About deployment of examples

[dvojtise]

Hi I 'm thinking about a way to improve how the examples are presented to the users:

currently, we have the following:

• 1 deployer plugin(language), that must be used in a first eclipse in order to install the language plugins
• 1 deployer plugin (model), that must be used in the second workbench in order to install model examples of the language

This approach is mainly focused on the language designer point of view, this is nice for people who already know about language development but probably less intuitive for a newcomer not from our community.

In Eclipse, it exists the following feature (that I find very useful in many situations): you can click on a plugin (in the plugin view) and then click on import as then "source project" or even "project from a repository" (the first simply install the source if the plugin have been distributed with their sources (ie. work 100% offline but with the cost of a larger binary), the second, knows the git repository location and automate the checkout of the source directly from git at the correct revision (ie. allows tthe user to easily fork the code but require a network connexion)) Using this when presenting, we could focus on the final user point of view first(ie. the model designer)

• the language is fully installed in Eclipse (as if the language was a normal feature to be added in eclipse using update site)
• 1 deployer plugin (model) so the user can play with the gemoc debugger directly

Then the associated documentation would explain how this was built by importing all the language plugin sources and give more information about language development. (the import would be done either by explaining how to get the source via the above UI, or automated via a wizard based on the same principle in order to import all the languages plugins in a single action)

What do you think of this ?