[Helma-user] Rocket the Super Rabbit + new approach to web development

[Daniel Ruthardt]

Hi Samuel and list,
> The different tasks like data validation, form rendering etc. can then 
> be handled decentralized by specialised classes - no problem with 
> that. (As long as the user don't have to install dozens of libs 
> because of all the dependencie s ;) )
I released a working (but not actively developed anymore, so probably 
clean-up and Helma specific updates needed) framework capable of doing 
would you describe above quite some time ago. It is based on a central 
xml file (configuring prototypes and properties validation) and 
decentralized xml files for each prototype describing form rendering.
As far as I know the download site doesn't exist anymore, but feel free 
to send my an off-list email to hear more about it or to get it.
> To start the modelling from an existing db is in fact a process I also 
> thought of and is not implemented yet only because of a shortage of time.
Did you see my warp repository (see helma-dev mailing list)? This is 
probably the most easiest way to start of Helma prototyping from an 
existing db.

> What I want is a central (typed) data model, that can be used to 
> automatise/ease most of the common web developer tasks. It should be 
> designed the way, that you only have to define that model add some CSS 
> and will have the most basic input/output/data handling functionality 
> ready.
> [...]
> There is so much momentum in the Javascript/Ajax development. Sadly 
> ROR is the big player on the server side development. But with the 
> right efforts, maybe Helma could catch up.
This sounds very similar to what I want to work on in the near future, 
although my approach is more academic and will not be Helma-based but 
using Helma as one of many web application framework alternatives.
I am not sure if anybody of you ever heard of Executable UML (little 
development and research effort so far), did you? Its an approach to cut 
the development effort for medium and big sized projects, keep modeling 
in sync with development, increase re usability and to develop without 
much architectural thinking.

In the most ideal case, software development - or in the special case we 
are all interested in which is web-development - the development effort 
is decreased to the point where you only need to model your web 
application using UML using specialised UML profiles. Once the modeling 
process is done, web application development should be finished and the 
model should be directly executable with for example Helma as so called 
model compiler.

And now imagine:

The database gets created and is kept in sync by the configured database 
model compiler (i.e. MySQL)
The web application gets created and is kept in sync by the configured 
logic model compiler (i.e. Helma)
and so on...

Helma update from 1.x to 2.0?
"Downgrade" (-; to PHP?
Convert to a Java GUI application?

Just change your model compiler!

Well, this is the big idea - I won't be finished by tomorrow although (-;

The really important thing about that all related to Helma is that the 
first available model compiler of course will be Helma and one other.
So this wouldn't only help Helma to just catch up with RoR but maybe 
overtake it by some steps (-;

Greetings,
Daniel