three tier development qliksense

[Thiarez]

I have worked on Qlikview and we were using the three tier development architecture in our previous project.
Now we are going to start on a Qliksense project and I need help on that.
What architectural approach  should we follow? Should we continue with the same approach as qlikview(Creating Qvd's, Creating model and finally creating application)? And How can apply It, app that loads qvd on stream?  there is a better approach in qliksense?
After going through some posts I found that we can not manage qvf files in our folders. Please suggest the best way to go ahead.
Thanks in Advance

[Rodj]

In general terms I'd say that sticking with a three tier approach is a sound idea. You just need to be cognisant of the obvious differences between Qlik Sense and QlikView. I believe there's some reasonable guidance in the QDF as to how to do it, but it really comes down to how  your environment is set up and what is going to work best in your context. If you have a data warehouse / data lake or similar and other data engineering tools there's always some interesting decisions to be made around how you architect the environment that may see you modifying the approach, but the QDF is a good place to start.

WRT to how you apply the QDF approach, the main question tends to be around where and how you publish your extract and transform layer apps. You could stick them all in a stream with the presentation app, put them in their own stream per project, have them all in one stream....it all depends on how you want to set up your streams and administer security. I'd suggest working backwards from how you want to manage security and presentation of your final user-facing apps. The stream and security model you build to support that will likely shape how you manage the rest.

I tried responding to this question earlier but then noticed my response had gone missing somehow. Hopefully I haven't ended up responding twice and contradicting myself!