to suit them?

And the people here are actually upset about this?

How is Sun the bad guy for not giving away the sourcecode to their product (when they've never had any intention of doing so) just because some other company (I imagine the 'Good Guy') thinks they should?

Embrace Openness and Freedom....

This is the exact reason that I do not use Java anymore. The Sun Community Source License (SCSL) is a farce. Is it *really* open-source? Not in my opinion. It could be closed at anytime, anyplace, without notice, leaving the community behind. It definitely *not* GPL compatible(see http://www.gnu.org/licenses/license-list.html#GPLI ncompatibleLicenses), so it is not promoting software freedom. So, why does not Sun leave Java as proprietary and closed-source? It is a hell of a lot better than upsetting a bunch of dedicated developers. If the "next big thing" is web services, Java does not have to be the only alternative to Microsoft .NET. There are others out there....

Lutris is saving the face on this

Lutris is using the license of the new J2EE draft as an excuse to close the OS initiative on the Enhydra project. The fact is, that license is only a DRAFT and not DEFINITIVE. Older J2EE especifications have different and more permisive licenses, it may be to prevent implementations on a not-yet-approved spec.

Even if this was true, Lutris hasn't ever tried to solve the problem. The attitude of "well, we aren't gonna keep on this, but it's not our fault, blame Sun" is not very clean.

If I had contributed to the OS part of a product that is now going to be closed up by Lutris, I would just be pissed.

ESR on the WTC Attack

Boneheaded, opportunistic comment of the day [salon.com]. Last week Jerry Falwell blamed the WTC attack on the ACLU, feminists, and gays. Here's what ESR has to say about it;

Raymond, the libertarian open-source guru, known for his love of firearms, suggested that if the passengers of the hijacked jets had had guns the four-plane tragedy might have been prevented: "We have learned today that trying to keep civilian weapons out of airplanes and other areas vulnerable to terrorist attack is not the answer either -- indeed, it is arguable that the lawmakers who disarmed all the non-terrorists on those four airplanes, leaving them no chance to stop the hijackers, bear part of the moral responsibility for this catastrophe."

The story about this took less than five minutes to be rejected by the editors. Apparently when your stock is circling the drain, a member of the Board of Directors saying something like that isn't something you necessarily want publicised.

Think air rage is bad now? Try arming those drunk businessmen and see what happens.

I worked at Lutris

I worked for Lutris not too long ago, in
their Enhydra side (they also have a service
oriented side to their business - building commercial websites etc)
They were excited about the use of Open Source, but
I could tell then that they didn't fully understand it, and they had troubles figuring out how to keep an open source Enhydra, even though they truly meant to keep it open source. The J2EE issue was the biggest obstacle they faced, but they never thought they would have to give up open source. I worked for a guy there that used to work for Sun, and had even worked on the Java language, so he knew what kind of a company Sun is... and it just turns out that they even if they made a J2EE product thats fully J2EE compliant, they wouldn't
be able to market it as such without an expensive license from Sun, and that was a big difficulty they were having. Now the J2EE problem appears to have "resolved" itself by elliminating the open source side of Enhydra, which is SAD!!!! But, like I said, they seemed to have a mildly slippery grasp of what open source is about.

Lutris is also having financial problems (who isnt?) which is another sad thing. They are a great company. It's sad that they had to give up the open source Enhydra.

More on the Lutris Situation

I have worked with Lutris in the past with both Enhydra (version 2.3-3.0 and spec'd migration to the J2EE framework in enterprise) and there is more to all of this than just licenses from Sun and the J2EE framework. Don't get me wrong, what has been pointed out previously about the reference implementation and redeployment with different license terms would be an issue, but its more of a business trying to remain afloat.

Lutris was a consulting services company to start with. Enhydra was developed by bringing together a lot of what they used to develop and deploy customer web applications in previous projects. Since they were a consulting services company first, an open source process served to both (marginally) push forward the development of the applications server with public support, but also create a low barrier to adoption for companies to get the services process in the door. I was a software director at one of those companies that adopted the process and then moved to bring in the consulting side to deploy a very large application on.

