5 Replies Latest reply on Jun 13, 2014 9:17 AM by rareddy

    Refreash Teiid Designer?

    jason.marley

      Hi All,

       

      I've been using Designer for a few years and I know there is a lot of stuff that goes on behind the scenes as far as getting everything to tie together well with Teiid Server, but as a general user I don't really get to see that richness through the UI. As a general user, I often find myself debating whether I should be doing something in Java or SQL, even though I've never done anything in Java in Designer mode; I think I get this vibe because of eclipse and its tooling.

       

      I don't think I've ever used Designer for more than anything I would do in mysql, jaspersoft,  mssql, etc. workbench clients. The one thing I think would possibly be helpful in eclipse is being able to share a project between developers, but I know that it is difficult to do and we are having a hard time finding a good way forward. I'm wondering though if we rolled our own client whether we could incorporate the best of all worlds. It seems to me that DBA's and the like (at least the ones I've been in contact with) have a difficult time with the tooling and I'm not sure whether it is because most of them aren't familiar with eclipse or whether the tooling is too different from common SQL workbench clients?

       

      To that end, I'm not familiar with why eclipse was chosen to be the front end and don't mean to be disparaging, but is it necessary going forward?

       

      Jason

        • 1. Re: Refreash Teiid Designer?
          rareddy

          There may be be other things that needs to be fixed, but eclipse. Eclipse is our base GUI framework everything else is written on. So that will not change. Even having own client that will not be big improvement over perception.  We did at one time had Designer as single plugin in the entire eclipse to give apperence as custom client.

           

          I do agree that, we need to make more efforts in GUI based SQL building as other tools provide, currently that is lacking main transformation area. Designer does fairly good job with type resolver and functions. Need to extend that to SQL based on context it is invoked in, to bring in capabilities as other SQL centric tools.

           

          Model sharing across team is something we actively looking for solutions. If you have concrete suggestions as to how UI should look or can be updated please do open up a discussion or JIRA.

           

          Thanks

           

          Ramesh..

          • 2. Re: Refreash Teiid Designer?
            blafond

            Jason,

             

            Great questions ... and we've been asking ourselves some of the same questions.

             

            Teiid Designer is a legacy Eclipse-based tool (MetaMatrix) that fit in nicely with JBoss Tools and the RH JBDS framework.  Choosing a tooling workbench other than Eclipse would have made it harder to integrate with other JBoss tooling (i.e. server functionality) at the time. So that was the basis for staying with Eclipse.  Another reason Eclipse is still use is our dependency on it's EMF framework that defines and manages the underlying business objects (metamodel).   Replacing this metamodel with non-EMF BO's would not be an easy exercise.

             

            The primary purpose of Teiid Designer is to provide a robust/flexible environment for creating and managing the primary Teiid artifact .... the VDB.

             

            I'm curious about your comment that DBA's either "aren't familiar with eclipse or whether the tooling is too different...".  Narrowing that down would be of great help moving forward because....

             

            We've started the groundwork for a new, more flexible/scableable Teiid Designer, currently code-named our future Komodo release.  Komodo development, which is starting to gear up this month, includes removing EMF as the fundamental metamodel and utilizing a common Teiid DDL-based metamodel. Both steps will allows us to create a headless VDB editing framework that will lighter weight and UI independent. One of our first steps will be to create a CLI editing interface for creating, editing and navigating a VDB as well as connecting to and executing against a server (current Preview VDB and Execute VDB modes).

             

            Though we'll be starting out providing an Eclipse-based UI for integration into JBoss Integration Stack for DB, one of our goals is to insure that our design-time engine is fully portable to other UI's as well, including web-based.

             

            You can find most of our Komodo discussions and ideas here: Teiid Designer - Komodo Release Development Plan

             

            Barry LaFond

            Teiid Designer Project

            • 3. Re: Refreash Teiid Designer?
              jason.marley

              I understand that eclipse is used to develop teiid designer and the logic for the metadata mapping, but it shouldn't be the "only" solution for the users going forward (IMO ). You know what I mean? MYSQL workbench doesn't have its users use Visual Studio in order to build database models and the like, but I'm sure they could.

               

              I confess, I'm not fully aware of all the nuances of the team sharing issues, but from what I've seen the issues happen when the data source connection profiles are named differently (even though the connection strings are the same), JBDS customizations in the metadata files and I've heard reorienting tables tweaks things as well; I'm sure there are others. If there is a list of the issues I'd be happy to provide my 2c.

              • 4. Re: Refreash Teiid Designer?
                jason.marley

                I understand about where Teiid Designer has been now, thank you. I haven't been a part of a project nearly this big or old, but am familiar with the weighing risks of re-factoring and it is never easy.

                 

                From the layman (myself), the Teiid Designer is connecting to a data source(s), pulls metadata about said sources, makes it available, allowing views to be created and metadata changed against those sources and then packages it up for deployment on the an application server so "things" can query those new views.

                 

                Now abstracting the fact that Teiid Designer does a lot more than just that, but from layman point of view it looks exactly what the other SQL/ETL workbench tools provide. I think it may be worth while to go through some of the SQL/ETL workbenches and see how easy (or not) it is to do those things above and lets see if we can do them better, because these are the tools our clients are used to using on a daily basis and they are bench-marking against Teiid Designer.

                 

                On that note, Komodo sounds brilliant! Can't wait to contribute.

                 

                Jason

                • 5. Re: Refreash Teiid Designer?
                  rareddy

                  Jason,

                  I understand that eclipse is used to develop teiid designer and the logic for the metadata mapping, but it shouldn't be the "only" solution for the users going forward (IMO ). You know what I mean? MYSQL workbench doesn't have its users use Visual Studio in order to build database models and the like, but I'm sure they could.

                  This part has nothing to do with the Designer. This is a feature we are planning for the 9.x version of the Teiid. Where you can fire up the sql work console start typing in DDL to define you VDB or alter it. Right now we only allow this during the deploy time, but we are planning to extend the metadata updates to runtime of the server.

                  I confess, I'm not fully aware of all the nuances of the team sharing issues, but from what I've seen the issues happen when the data source connection profiles are named differently (even though the connection strings are the same), JBDS customizations in the metadata files and I've heard reorienting tables tweaks things as well;

                  These are issues we need to fix/patch in more immediate future. I would like to hear more concrete example issues here that can be addressed. You can open either separate discussions or JIRAs on those subjects.