incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alex Karasulu" <aok...@bellsouth.net>
Subject RE: [PROPOSAL] PMC Vote to incubate Directory Project
Date Mon, 22 Sep 2003 07:22:49 GMT
Phil,

The LDAPd server in its present state could eventually support X.500 over
TCP/IP.  In fact both X.500 and LDAP seem to be coming closer every day
since X.500 made the jump to using TCP/IP.  I think the two will eventually
come back together.  For the time being when we speak about a directory
project we could presume any searchable hierarchical namespace to be a form
of directory.  Let's not limit our selves to just LDAP.

As you pointed out there will be more to this endeavor as a consequence of
JNDI.  It's very natural to associate JDBC related source with the
db.apache.org TLP.  Likewise a similar association exists between JNDI and
the directory project side whatever it may be called.

And again as you mentioned the potential for other applications on top of
the directory services are possible.  An identity management system project
is a natural progression.  As a matter of fact I founded the LDAPd Group
specifically for this purpose.  Necessity is the mother of invention right?
Other examples of potential subprojects could be an ASN.1 runtime library
and its associate stub compiler.  LDAPd already has ASN.1 runtime clone of
SNACC called 'snickers' - its just missing a compiler.  A meta-directory, a
virtual directory and a JNDI LDAP provider are all potential subprojects.

The embeddable nature of the server brings about incredible synergy between
it and other Apache projects.  Integration effort will probably be
facilitated by the directory people however obviously the end result will
reside within its own sphere.  Take for example the potential for white
pages backed by LDAP in JAMES.  Integration could serve as the basis for
Geronimo's security subsystem.  Integration with Struts and an identity
management system built on top of the directory server could allow for
authorization driven rendering.  There's also UDDI backed by LDAP.  All
these synergies are very attractive and integration efforts could bring
about collaboration with many projects however the end result will rest
within the respective domain.



Roy,

Other synergies certainly exist outside of an embeddable configuration.
Obviously certain language barriers are overcome by the protocol.  The httpd
mod_ldap DSO module can use the directory server.  Other plugins can be
added to the list like a mod_ldapconf module that pulls Apache virtual host
configuration information from a directory.  ISP's and ASP's would really
love this one. Several possibilities exist and I'll stop here before I bore
you.

The space is huge and the probability of budding projects and some in other
languages is certainly high.  Language will be transcended and if API's need
to be carved out for other languages they probably will out of necessity.
The RFC based protocol makes the server's implementation irrelevant.  Any
client should be able to talk to it.  Plus I'm sure with an Avalon.Net
quickly emerging someone with an itch might toy with bringing it over to C#
with a JNDI analog in that realm.

Naming is so important to directories so I guess I'll put in my $0.02 about
what the project could be called.  Directory sounds good and so does
directory services.  It's general enough to allow the full gambit of
coverage for the related spin-offs.  As a side note, just for the name of
the LDAP server (not the overall project) you probably couldn't use AD for
the Apache Directory I would imagine ;-).  A few months ago while working on
the new architecture for LDAPd I looked at AD in Application Mode or ADAM
(a.k.a. AD/AM) while still in beta.  It suddenly dawned upon me that I had
to code name the new LDAP daemon architecture 'Eve' which would be analogous
to Tomcat's 'Catalina'.  I have a soft spot for her ;-).  Folks me know what
you think about 'Eve' in particular.  

Cheers,
Alex

> -----Original Message-----
> From: Phil Steitz [mailto:phil@steitz.com]
> Sent: Monday, September 22, 2003 1:56 AM
> To: general@incubator.apache.org
> Subject: Re: [PROPOSAL] PMC Vote to incubate Directory Project
> 
> Roy T. Fielding wrote:
> >>> Greg posted a message back on the 18th noting that a PMC vote on the
> >>> entry of the project to the incubator would be kicked off under the
> >>> private pmc@incubator list.  I don't know the specifics of Incubator
> >>> voting policies but I guessing we will see a vote result early next
> >>> week.
> >>
> >>
> >> I saw that.  Nonetheless, I still find it odd that there has been
> >> virtually no "public" discussion of the merits of the proposal and I
> >> felt the need to express my opinion.  No disrespect for the process,
> >> de facto or documented, intended.
> >
> >
> > Hmm, in general, the only discussions we ever hold in private are those
> > relating to non-disclosure issues or personnel issues.  We would have a
> > discussion and vote about the directory project on the general list.
> > The reason we haven't is because it is being suggested to incubator
> itself
> > rather than some other project or the board, which means it is waiting
> for
> > us to arrange for a new chair.
> >
> > FTR, there is no such vote taking place in private.
> 
> Thanks for explaining this.
> 
> >
> > Personally, I think it is an excellent proposal except for one item:
> > Java is not the center of the universe.  Either this project should
> > consider all languages (seems unlikely, unless you already have those
> > other people in place), or the project should be named after the
> > product name and not the whole category of software.  A name like
> > "jeldap" (I've only seen that word used in relation to turbine)
> > would get you past a number of hurdles.
> 
> After re-reading it, I agree that the proposal does emphasize the Java
> aspects a bit more than the "Apache Directory" name would suggest.
> There are three reasons for this: first, the initial code base is Java,
> second JNDI and some of the extended APIs that the project will support
> are Java APIs and third there is no small measure of delight being
> expressed over the fact that recent JDK advances make a high-performance
> Java-based LDAP server thinkable :-))
> 
> I would personally have no problem with changing the name to reflect the
> Java/JNDI emphasis. While I would personally favor a nice animal (e.g.
> or culturally neutral place name, I could also support a JAMES-like
> moniker that brings the "J" out.  I don't much like "jeldap" or "JLDAP"
> or anything else based on LDAP by itself, because I expect the naming
> and identity services to emerge as equally important parts of the
> project, enabled by the core directory services.  The only ideas that I
> can come up with immediatedly are JANIS (Java Apache Naming and Identity
> Server), which is a bit extreme (not mentioning LDAP at all) or JED
> (Java Enterprise Directory), which doesn't say much. Any better ideas?
> 
> Thanks for the feedback.
> 
> Phil
> >
> > ....Roy
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> > For additional commands, e-mail: general-help@incubator.apache.org
> >
> 
> 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
For additional commands, e-mail: general-help@incubator.apache.org


Mime
View raw message