3 Replies Latest reply: Feb 9, 2017 5:45 AM by Georg Gottsmann RSS

    How to handle small, decentral QV-applications seperated from QDF?

    Georg Gottsmann

      Dear all,

       

      we have installed QDF on a server and it runs well. However, we often face the problem where this solution seems to be "oversized" and a local application on the user's computer would be sufficient. Another problem is that server cannot access this local data.

       

      Is there a best practice to use a QDF Container as an "offline container" for example which will be given to a local user? We want to keep the good structure and the possibility to just copy the Container to the Server in case the Application gets bigger. QDF is great, but it assumes all developers to work only on the server (which makes perfect sense). In our company unfortunately, we have only one member of the IT responsible for that and our strategy is to mostly enable users to maintain their own applications. Could you recommend a solution for this?

       

      Thank you in advance.

        • Re: How to handle small, decentral QV-applications seperated from QDF?
          Damian Waldron

          Hi Georg,

           

          I am trying to understand your process.

           

          * Are you indicating that users develop on their local machine. Changes then need to be migrated into the production environment

          * Only the IT administrator can update production

          * The server then reloads and publishes the application from the server.

           

          Therefore is the problem lies around how the developers ensure that the correct artefacts are moved to production? I have seen the following as solutions

          * Manual change logs identify which files need to be moved.

          * Rather than manually tracking changes you could leverage SyncToy to compare files. SyncToy - Wikipedia

          * Finally you could implement a more robust solution using GitHub.

          1. Master Container is stored in GitHub.

          2. Developer checks out's repository. makes changes and then pushes the changes

          3. IT Admin pulls updated repository from github

           

          Using Git you would want to use gitignore to only sync code files. I.e. ignore QVDs and data files.

           

          I hope this is useful.

          • Re: How to handle small, decentral QV-applications seperated from QDF?
            Magnus Berg

            Hi Georg, As Damian stated versioning is a good way of controlling local development. You can also create "self-service" containers, one per user (usually under a sub folder). See to it that the users have access to the the personal central container, now they can copy the container to the laptop start develop and copy the container back up again replacing the old.

            Regards

            Magnus