Re: [xwiki-devs] [Discussion][Skin] Bootstrap integration inside platform

Hi Cathy, To integrate bootstrap with XWiki, several approaches has been investigated: A. Jerome, with the Lyrebird skin, has adapted the bootstrap CSS to the current XWiki markup and templates B. Myself, with the Bluebird skin, has adapted the XWiki templates to the bootstrap CSS to partially target the Flamingo skin look. C. You, Cathy, with the Junco skin, has bridge the bootstrap CSS to adapt it to the XWiki markup and templates The inconvenience with A. was the inability to use any of the existing bootstrap variations available, and to follow the bootstrap evolution easily. So, A is no go for me.

The inconvenience with B. is that it require the whole XWiki markup to be adapted, including those produced by existing extensions, providing no compatibility with the past. The inconvenience with C. is that it require any bootstrap variation to be "recompiled" (using less) to be adapted. My personal idea, that I have defended during the last seminar, is that we should target a state where we are free to change the HTML markup without breaking existing extensions, and the way I propose to do so was to provide more abstraction then we have currently, by providing XWiki macro to generate most of the HTML needed. From the discussion we had, I have understand that reaching that point will be not easy, and could be an unreachable dream, but I am still convinced it could be achieved for a lot of simple applications, so most of them. What I would find a very annoying situation, is that we refrain to change our skin, and our html markup, simply to keep the compatibility with the past. The currently generated markup is sometime over complex, and followed our own rules. I see the bootstrap integration as the occasion to choose a more popular html markup style, that is already widely used on the net, and that could be easily followed by newcomers. Bootstrap propose to use a very semantic html markup while providing a complete set of UI features. It is a far more clean and simple HTML than we have currently IMO. Also, using a purely bootstrap based solution, without any adaptation, open us to the wide range of bootstrap variation, either paid or free. Based on our previous experiences, I have the feeling that we have all what we need to get the best of both worlds, and migrate smoothly. We may use method C. (Junco) to provide a good compatibility level for existing extension and to be upgraded templates, while we could use method B. to evolve our skin and simplify our markup. And we may also improve at the same time, our feature that prevent us to manually generate the HTML. Regarding the JS aspect, Louis-Marie is perfectly right, we have to vote the deprecation of Prototype.js and the usage of JQuery as our default JS framework. I am confident we could reach a consensus on that point, and for those not yet convince, I remind you this post by Sam Stephenson the creator of Prototype.js initially mentionned by Jean-Vincent: http://sstephenson.us/posts/you-are-not-your-code So I invite you to open an independent thread for that vote. Thanks, On Wed, Feb 19, 2014 at 7:43 PM, Joshua Marks <jmarks(a)curriki.org> wrote: