On Wed, Sep 25, 2013 at 11:13 AM, Ludovic Dubost <ludovic@xwiki.com> wrote:

Great job indeed. I installed on a test wiki on a 4.5 farm and it looks quite nice.

I have a few questions.

1/ If I understand properly we have a legacy.css where lives all CSS of HTML that would not have been made clean plain bootstrap compatible HTML. Is that so ?

I've took colibri.css: - stripped it from any rules that Bootstrap already covers (reset, standard html elements, forms, etc.); - mapped what I could to Bootstrap components (tables, boxes, panels, etc.); - stripped layout elements and tried to get rid of ColorThemes values (in the end realizing that I can't, without loosing functionality); - cleaned the remaining selectors and properties (removed IE6,7 rules; removed rules that maybe are covered by Bootstrap; etc.) What remained are classes that will be needed by any new skin; classes that kind of represent the CSS API (although there are still parts that could be moved to their specific component file). I though that if someone makes a change in Colibri and that selector is also in legacy.css it should be ported here (that's why I left the selectors as original). Everything that is in legacy.css could be rewritten in junco.css, but it will create a bigger gap between the two skins.

2/ In terms of work, how much and what is needed to get to a point where this could become our default skin, knowing that we should not have feature regression, for instance users should still be able to build their color themes in Wysiwyg like they can with colibri.

I'm sure we can find a solution to do on the fly compiling when saving color themes so to solve the issue of needing LESS.

The integration problems are presented above (__Platform integration problems__). One is the need to be able to build LESS files and update them 'on the fly'. Regarding ColorThemes values, there are 2 files: - global/xwiki-colorthemes.less - which is kind of a colorThemeInit.vm https://github.com/evalica/bootswatch/blob/junco-themes/global/xwiki-colorth... - <theme>/xwiki-colortheme.less - which overrides the initialization and provide colors for a specific theme Blueberry example: https://github.com/evalica/bootswatch/blob/junco-themes/blueberry/xwiki-colo... So when changing ColorThemes values, we need to make sure we update this files.

3/ How much is needed to get rid of legacy and have a fully native skin with the new system ?

As stated above, legacy.css is not a limitation. We could rename it from legacy.css to base.css or api.css, we could maybe write some more performant selectors or maybe write them using LESS nested selectors. The file contains the common denominator for Colibri and Junco and some parts of the layout could be removed from it, in order to make it a true API file. All the selectors that will be removed from here will end up in the xwiki-common.less https://github.com/evalica/bootswatch/blob/junco-themes/global/xwiki-common.... If we rewrite it whole, the only problem is that it will be harder to update both Colibri and Junco.

4/ What is the migration path for a wiki where a custom skin has been built on colibri

Because of the mapping between XWiki and Bootstrap the Colibri classes will be covered in Junco. The problem are base HTML elements, that may have other default values. I'm sure tables will produce some problems, because in Colibri they didn't had any default padding, while after mapping they do. If someone use table for layout, there will be some extra spacing. There is not migration path for skins like that. Although we should have minor problems, they will need to be tested and fixed manually.

5/ What are the potential consequences on future compatibility with colibri based skin, particularly if we start modifying our HTML produced by different modules ?

So the current HTML structure is covered by legacy.css and xwiki-mapping.less. If someone will want to change the HTML structure, you need to: - make sure that the classes you delete are not 'API', this means they are not called in legacy or xwiki-mapping. If a class is present in these files means it also has a significance in Colibri; - if you find a certain class, don't delete it. Instead just append your new class or Bootstrap class. Keeping the legacy class, it will have a fallback if Colibri skin will be used. Of course there will be differences between skins if we start using Bootstrap components and that is normal (because Colibri will not know of their existence). Thanks, Caty

Ludovic

2013/9/20 Ecaterina Moraru (Valica) <valicac@gmail.com>

...