Things were going great there when the economy was going great -- consulting services paid all the bills for the engineering crew to continue the primary development of the app server. The problem is when the economy turned south, the first thing to be cut were the consulting groups. Lutris had their contracts drying up and couldn't continue to pay the bills that way. Pretty soon they were left with a model that wouldn't work in an economy without a lot of free cash. There had to be another way to generate revenue or to go out of business. That model had to concentrate on traditional software development and open source companies haven't weathered that storm very well when there were commercial or other products that had more functionality or more entrenched customer bases. The quickest way to catch up was to push the enhydra enterprise process, use as much as possible to get it to a finished state (Sun ref implementation) and try to pull in product revenue with traditional sales. This couldn't be rectified with the open source licenses they were previously working on.

It's economics. Sure Sun's license for using their implementation of things is going to effect that, but its an after the fact reason. The underlying problem is that a consulting services company with no contracts isn't going to stay in business... A software company at least has a fighting chance.

I had friends that work(ed) there and this is not necessarily what they wanted out of things, but the survival instinct can be a powerful one. Has the discovery channel taught us nothing?

Who needs Enhydra?

With things like Resin (very fast servlet runner), Tomcat, jboss, Jonas, and OpenEJB, why would we care about Enhydra? It was always kind of a bizarre product anyway, with one of the lamest templating languages I've seen. Open source Java is alive and well.

Zope Enterprise is still open

I'd go for Zope... has the better feelgood factor.

SCSL is a commercial virus

I don't think the issue is Lutris. Lutris probably tried to get the license and failed; after all, they'd be in a much better position if they were open-source and J2EE; that would be a completely unique market position. Per Sun, if you read the J2EE spec, you agree to do nothing unless you license J2EE from them. SCSL is the only J2EE license, and Sun won't let you sign it unless they're sure your open-source (good) won't conflict with theirs (bad). Sun is the only game in town because to interoperate with them, you've got to use their API's, under license. Get the gospel? With the JCP, they lock up all the priests. So we're back to the cathedral, with Sun in charge of gospel and clergy alike. Sun squelches competition in enterprise Java software by reaching backwards with the JCP process and viral licenses that require more licensing to deploy and forwards with SCSL's trademark-based scheme (can't deploy without the trademark; need the license to get the trademark). You want viral? Check out the J2EE spec license, JAXP 1.1, heck, even JavaHelp - they all say you can't implement javax.* namespace without getting a license from Sun. The only exceptions are the special Apache licenses which are designed to get us all onto java based two-tier systems. But Sun would not let lutris do this. Why? My bet is that Lutris had the open-source BUSINESS MODEL of giving it away and making money off services and add-on's. That goes completely against Sun's license-based scheme for using app-server vendors to reach into the pockets of developers. I doubt any vendor with netscape's razors-and-blades model would get a license from Sun, because that would hurt Sun and the other licensees too much. Sun prefers value to be in IP that can be licensed (and contained); they can't reasonable reach into services and other revenues. Licensing law is worse that copyright and patent in this regard; at least for those, there are objective standards. Here, Sun can impose any standard it wishes in its "licenses", even if the only thing it's licensing is the opportunity to play the game. At least when MS was setting de facto standards, they had the burden of delivering a quality implementation that beat the others in the marketplace first.

Enhydra is full of....

I would really like to see an actual explaination of how the J2EE license prevents then for going open source on their code. Particualrly intersting to note is that Sun itself donated an open source app server to Apache ("Tomcat").

IMO Enhydra has decided that they can't make money in an open source model and are tyring to blame Sun in order to avoid the PR backlash.

philosophical differences

Enhydra first came to my attention almost a year ago when the company I was working for at the time was looking at JSP vs. XML for wireless and web presentation development. The folks at Lutris made no secret of their disdain for JSP and J2EE technology generally and their preference for their proprietary XML generating technology for web applications. There's an article about it at IBM's developerworks website at Objects, objects everywhere [ibm.com] and another even more relevant one at JSP technology -- friend or foe? [ibm.com]

At this point, Lutris has too much ground to make up against the J2EE server leaders, and no one is jumping onto their proprietary XML binding bandwagon, so Enhydra needs a way to distinguish itself from other Java application servers to get attention. My own evaluation of Enhydra gave me serious reservations with its architecture, including some issues around its scalability. Pointing to Sun and crying foul over the J2EE licensing issues (and that's all it is: an spat over whether or not their product can have the official J2EE compliant label or not), is just poor form.

