You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tuscany.apache.org by Luciano Resende <lu...@gmail.com> on 2007/07/25 00:55:00 UTC
[LDAP DAS] Current API and EMF dependencies...
Hi Ole
After you mentioned you are almost done with a initial version of
LDAP DAS that provides CRUD operations [1] I went to your sandbox [2]
to take a quick look at the current implementation and have couple
questions :
- Where is the latest implementation of LDAP DAS available ? Is the
code you have in your sandbox current and the most update one ?
- Your sandbox looks like still using a lot of EMF dependencies. Do
you have any plans to base the LDAP DAS implementation on current
Tuscany SDO (SDO 2.1 specification implementation) ?
- The beauty of "Heterogenous DAS" is to have a consistent
programming model and a single set of APIs to access data from
heterogeneous data sources, such as RDB and LDAP. Looks like the
current implementation of LDAP DAS is using a different set of APIs
for it's implementation, thus introducing a new programming model and
a new set of APIs. What are your plans here ?
[1] http://www.mail-archive.com/tuscany-dev%40ws.apache.org/msg20553.html
[2] https://svn.apache.org/repos/asf/directory/sandbox/oersoy/das.ldap.parent/
--
Luciano Resende
Apache Tuscany Committer
http://people.apache.org/~lresende
http://lresende.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org
Re: [LDAP DAS] Current API and EMF dependencies...
Posted by Ole Ersoy <ol...@gmail.com>.
SNIP
> Good, I'll wait for your updates to take a further look. Please let me
> know when you committ these updates.
Cool - I'll check in and update in a few days.
SNIP
> I have no idea of when SDO 3.0 will be available, and we are not sure
> that we will have all the necessary APIs you think is missing on that
> release, right ?
Sure - I'm sure 3.0 vs. 2.1 will be a minor thing overall. Once I'm done with the design document, we can review and consider an approach for the gaps that we see between the APIs. There are something like 12 classes that use EMF, of which 6 are core classes, so I think solutions to make the LDAP DAS SDO 2.1 compliant can be implemented quickly.
SNIP
>
> We should also involve the SDO folks here, and get their input on this
> subject, but, having RDB DAS and LDAP DAS returning incompatible SDO
> would be a issue to really take in consideration when making the final
> decision here.
Sure - I think once we understand the gaps, we'll quickly come up with solutions, and then we can put them all in a road map, and start making the LDAP DAS work with the SDO Spec.
SNIP
> This is not really around SDO, but making usage of DAS Interfaces. I'd
> expect that the LDAP DASImpl [1] would implement a variation of
> Tuscany DAS Interface [2], and all other Tuscany DAS implementations
> would do the same (e.g RDB, XQuery, or any other that comes in the
> future). A similar interface is being used for DAS C++. This allows
> for a common programming model and api on the DAS level. Any plans for
> this ?
Yes - That's the plan.
>
> [1]
> https://svn.apache.org/repos/asf/directory/sandbox/oersoy/das.ldap.parent/das.ldap/src/main/java/org/apache/tuscany/das/ldap/impl/DASImpl.java
>
> [2]
> https://svn.apache.org/repos/asf/incubator/tuscany/sandbox/lresende/das/api/src/main/java/org/apache/tuscany/das/DAS.java
>
>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
>> For additional commands, e-mail: tuscany-dev-help@ws.apache.org
>>
>>
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org
Re: [LDAP DAS] Current API and EMF dependencies...
Posted by Luciano Resende <lu...@gmail.com>.
Ole
On 7/24/07, Ole Ersoy <ol...@gmail.com> wrote:
> Hey Luciano,
>
> Luciano Resende wrote:
> > Hi Ole
> >
> > After you mentioned you are almost done with a initial version of
> > LDAP DAS that provides CRUD operations [1] I went to your sandbox [2]
> > to take a quick look at the current implementation and have couple
> > questions :
> >
> > - Where is the latest implementation of LDAP DAS available ? Is the
> > code you have in your sandbox current and the most update one ?
>
> As soon as I have finished up the DAS interface code, commented everything, and removed left over scraps I'll do another check in the "Working" version. The sandbox is still the old stuff from when I first got reading and writing graphs working.
>
> >
Good, I'll wait for your updates to take a further look. Please let me
know when you committ these updates.
> > - Your sandbox looks like still using a lot of EMF dependencies. Do
> > you have any plans to base the LDAP DAS implementation on current
> > Tuscany SDO (SDO 2.1 specification implementation)?
>
> I think we might want to shoot for having it compliant with the SDO 3.0 spec. 2.1 seems to be missing some key API features such as something equivalent to getEIDAttributes() which returns the EAttribute where id is true. Also another thing that both EMF and SDO 2.1 are missing is getEAllCrossReferences (which is the opposite of getEAllContainmentReferences()...EMF does provide an implementation specific way of doing this though, although it's not part of the API). Having these in the SDO spec would give us a head start of having a consistent programming model across DASs.
