[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: JDO is dead (was: Re: *URGENT* -- Call for support)



On Fri, 19 Jan 2001, Eric Samson wrote:
> Hi Carl Falko and all,
> 
> I think that you interpreted my mail like an offence
> it wasn't my purpose, sorry I you feel it like that.

No, of course we did not interpreted anything as an offence. Sorry, if you got
that feeling!

The background is that we, Carl and me, discussed so many times about ODMG,
JDO, the OODBMS market and all that stuff that we came to more or less clear
conclusions (at least for us). You are right, we did throw those conclusions in
the dicussion somewhat 'uncommented'.

Again, I'm open minded for all ideas. And if we all decide that JDO is
important for ozone, then I will support this.


Falko

> 
> I just want to clarify few things :
> 
> * ODBMS is a very tough and thin market, for both OpenSource and commercial
> products
> * at least 10 to 15 years are necessary to impose a new technology like
> ODBMS
> * things are changing slightly now, owing to Java and EJB
> * JDO is a real chance for ODBMS, because
> 	* ODMG was only a standard for ODBMS
> 	* JDBC is only a standard for RDBMS
> 	* JDO addresses the persistence in a different way, for any data source
> 	* it's the best way to educate the market about transparent persistence
> 	* Oracle won't support JDO soon because it's a real danger for them
> 	* but some OpenSOurce initiatives will provide Oracle JDO drivers
> 	* see castor.exolab.org
> 	* JDO also addresses the J2EE perspective
> 	* few days ago I was still a Versant employee and I can tell you that JuDO
> is not burried
> 	* I'm also a new member of the JDO expert group and I can tell you that JDO
> is not dying
> * EJB are junk : current Entity Beans specs suck, you're right, but the
> whole J2EE concept is good. You can imagine in the future, session beans
> manipulating pure Java objects, transparently stored in an ODBMS, or even
> better, evolutions of the Entity Bean model to fully support complex and
> rich object models through a complete JDO persistence. All EJB scalability
> issues are very nice for ODBMS, because access by navigation is faster than
> access by query. The current spec is a perfect explanation of why RDBMS are
> not able to cope with Internet.
> * Quality : I agree with you, but I'm convinced that OpenSource participate
> to quality
> 
> * Smalltalk is the best language : this is absolutely true. Microsoft should
> participate in making Java the perfect environment instead of creating a
> proprietary alternative, with no added-value, in a destructive spirit.
> * I still think that finding contributors for core database technologies is
> not so easy. You say the opposite but it doesn't solve the current Falko's
> issue which seems to confirm my point of view
> 
> Why don't you merge all your separate initiatives in a bigger OpenSource
> ODBMS project ?
> Why not trying to insert Ozone in a bigger OpenSource project (Apache, ...),

this is not that easy. apache fights against internal problems which prevent
their organisation from scaling as needed.

On the other hand ozone does not fit into apache. they are doing web stuff.
ozone is a general OODBMS. 

> as Enhydra recently did with InstantDB ?
> Is there no way to get european funds to develop and market the product ?
Get money for open source development? Sure, then you have to get money out of
service, which means that you can just put 20% (at max) of your time into OS
development, which in-turn means that all participating parties must help with
the OS development. This indeed is our core problem right now ;)


