PAW created/edited TI processes are backward compatible with Architect

I used Planning Analytics Workspace to make a change in a TI that used a cube datasource.  I did a save as, so I could ask a co-worker about a change and not overwrite the original.   My co-worker was using architect when he opened the TI and made a small change and saved, then he was prompted saying the process was not valid.   On the variable tab in architect all the data types were set to ignore.

After some research we found the following differences with PAW TI processes that are created or edited when architect was later used:

1. On new data source columns - Variable names and variables types are not transferred to architect.  If the variables are updated in architect, they show correctly in PAW.  
2. On changed data source columns - Changes in data types or names of variables are not transferred back to architect, but values originally set in architect are still there, and if saved will reset PAW changes.
3. PAW does not create the "#‌Begin Auto Generated Code" and the "#‌End Auto Generated Code" on MetaData and Data tab/section.   If you put code on lines 1 and 2 and save it in PAW, then open it in architect, it will add the auto generated comments and wipe out any code on lines 1 and 2.

Workarounds: If you build a process in PAW, at a later point edit it with architect and manually set values.  If you need to change variables/types do it in architect.

We have people on the team with different levels of comfort with PAW vs Architect,  allowing people to use either tool for support is of value.  When PAW is down allowing support to continue using architect is another benefit.