Hi devs,

For the past weeks I've been working on a skin based on Bootstrap[1]. You can read more about it and test the new Junco Skin at http://extensions.xwiki.org/xwiki/bin/view/Extension/Junco+Skin

This proposal is about using the Junco Skin (Blueberry Theme<

http://extensions.xwiki.org/xwiki/bin/view/Extension/Junco+Skin#HTheme:Blueb...

...

...

) for xwiki.org. I prepared some responsive screenshots for the xwiki.org Homepage and an Extension page.

http://incubator.myxwiki.org/xwiki/bin/view/Improvements/XWikiOrgJuncoHomepa...

...

This is my +1

Please report any issues on https://github.com/evalica/bootswatch/issues

__Advantages__ - a change is always welcomed, shows the users there is activity on the website; - the skin is responsive; - by using Bootstrap we have the whole framework's power to use (grids, components, etc.); - we have the chance to test a bit the skin in production and see the possible bugs, in order to later integrate; - IMO the skin looks nice :)

__Disadvantages__ - the only disadvantage is that there will be bugs and they will take

some

...

time to be detected and fixed;

__Platform integration problems__ 1. currently the new skin uses the HTML5 doctype. This is needed if we were to use some Bootstrap JS components (carousel, menus, etc.) - which we currently don't on xwiki.org, so I could use the old doctype (but lose some of the testing purpose). Because of the HTML5 doctype, the HTML validation fails. See http://jira.xwiki.org/browse/XWIKI-7552

2. Junco Skin currently doesn't have support for changing the ColorThemes on the fly. I would need the help of a developer to fix this problem. Is not so much a problem for xwiki.org (I could fix them by duplicating some code in the xwiki.org skin), but it is a problem for the integration.

3. Being build on Bootstrap, it needs LESS to compile the files into CSS. Again I would need a developer to see how we could integrate the building of the themes in platform. Right now this is done locally, partially manual by using Grunt.

Tell me what you think and take some time to test the skin, Caty

[1] http://getbootstrap.com/ _______________________________________________ devs mailing list devs@xwiki.org http://lists.xwiki.org/mailman/listinfo/devs

-- Ludovic Dubost Founder and CEO Blog: http://blog.ludovic.org/ XWiki: http://www.xwiki.com Skype: ldubost GTalk: ldubost _______________________________________________ devs mailing list devs@xwiki.org http://lists.xwiki.org/mailman/listinfo/devs

Hi Caty, Just before giving some arguments below, let me say that I'm very happy that you're working on xwiki.org and I globally like very much what you're doing, no problem there :) On Sep 27, 2013, at 7:22 PM, Ecaterina Moraru (Valica) <valicac@gmail.com> wrote:

Hi,

Thank you for your votes. I've integrate the Homepage changes and we are using the Junco skin on xwiki.org

__Discussion 1__ So one of the problems is 'Clearing your Cache'. In order to see the changes correctly you need to refresh the page and make sure the cache is updated with the new style rules. It would be great if someone would know how we could force the clearing of the cache. Maybe http://jira.xwiki.org/browse/XWIKI-6073 ?

__Discussion 2__ On a related topic, Vincent thinks we should not use a 'centered' design, instead have (just like Colibri has) a 'full width 'design.

TBH I don't know yet what is better yet but I'm asking myself and the comments below are my comments only from first sight. I'm sure I can adapt or have a different opinion given time :)

I've made some screenshots (please take a look at them in full view): * Centered: http://incubator.myxwiki.org/xwiki/bin/download/Improvements/XWikiOrgJuncoHo... **Full width: http://incubator.myxwiki.org/xwiki/bin/download/Improvements/XWikiOrgJuncoHo...

Try with these too: * http://extensions.xwiki.org/xwiki/bin/view/Main/WebHome * http://www.xwiki.org/xwiki/bin/view/ReleaseNotes/ * http://www.xwiki.org/xwiki/bin/view/Main/News

Why I disagree (having full width and prefer the centered version): - xwiki.org is a documentation site so most of its content needs to be read;

Centered is nice for websites, ie site for which you control the content and can lay it out nicely. This is not really our case and it's hard to have something that looks nice for all pages.