> >
I have no idea of when SDO 3.0 will be available, and we are not sure
that we will have all the necessary APIs you think is missing on that
release, right ? Also, I think Tuscany SDO is trying to move off EMF
dependencies, but I guess the SDO team would be able to give a better
explanation of the directions here.
One thing I always had in mind with multiple DAS implementations, was
the scenario where you had your data stored in one type of data source
(e.g your person records would come from a HR Table), but later, you
start having that info coming from LDAP, and you would only have to
change your DAS Config files, to update connection information, and
command syntax. By having RDB DAS producing SDO 1.x (I guess this is
what EMF supports) and then having DAS RDB producing SDO 2.1, can we
still accomplish this ?
We should also involve the SDO folks here, and get their input on this
subject, but, having RDB DAS and LDAP DAS returning incompatible SDO
would be a issue to really take in consideration when making the final
decision here.
BTW, the gap between EMF and SDO 2.1, can they be workarounded in the
LDAP DAS implementation ?
> > - The beauty of "Heterogenous DAS" is to have a consistent
> > programming model and a single set of APIs to access data from
> > heterogeneous data sources, such as RDB and LDAP. Looks like the
> > current implementation of LDAP DAS is using a different set of APIs
> > for it's implementation, thus introducing a new programming model and
> > a new set of APIs. What are your plans here ?
>
> I totally agree with you. The LDAP DAS should work with the standard SDO API asap. So we need to identify the gaps between the EMF API parts used currently and the SDO API and bridge them. Meanwhile, those users needing a common interface for both the LDAP DAS and RDB DAS would have to use the EMF SDO implementation.
>
>
This is not really around SDO, but making usage of DAS Interfaces. I'd
expect that the LDAP DASImpl [1] would implement a variation of
Tuscany DAS Interface [2], and all other Tuscany DAS implementations
would do the same (e.g RDB, XQuery, or any other that comes in the
future). A similar interface is being used for DAS C++. This allows
for a common programming model and api on the DAS level. Any plans for
this ?
[1] https://svn.apache.org/repos/asf/directory/sandbox/oersoy/das.ldap.parent/das.ldap/src/main/java/org/apache/tuscany/das/ldap/impl/DASImpl.java
[2] https://svn.apache.org/repos/asf/incubator/tuscany/sandbox/lresende/das/api/src/main/java/org/apache/tuscany/das/DAS.java
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
> For additional commands, e-mail: tuscany-dev-help@ws.apache.org
>
>
--
Luciano Resende
Apache Tuscany Committer
http://people.apache.org/~lresende
http://lresende.blogspot.com/
---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org
Re: [LDAP DAS] Current API and EMF dependencies...
Posted by Ole Ersoy <ol...@gmail.com>.
Hey Luciano,
Luciano Resende wrote:
> Hi Ole
>
> After you mentioned you are almost done with a initial version of
> LDAP DAS that provides CRUD operations [1] I went to your sandbox [2]
> to take a quick look at the current implementation and have couple
> questions :
>
> - Where is the latest implementation of LDAP DAS available ? Is the
> code you have in your sandbox current and the most update one ?
As soon as I have finished up the DAS interface code, commented everything, and removed left over scraps I'll do another check in the "Working" version. The sandbox is still the old stuff from when I first got reading and writing graphs working.
>
> - Your sandbox looks like still using a lot of EMF dependencies. Do
> you have any plans to base the LDAP DAS implementation on current
> Tuscany SDO (SDO 2.1 specification implementation)?
I think we might want to shoot for having it compliant with the SDO 3.0 spec. 2.1 seems to be missing some key API features such as something equivalent to getEIDAttributes() which returns the EAttribute where id is true. Also another thing that both EMF and SDO 2.1 are missing is getEAllCrossReferences (which is the opposite of getEAllContainmentReferences()...EMF does provide an implementation specific way of doing this though, although it's not part of the API). Having these in the SDO spec would give us a head start of having a consistent programming model across DASs.
>
> - The beauty of "Heterogenous DAS" is to have a consistent
> programming model and a single set of APIs to access data from
> heterogeneous data sources, such as RDB and LDAP. Looks like the
> current implementation of LDAP DAS is using a different set of APIs
> for it's implementation, thus introducing a new programming model and
> a new set of APIs. What are your plans here ?
I totally agree with you. The LDAP DAS should work with the standard SDO API asap. So we need to identify the gaps between the EMF API parts used currently and the SDO API and bridge them. Meanwhile, those users needing a common interface for both the LDAP DAS and RDB DAS would have to use the EMF SDO implementation.
---------------------------------------------------------------------
To unsubscribe, e-mail: tuscany-dev-unsubscribe@ws.apache.org
For additional commands, e-mail: tuscany-dev-help@ws.apache.org