Hello Everyone ,
The weight of the cms files inside jason's branch now stands at 13+ MB
The media files alone takes up 9+ MB. I think the increase in size is due to the new richtext editors.
IMO we should distribute the cms editors with only the default language (en). We can then provide the full multi-langage versions of the editors in a separate down. We also need to look for common features between the two editors and merge were possible things like the simley icons, etc.
I'm thinking do we really need two editors? From the looks of things I would have to say that FCKEditor has the larger developer base of the two editors. It's ranked as #3 on SF as the most active project. It has more HTML features than TinyMCE (but might not be as simple). The plugin interface seems very easy to adopt. And it seems to process things faster than TinyMCE. Having a single editor would mean that we only have to create one plugin to extend that editor. Having two editor means that one editor might lag behind the other in terms of features.
What do you think about standardizing on a single Richtext editor?
And as much as I like FCK's latest features, I'm still torn between editors, as I really like a lot of the features in TinyMCE.
I'd prefer if we kept this option, except I propose we provide a little more abstraction and the ability to plug-in/roll-your-own editor. We can slim down these installs for sure (I'll clean up what isn't needed). Then I suggest we offer our modified editor installs as separate add-ons to the product. In other words, don't ship with any RichText editor, and allow users to choose from the array of pluggable editors available on our site.
Abstraction points include:
* Custom configuration properties for individual editors (perhaps on their own manager tab?)
* TV rich-text control scripts / mergeContent() in document.parser
* Manager document editing / mutate_content.dynamic.action
Then later we can add an auto-installer in the Manager for the packages along with a definition for defining an editor package for MODx.
If Ryan will add the abstraction points as tasks for next release, I'll see if I can get 'er done, unless there are objections to this approach?