- From a readability point of view, if the line is too long it is harder for the eye to read it, please read http://baymard.com/blog/line-length-readability ;

I haven't read this link yet but what I can tell you is that right now I have a hard time reading content because: * I see only a really small portion of it at once * The font is quite larger than before I don't know yet if it's better or worse than before but I certainly need some adaptation time. I think the main issue is that our UI is too crowded for having a centered + large font UI. For example I like this site quite a lot (Note that they use the full width too): http://docs.vagrantup.com/v2/installation/index.html It's large too but the interface is very clean and I find it nice. In our case we have various sizes everywhere (look at menu sizes vs content size) and lots of eye distraction with all the menus.

- Being a responsive skin, for smaller devices the content goes full width, so the centered width is just for large resolutions;

Thanks -Vincent

Let me know what you think, Caty

On Thu, Sep 26, 2013 at 11:08 AM, Ecaterina Moraru (Valica) < valicac@gmail.com> wrote:

...

On Wed, Sep 25, 2013 at 11:13 AM, Ludovic Dubost <ludovic@xwiki.com>wrote:

...

Great job indeed. I installed on a test wiki on a 4.5 farm and it looks quite nice.

I have a few questions.

1/ If I understand properly we have a legacy.css where lives all CSS of HTML that would not have been made clean plain bootstrap compatible HTML. Is that so ?

I've took colibri.css: - stripped it from any rules that Bootstrap already covers (reset, standard html elements, forms, etc.); - mapped what I could to Bootstrap components (tables, boxes, panels, etc.); - stripped layout elements and tried to get rid of ColorThemes values (in the end realizing that I can't, without loosing functionality); - cleaned the remaining selectors and properties (removed IE6,7 rules; removed rules that maybe are covered by Bootstrap; etc.)

What remained are classes that will be needed by any new skin; classes that kind of represent the CSS API (although there are still parts that could be moved to their specific component file). I though that if someone makes a change in Colibri and that selector is also in legacy.css it should be ported here (that's why I left the selectors as original).

Everything that is in legacy.css could be rewritten in junco.css, but it will create a bigger gap between the two skins.

...

2/ In terms of work, how much and what is needed to get to a point where this could become our default skin, knowing that we should not have feature regression, for instance users should still be able to build their color themes in Wysiwyg like they can with colibri.

I'm sure we can find a solution to do on the fly compiling when saving color themes so to solve the issue of needing LESS.

The integration problems are presented above (__Platform integration problems__). One is the need to be able to build LESS files and update them 'on the fly'. Regarding ColorThemes values, there are 2 files: - global/xwiki-colorthemes.less - which is kind of a colorThemeInit.vm

https://github.com/evalica/bootswatch/blob/junco-themes/global/xwiki-colorth... - <theme>/xwiki-colortheme.less - which overrides the initialization and provide colors for a specific theme Blueberry example: https://github.com/evalica/bootswatch/blob/junco-themes/blueberry/xwiki-colo...

So when changing ColorThemes values, we need to make sure we update this files.

...

3/ How much is needed to get rid of legacy and have a fully native skin with the new system ?

As stated above, legacy.css is not a limitation. We could rename it from legacy.css to base.css or api.css, we could maybe write some more performant selectors or maybe write them using LESS nested selectors. The file contains the common denominator for Colibri and Junco and some parts of the layout could be removed from it, in order to make it a true API file. All the selectors that will be removed from here will end up in the xwiki-common.less https://github.com/evalica/bootswatch/blob/junco-themes/global/xwiki-common....

If we rewrite it whole, the only problem is that it will be harder to update both Colibri and Junco.

...

4/ What is the migration path for a wiki where a custom skin has been built on colibri

Because of the mapping between XWiki and Bootstrap the Colibri classes will be covered in Junco. The problem are base HTML elements, that may have other default values. I'm sure tables will produce some problems, because in Colibri they didn't had any default padding, while after mapping they do. If someone use table for layout, there will be some extra spacing. There is not migration path for skins like that. Although we should have minor problems, they will need to be tested and fixed manually.

