Raise your level of abstraction
It's great work, and very high-profile. Well done!
That was quicker than I thought. No real surprise though, after Lightswitch and AppFactory and eZ-App-Factory and whatever else. The once empty room slowly becomes crowded.
This is an interesting way how to create applications!
I thought about similar features as those about UML modelling or DSL coding or drawing may be too complicated for an end user. I would create a data driven approach first - eg. creating tables and rows, then select tables to define forms for it. And finishing it with menu structuring steps.
Using template choosing and let the user modify the theme is a way to get started from the look and feel point. Also the navigational aspect is a good starting point for beginners - simple tree like structure.
But three weak points for the more experienced 'programmer' are:
I am not able to export a model in a XML format, nor am I able to import one. So for me as a tool vendor be not able to add support from my UML modelling approach first. Eg. transforming my model to your target XML format and then simply use your website to generate my application with your code generators.
So this would be a missing integration point in my view.
Another missing point, but not that important and maybe due to the beta state is the available data sources. It is not clear to me why you do not support for sample simple REST based data sources to any JSON formatted URI. For sample, Microsoft supports at least this feature and I was able to directly see the data in the design (even if they only fetch 5 rows). See Microsoft App Studio for it.
I have reviewed the generated code and figured out the repository implementation of the data source. It would be easy for me to pick up related URI's and point them to my data sources. So it is ok to have only one local and one cloud data source.
The third point from my view is the important one:
I have a more complex data model, say, within my UML model. What do you think about an entity schema import in any format you define and then let the user structure forms and menu around the schema?
Conclusion: At the current state, it is a nice try worth to see what is possible from the end user aspects, but for me, I feel there is missing an integration approach to enable other tools to integrate more easily. In my view this missing integration feature is a reason not to use it.
Do you plan to add such an integration approach / feature?
Thank you very much Lothar for your interest & comments on the topic.
A lot of things has already happened and product has evolved a lot from the initial publication.
I invite you to take a look to the last version of the technology at:
Both of them: backend and mobile apps can be connected also.
The current versions addresses some of your concerns.
Take a special review on datasources as managed on Appnow (allowing users to use Excel files with real data as input models).
It is true that more export/import capabilites can be added to Radarc-Online for advanced users, but at the same time we try to keep it as simple as possible for Citizen Developers.
We keep pushing features on the roadmap. Thanks for your valuable feedback!
All the best!
I have had a review of Radarc Online. Let's see what happens in the future.
The AppNow is interesting from the bidirectional modeling aspect. I suppose there will be additional backend options in the future...
What technology is behind the parser? Do you used any parser generator, or a more complex thing like a language workbench? (eg. JPS/Spoofax)
What is about other parsers like GOLD parser (CGT based, see http://goldparser.org/doc/index.htm)?
Eg. I am using C++ thus the GOLD parser is an interesting solution, because of the independent language engines available in different implementation languages.