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

Re: Fwd: XML Server



Hi Falko,

I fully agree with what you're quoted (except the part about  OODBMS
which I do not know enough).

Just one question, why do you say that

> he also pointed out that something like "THE XML server" will 
> never be developed.

when on the contrary Ismael says 

> > If these guys did that, I see absolutely no reason to think that the 
> > XML repository will not be the next back-end of web-aware IT infrastructures.

which is the idea which made me start the list about our XML server
dream ?

Eric

Falko Braeutigam wrote:
> 
> On Tue, 26 Oct 1999, Falko Braeutigam wrote:
> > On Tue, 26 Oct 1999, Hannes Haug wrote:
> > > Falko Braeutigam wrote:
> > > >
> > > > But there is a small"but". SoftwareAG, a german database company, has been
> > > > providing its database software for 20+ years I think. So they seem to be very
> > > > experienced with database development. They never started an OODBMS - but the
> > > > _have_ a XML server called tamino. (somewhere at http://www.softwareag.com)
> > > > So SQL is indeed the wrong choice but maybe not the whole relational thing.
> > >
> > > Tamino is at http://www.softwareag.com/tamino
> > > Software AG doesn't have an OODBMS, but their "real" product,
> > > Adabas (http://www.softwareag.com/adabas), is a hirarchical
> > > database not a relational one.
> > Hmm... IMO I did neither say that adabas is an oodbms nor that tamino is based
> > on an RDBMS (see Erics mail). All I wanted to say is that there are other ways
> > to efficently store XML besides OODBMSs (ozone).
> Sorry if this was not that clear. To clarify this a bit more I quote from a
> discussion with Ismael of Exoffice. He is interested in ozoneXML but he also
> pointed out that something like "THE XML server" will never be developed. We
> should try to identify the application domains where we want to use ozone(XML)
> and then optimize it to fit exactly the needs of these applications.
> 
> Ismael:
> > Most of my confidence in this very specific field comes from following what
> > does a German company: Software AG. This very old independent software
> > vendor (30 years old) has a very long history in the database field. They
> > started to make a very powerful hierarchical database engine, and then made
> > the most powerful relational then object-relational database engine: Adabas
> > D. Adabas D has been for a long time the only database natively supported by
> > SAP R/3 with Oracle. And more interesting: Adabas D has the unique ability
> > to store the data in a way that makes useless its regular re-organization.
> > To make it simple: these guys are quite good developers :-) The interesting
> > point is that these guys have never been interested by developing an ODBMS.
> > Not that it was not fun. I just like to think they knew from their own
> > experience that it was quite impossible to make a scalable ODBMS which could
> > be used in a mission-critical production environment supporting distributed
> > transactions in a very secure way. And they were right... But at the same
> > time, these guys, for any kind of reason, decided to take the core engine of
> > Adabas D, to make some serious hacks on it, and to use it as the core engine
> > of the X-Machine, the core engine of Tamino, there brand new XML repository
> > [http://www.softwareag.com/tamino/default.htm]. If these guys did that, I
> > see absolutely no reason to think that the XML repository will not be the
> > next back-end of web-aware IT infrastructures.
> ...
> > My wish would be to get Ozone to be the PostgreSQL of XML repositories...
> > And the work you have done with Vincent and Ann make me think that this
> > could become a reality quite soon. Doing so, developers will get the ability
> > to develop there application using XML as a core way to model their data and
> > deploy it using Open Source software only. It might not have the scalability
> > and the fault-tolerance of products like Tamino or the XML repository I
> > expect Oracle to deliver in the future, but if they need this kind of
> > high-end features, they will certainly have the money to buy these products.
> > In the meantime, the availability of Ozone and extensions like CDQA, XOQL,
> > the XPath and XQL front-ends, and MonsterDOM will dramatically ease this
> > kind of ideas to spread into the community of web-applications developers,
> > which is IMHO the most important thing for most of us.
> 
> I don't agree with him about OODBMSs in general. But I agree that a SQL/C++
> based solution, as well as ozoneXML, as well as any other solution, will not be
> the "ideal" XML server. (but ozoneXML will be "more ideal" than SQL;)
> 
> This idea was not clear from my initial posting, sorry again.
> 
> Falko
> --
> ______________________________________________________________________
> Falko Braeutigam                         mailto:falko@softwarebuero.de
> softwarebuero m&b (SMB)                    http://www.softwarebuero.de

-- 
------------------------------------------------------------------------
Eric van der Vlist                                              Dyomedea

http://www.dyomedea.com                          http://www.ducotede.com
------------------------------------------------------------------------