...

5/ What are the potential consequences on future compatibility with colibri based skin, particularly if we start modifying our HTML produced by different modules ?

So the current HTML structure is covered by legacy.css and xwiki-mapping.less.

If someone will want to change the HTML structure, you need to: - make sure that the classes you delete are not 'API', this means they are not called in legacy or xwiki-mapping. If a class is present in these files means it also has a significance in Colibri; - if you find a certain class, don't delete it. Instead just append your new class or Bootstrap class. Keeping the legacy class, it will have a fallback if Colibri skin will be used.

Of course there will be differences between skins if we start using Bootstrap components and that is normal (because Colibri will not know of their existence).

Thanks, Caty

...

Ludovic

2013/9/20 Ecaterina Moraru (Valica) <valicac@gmail.com>

...

Hi devs,

For the past weeks I've been working on a skin based on Bootstrap[1]. You can read more about it and test the new Junco Skin at http://extensions.xwiki.org/xwiki/bin/view/Extension/Junco+Skin

This proposal is about using the Junco Skin (Blueberry Theme<

http://extensions.xwiki.org/xwiki/bin/view/Extension/Junco+Skin#HTheme:Blueb...

...

...

) for xwiki.org. I prepared some responsive screenshots for the xwiki.org Homepage and an Extension page.

http://incubator.myxwiki.org/xwiki/bin/view/Improvements/XWikiOrgJuncoHomepa...

...

This is my +1

Please report any issues on

https://github.com/evalica/bootswatch/issues

...

__Advantages__ - a change is always welcomed, shows the users there is activity on the website; - the skin is responsive; - by using Bootstrap we have the whole framework's power to use (grids, components, etc.); - we have the chance to test a bit the skin in production and see the possible bugs, in order to later integrate; - IMO the skin looks nice :)

__Disadvantages__ - the only disadvantage is that there will be bugs and they will take

some

...

time to be detected and fixed;

__Platform integration problems__ 1. currently the new skin uses the HTML5 doctype. This is needed if we were to use some Bootstrap JS components (carousel, menus, etc.) - which we currently don't on xwiki.org, so I could use the old doctype (but lose some of the testing purpose). Because of the HTML5 doctype, the HTML validation fails. See http://jira.xwiki.org/browse/XWIKI-7552

2. Junco Skin currently doesn't have support for changing the ColorThemes on the fly. I would need the help of a developer to fix this problem. Is not so much a problem for xwiki.org (I could fix them by duplicating some code in the xwiki.org skin), but it is a problem for the integration.

3. Being build on Bootstrap, it needs LESS to compile the files into CSS. Again I would need a developer to see how we could integrate the building of the themes in platform. Right now this is done locally, partially manual by using Grunt.

Tell me what you think and take some time to test the skin, Caty

[1] http://getbootstrap.com/ _________________________

Hi, On Sat, Sep 28, 2013 at 9:54 AM, Thomas Mortagne <thomas.mortagne@xwiki.com>wrote:

On Fri, Sep 27, 2013 at 7:22 PM, Ecaterina Moraru (Valica) <valicac@gmail.com> wrote:

...

Hi,

Thank you for your votes. I've integrate the Homepage changes and we are using the Junco skin on xwiki.org

__Discussion 1__ So one of the problems is 'Clearing your Cache'. In order to see the changes correctly you need to refresh the page and make sure the cache is updated with the new style rules. It would be great if someone would know how we could force the clearing of the cache. Maybe http://jira.xwiki.org/browse/XWIKI-6073 ?

Yea this one is a tough one we have since a while now...

...

__Discussion 2__ On a related topic, Vincent thinks we should not use a 'centered' design, instead have (just like Colibri has) a 'full width 'design.

I've made some screenshots (please take a look at them in full view): * Centered:

http://incubator.myxwiki.org/xwiki/bin/download/Improvements/XWikiOrgJuncoHo...

...

**Full width:

http://incubator.myxwiki.org/xwiki/bin/download/Improvements/XWikiOrgJuncoHo...