Falko

> 
> Best Regards,
> Eric Samson
> LIBeLIS, Liberty for Large IS
> President
> +33 6 0302 5341
> Eric@libelis.com <mailto:Eric@libelis.com>
> 
> 
> -----Message d'origine-----
> De : ozone-users-owner@ozone-db.org
> [mailto:ozone-users-owner@ozone-db.org]De la part de Carl Rosenberger
> Envoyé : jeudi 18 janvier 2001 23:46
> À : ozone-users@ozone-db.org
> Objet : JDO is dead (was: Re: *URGENT* -- Call for support)
> 
> 
> ----- Original Message -----
> From: "Eric Samson" <eric@libelis.com>
> 
> Hi Eric,
> 
> I would like to comment on some of your statements.
> Maybe my opinion could also be interesting for others in this group.
> 
> A comment to *URGENT*:
> Although I am a competitor, Falko and I are mailing on a regular basis.
> I habe been visiting Falko in Leipzig and he really is a very nice guy.
> He deserves all the support you can give.
> I sent him a private sympathy mail an hour ago.
> 
> There is no other open source Java OODB of importance besides ozone.
> XML support is of great importance today.
> Please help to continue the ozone project.
> 
> 
> To your mail:
> 
> > * but real database kernel skills is not something so common
> 
> I don't think so.
> Domain knowledge is the main problem in programming.
> Working on databases is fun, smooth and easy because the functionality can
> be very well defined.
> 
> 
> > * suppose that Open Source intiatives can take 10 % of a market
> > * and today ODBMS is <2% of the whole DBMS market
> > * => so the target for an Open Source ODBMS is no more than <0,2 % which
> is
> > not huge
> 
> Mixed calculations like this never make sense.
> An excellent product defines it's own market.
> 
> A VC did a similar calculation for my project:
> 1% mobile devices run Java
> 50% have an application that need a database
> 10% would use an OODB
> Market: 115.750 EUR
> 
> What a joke!
> I have a request for 2 million licences at 1$ and that's only 1 request
> among 30.
> 
> Marketing is what ozone needs.
> Specifically the XML area can be pushed.
> This is the trouble with Open Source:
> You typically don't have marketing experts that try to get hype going.
> 
> 
> > * analysts says that some technos are ready for Open Source, while others
> > are not
> 
> The question is not "Open Source or Closed Source", the question is one of
> the quality of a product.
> Quality can only be reached with work, work and work again.
> Open Source and Closed Source only define, who works on a project and who
> receives the rewards.
> 
> I see a problem for Open Source since programmers can get so much money for
> their work that it seems stupid to work for free.
> Could it be an option to *pay* Falko for his work from the ozone community?
> It would be very cheap, looking at the work he does.
> I think there are 200 people on the mailing-list, is that right?
> I suggest to invest 1000$ each for Falko to keep the work going.
> 
> 
> > * I'm quite convinced that ODBMS market is having a new birthday owing to
> > Java and J2EE
> 
> I believe that all technologies too specific to Java might get a problem.
> C# is coming strong and MS VS .NET will be a beautiful development
> environment.
> Normal Java code can easily be ported to C#.
> The more complex specific APIs are, the more proprietary to the Java
> platform they will remain.
> 
> Smalltalk clearly is the superiour language to Java and to C#.
> 
> 
> > * EJB market is a huge potential for ODBMS
> 
> EJB are junk.
> A proxy mess like that has no right for a future.
> Even some Sun- and Java-fans have a similar opinion.
> 
> 
> > * there is NO "xml storage market" (see Excelon troubles these days) !!!!
> 
> Excelon got famous by becoming the market leader with their terrible object
> database ObjectStore.
> They ruined the market for everyone.
> 
> I would not expect their XML product to be that good.
> 
> 
> > * xml is definitely not a storage format, it's only an exchange format
> 
> I fully agree here.
> It's only an exchange format for small documents.
> 
> 
> > * ODMG is dead ... (yes it is)
> 
> Yes.
> 
> > * ... long live to JDO !
> 
> JDO is dead.
> http://access1.sun.com/jdo/
> Even Versant now degrades it's "Judo" interface to a research project.
> 
> 
> Sorry for my harsh words, but I can't help my opinion.
> 
> 
> Kind regards,
> Carl
> ---
> Carl Rosenberger
> db4o - database for objects - http://www.db4o.com
-- 
______________________________________________________________________
Falko Braeutigam                              mailto:falko@smb-tec.com
SMB GmbH                                        http://www.smb-tec.com