What do you think @lenaatsafe and @dmitribagh?
Hi @bmdull
it is a brilliant idea! We are discussing it now. You would only need to update attribute values, am I correct? I.e. you don't need to alter geometries, right? If a geometry needs to be deleted, it can be simply marked - and the workspace will take care of actually deleting it from the main dataset.
I wonder how you imagine the password protected models will work. Would you prefer the user to type in the password before loading the model? Or would it make sense to have FME AR login - and once the model update is pushed back to the server, the app will also send user ID/name?
Regarding the workspace that should take care of the submitted changes: would you prefer the workspace to be a Server App with URL added to the model? This would allow having different models with different Server Apps responsible for processing. Another option would be to have the Server App URL as FME AR setting with an easy switching between different URLs.
These are very rough ideas for now. If you have a different vision or have a list of requirements in mind - please share with us.
Your input is much appreciated.
Hi @lenaatsafe. Thank you for responding.
I would preferably like to update both attribute values and geometries. After the features are collected, it's not uncommon for our field crew to go out and collect what we call redlines. During the redline process, the field technician might move a splice enclosure from one pole to the other across the street, delete it all together, create a new feature, or change the attachment height attribute value from what was previously collected. So, what I have in mind is for the features and attribute values to be field verified and updated within the AR model, which then gets validated in our PostGIS database. In effect, this would operate similar to a field collection app, like Arc Collector, especially if we can add feature creation based on its type. For instance, creating new pole features if they weren't previously collected. It would also be great to be able to add additional attributes to each model feature, including: change type (create, delete, modify geometry, and modify attribution), username and date edited, etc in addition to the attributes already associated with each feature type based on the PostGIS table's schema.
Regarding password protected models, I like the idea of inputting a password prior to being able to download the model. Alternatively, this can be done within a Server app such that the user is able to submit a job to download the FME AR model if they have the proper credentials.
Regarding the reconciliation workspace, I envision filtering the changes by type. Those typed as modified would can use a WHERE clause of "id" = "id". Features that are either create or delete change typed would output into a temporary PostGIS table for someone on the design team to filter through and validate the changes. When these changes are validated, the designer can run another a Server app to quickly push those changes to the actual PostGIS database tables.
Hi @bmdull,
this all sounds great! I have one concern re adding geometries in AR mode. In general, in AR sometimes it is hard to tell whether an object in the model is higher or further away. Occlusions help a lot with interpreting the spatial relations between different objects. Textures, shadows, lighting options, etc. are also helpful. However, adding a new geometry in 2D might still be easier or more intuitive. I.e. one of the options would be to have 2D view to add new geometries. We will also look into adding features in AR mode itself.
Thank you for providing more details about your scenario. It is very important for us to understand users' workflows and tasks to make sure functionality we are going to implement will meet users' requirements.