...

Why I disagree (having full width and prefer the centered version): - xwiki.org is a documentation site so most of its content needs to be read; - From a readability point of view, if the line is too long it is harder for the eye to read it, please read http://baymard.com/blog/line-length-readability ; - Being a responsive skin, for smaller devices the content goes full

width,

...

so the centered width is just for large resolutions;

Let me know what you think,

Centered is one thing but right now the width of the content is way too small for me. This can be OK on a website with very few information but xwiki.org is a wiki, not just a website with a few pages designed specifically for small width.

Sure I don't need to move my eyes much in http://www.xwiki.org/xwiki/bin/view/ReleaseNotes/ but at the same time it's totally unreadable anyway.

Whatever the scientific arguments I can say that a layout taking the full available width is way more comfortable for me.

I was looking at the site and I think a fixed width would be good for plenty of pages. For instance on this page: http://www.xwiki.org/xwiki/bin/view/ReleaseNotes/ReleaseNotesXWiki43#History Once you scroll down the page on a largish screen (mine's 1680px wide) you see plenty of one-line text interspersed between images, which is hard to read. Thus I'm more in favor of a fixed width website, but maybe slighyly larger than 960px. Maybe around 1164px wide? Guillaume

...

Caty

On Thu, Sep 26, 2013 at 11:08 AM, Ecaterina Moraru (Valica) < valicac@gmail.com> wrote:

...

On Wed, Sep 25, 2013 at 11:13 AM, Ludovic Dubost <ludovic@xwiki.com

wrote:

...

...

Great job indeed. I installed on a test wiki on a 4.5 farm and it looks quite nice.

I have a few questions.

1/ If I understand properly we have a legacy.css where lives all CSS of HTML that would not have been made clean plain bootstrap compatible

HTML.

...

...

Is that so ?

I've took colibri.css: - stripped it from any rules that Bootstrap already covers (reset, standard html elements, forms, etc.); - mapped what I could to Bootstrap components (tables, boxes, panels, etc.); - stripped layout elements and tried to get rid of ColorThemes values (in the end realizing that I can't, without loosing functionality); - cleaned the remaining selectors and properties (removed IE6,7 rules; removed rules that maybe are covered by Bootstrap; etc.)

What remained are classes that will be needed by any new skin; classes that kind of represent the CSS API (although there are still parts that could be moved to their specific component file). I though that if someone makes a change in Colibri and that selector is also in legacy.css it should be ported here (that's why I left the selectors as original).

Everything that is in legacy.css could be rewritten in junco.css, but it will create a bigger gap between the two skins.

...

2/ In terms of work, how much and what is needed to get to a point where this could become our default skin, knowing that we should not have feature regression, for instance users should still be able to build their color themes in Wysiwyg like they can with colibri.

I'm sure we can find a solution to do on the fly compiling when saving color themes so to solve the issue of needing LESS.

The integration problems are presented above (__Platform integration problems__). One is the need to be able to build LESS files and update them 'on the fly'. Regarding ColorThemes values, there are 2 files: - global/xwiki-colorthemes.less - which is kind of a colorThemeInit.vm

https://github.com/evalica/bootswatch/blob/junco-themes/global/xwiki-colorth...

...

- <theme>/xwiki-colortheme.less - which overrides the initialization and provide colors for a specific theme Blueberry example:

https://github.com/evalica/bootswatch/blob/junco-themes/blueberry/xwiki-colo...

...

So when changing ColorThemes values, we need to make sure we update this files.

...

3/ How much is needed to get rid of legacy and have a fully native skin with the new system ?

As stated above, legacy.css is not a limitation. We could rename it from legacy.css to base.css or api.css, we could maybe write some more performant selectors or maybe write them using LESS nested selectors. The file contains the common denominator for Colibri and Junco and some parts of the layout could be removed from it, in order to make it a true API file. All the selectors that will be removed from here will end up

in

...

the xwiki-common.less

https://github.com/evalica/bootswatch/blob/junco-themes/global/xwiki-common....

...