Open Source J2EE AS Reviews?

Does anyone know of any reviews of the current leading open source platforms?
The only link I could source the explicitly mentions JBoss is CSIRO Australia report [csiro.au].
Suggestions?

Try Zope...

Probably the only application server you can have up and running in 2 minutes. Plus since its all python based, you can get tracebacks and debug the thing live. Lots of cool new method-specific caching features all built in. DAs for lots of different DBs, transactions etc. Solid community. Check it out! [zope.org] Ya might like it.

Seems like an easy fix

Lutris used Sun's reference implementation, hence the problems... couldn't they switch in an open source impl of that component (e.g. JBoss's) for an OS release?

IMO this needs saying....

BS like this only flies because people who know nothing shoot off their mouths like they do, and others belive them because its easier then doing your own research.

The SCSL and the J2EE licenses are totally seperate and dtstinct legal documents and thus seperate and distinct issues.

If you'ld like to actually learn something about them, based on the posts I've seen here I wouldn't try slashdot. I

Instead try the actual pointers contained in the post referenced below:

http://slashdot.org/comments.pl?sid=21699&thresh ol d=-1&commentsort=1&mode=thread&pid=2312751#2312969

Re:Java is dead anyway

No way. Java is more prevalent than it has been ever before; cell phones, pagers and now cars rely on it more than ever. Java isn't going anywhere, you silly troll.

Re:Oh really?

call SAS

You mean the Scandinavian Air Lines [scandinavian.net]?

Well, they're about to go under so perhaps you could convince them to drop a few planes on Osama and claim the insurance instead...

Re:Newsforge

Actually when I hear Newsforge I think of a forge -- used of old to create iron and steal in the heat of the forge... you get the idea.

Foregery may have a similar root but the use of Forge here seems pretty strong -- a good name if ever I've seen one.

Re:Why not provided two versions ?

The risk is too high that somebody like Microsoft with a huge blind user base will jump on that one and ship an implementation that behaves differently than the reference one from Sun. This would split and kill Java right away.

Re:Something Is wrong

There is something inherently wrong with a tool maker restricting what you can and cant do with your own code.

The problem is that Lutris used the Sun reference code to build the J2EE Enhydra server. It would be kind of like Mono releasing a GPL'ed .NET based on Microsoft's reference code submitted to ECMA.

Sun, on the other hand, has no problem with JBoss, which is a clean-room implementation of the J2EE spec.

Re:economics stupid

Well, I really think the situation now is quite self-inflicted.

The companies that now sees their stocks go through the floor just DOESN'T MAKE MONEY. And if you don't make money you can't justify that others should by your stocks for lots of $$$. And in the long run you can't pay your salaries and bills.

Companies needs to make money, it's as simple as that.

sig: for your daily dose of fanatism visit www.taliban.com, www.hamas.org or www.gnu.org

Re:Burned by Enhydra Before...

Cloudscape is $99 for a deployment license, too, according to Informix's press release for Cloudscape 3.6 (2/20/2001). So you're not saving any money by switching to Cloudscape, unless you never deploy it, or if you plan to deploy anyway in violation of Informix's license.

Anyway, Lutris never did release the source of InstantDB, and they never made an official statement any stronger than a vague intention to release the source of InstantDB. They own InstantDB, and they have the right to do with it what makes sense for their business.

What some people overlook is that companies do own their software, they don't sell it to you, they license it to you for specific uses, with conditions. This is true of MS Windows, and Java, and InstantDB, and almost everything -- including most open source software!

Re:Sad to see

sadly, I think no one from outside Lutris ever contributed any code to the enhydra enterprise project.