If we rewrite it whole, the only problem is that it will be harder to update both Colibri and Junco.

...

4/ What is the migration path for a wiki where a custom skin has been built on colibri

Because of the mapping between XWiki and Bootstrap the Colibri classes will be covered in Junco. The problem are base HTML elements, that may have other default values. I'm sure tables will produce some problems, because in Colibri they

didn't

...

had any default padding, while after mapping they do. If someone use table for layout, there will be some extra spacing. There is not migration path for skins like that. Although we should have minor problems, they will need to be tested and fixed manually.

...

5/ What are the potential consequences on future compatibility with colibri based skin, particularly if we start modifying our HTML produced by different modules ?

So the current HTML structure is covered by legacy.css and xwiki-mapping.less.

If someone will want to change the HTML structure, you need to: - make sure that the classes you delete are not 'API', this means they are not called in legacy or xwiki-mapping. If a class is present in these files means it also has a significance in Colibri; - if you find a certain class, don't delete it. Instead just append your new class or Bootstrap class. Keeping the legacy class, it will have a fallback if Colibri skin will be used.

Of course there will be differences between skins if we start using Bootstrap components and that is normal (because Colibri will not know of their existence).

Thanks, Caty

...

Ludovic

2013/9/20 Ecaterina Moraru (Valica) <valicac@gmail.com>

...

Hi devs,

For the past weeks I've been working on a skin based on Bootstrap[1]. You can read more about it and test the new Junco Skin at http://extensions.xwiki.org/xwiki/bin/view/Extension/Junco+Skin

This proposal is about using the Junco Skin (Blueberry Theme<

http://extensions.xwiki.org/xwiki/bin/view/Extension/Junco+Skin#HTheme:Blueb...

...

...

...

) for xwiki.org. I prepared some responsive screenshots for the xwiki.org Homepage and an Extension page.

http://incubator.myxwiki.org/xwiki/bin/view/Improvements/XWikiOrgJuncoHomepa...

...

...

This is my +1

Please report any issues on

https://github.com/evalica/bootswatch/issues

...

__Advantages__ - a change is always welcomed, shows the users there is activity on

the

...

website; - the skin is responsive; - by using Bootstrap we have the whole framework's power to use (grids, components, etc.); - we have the chance to test a bit the skin in production and see the possible bugs, in order to later integrate; - IMO the skin looks nice :)

__Disadvantages__ - the only disadvantage is that there will be bugs and they will take some time to be detected and fixed;

__Platform integration problems__ 1. currently the new skin uses the HTML5 doctype. This is needed if we were to use some Bootstrap JS components (carousel, menus, etc.) - which we currently don't on xwiki.org, so I could use the old doctype (but lose some of the testing purpose). Because of the HTML5 doctype, the HTML validation fails. See http://jira.xwiki.org/browse/XWIKI-7552

2. Junco Skin currently doesn't have support for changing the ColorThemes on the fly. I would need the help of a developer to fix this problem. Is not so much a problem for xwiki.org (I could fix them by duplicating some code in the xwiki.org skin), but it is a problem for the integration.

3. Being build on Bootstrap, it needs LESS to compile the files into CSS. Again I would need a developer to see how we could integrate the building of the themes in platform. Right now this is done locally, partially manual by using Grunt.

Tell me what you think and take some time to test the skin, Caty

[1] http://getbootstrap.com/ _______________________________________________ devs mailing list devs@xwiki.org http://lists.xwiki.org/mailman/listinfo/devs

-- Ludovic Dubost Founder and CEO Blog: http://blog.ludovic.org/ XWiki: http://www.xwiki.com Skype: ldubost GTalk: ldubost _______________________________________________ devs mailing list devs@xwiki.org http://lists.xwiki.org/mailman/listinfo/devs

_______________________________________________ devs mailing list devs@xwiki.org http://lists.xwiki.org/mailman/listinfo/devs

-- Thomas Mortagne _______________________________________________ devs mailing list devs@xwiki.org http://lists.xwiki.org/mailman/listinfo/devs