Re: RFC: CForms Roadmap

2007-01-15 Thread Alexander Klimetschek

Jason Johnston schrieb:

Can't you just escape the .?

#myform\.somefield {...}

I remember CSS selectability was discussed back when the id naming rules 
were being proposed, and I thought I remembered this working correctly 
in the major browsers.


Here's some thread linkage: 
http://marc.theaimsgroup.com/?l=xml-cocoon-devw=2r=1s=css+escape+cforms+widget+idq=b 


Thanks, escaping works. Would be cool to have that hint in the CForms 
documentation.


Alex

--
Alexander Klimetschek
http://www.mindquarry.com



Re: RFC: CForms Roadmap

2007-01-10 Thread Carsten Ziegeler
Max Pfingsthorn wrote
 By the way, has anything improved authentication-wise in Cocoon? 

We have the simple CAuth framework in Cocoon which you could use for
authentication, but I think especially with Cocoon 2.2 which comes with
Spring, using Spring Acegi Security is a good choice.

Carsten
-- 
Carsten Ziegeler - Chief Architect
http://www.s-und-n.de
http://www.osoco.org/weblogs/rael/


Re: RFC: CForms Roadmap

2007-01-10 Thread Alexander Klimetschek

Hi Jeremy, hi all,

I have another feature request for Cforms: change the widget hierarchy 
separator from . to something else (_), eg. having a form called 
myform containing a widget named somefield would result in the fully 
qualified widget id myform_somefield.


The problem is that having a dot in the ids makes it very hard to do CSS 
styling for the final HTML version of the form. This is because CSS id 
selectors cannot contain a ., that is reserved for class selectors. 
For example:


#myform.somefield {
  // styling...
}

is incorrect CSS or at least is interpreted as

#myform .somefield {
  // styling..
}

aka element with id myform and below any element with the class somefield.

A way to work around is either to define your special classes via 
fi:styling or to write complicated selectors that take the structure of 
the HTML document into account (div.someclass div.foobar input.forms 
...). Both are very limited: fi:styling is difficult if you are using 
custom styled elements (=custom XSL) where you cannot easily set a class 
and it does not work if you want to use it together with the standard 
classes like active or output. And it requires the CSS designer to 
change the form templates probably on each modification, which IMHO 
violates separation of concerns. The other solution I don't wanna talk 
about... ;-)


Jeremy you said that the dot separator is basic assumption for all the 
cforms java code, so that it would be a massive change. Nevertheless I 
find it quite a major flaw in the cforms design (the only one I know of ;-).


WDYT?

Alex

Jeremy Quinn schrieb:

Hi All

Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid 
platform to complete the modernisation of CForms client-side code.


I hope the main outcomes of this will be :
Leaner: only the resources that are used will be loaded by cforms
Richer: more interactive widgets with wider x-platform support
Cleaner: eliminate most of the messy script tags scattered through 
our forms
Simpler: use more html templates to simplify our xslt (and simpler 
to adapt the widgets)


Here is a list of specific goals I would love us to achieve for the next 
release.
I hope to be working on this stuff, you would be very welcome if you'd 
like to join in 
If you would like to take on something from this list (or something I 
missed out) please discuss it on the dev list so no-one is duplicating 
effort.



Replacements :

Date/Time widget : replace MattKruse stuff with Dojo's implementation.
Help  validation popups: replace MattKruse stuff with a new Dojo 
implementation.

Tabs: replace with Dojo Tabs
RichText: replace htmlarea with Dojo Editor, using a new fi:styling, so 
htmlarea can still be used

MultiValue Editor: re-implement as a Dojo widget
MultiList (OptionTransfer) Selector: replace with a new Dojo widget
Maps: not even sure our current one is working, replace with Dojo (Yahoo 
and Google)



Possible Additions :

Easy graphically rich buttons, dialogs, menus etc.
Charts to plot user data
Colour picker
Re-sizable textarea
Sliders: graphical selector for number ranges etc.
Spinner: adjust values up and down with ± buttons
Validation: plug in client-side validation where common datatypes exist 
between CForms and Dojo, make new ones


Issues:

We need to do this work in such a way that has the absolute minimum 
impact on existing cforms projects.
eg. adapting Dojo widgets to work the CForms way and not the other way 
around.


We need to make it easier to customise the style, layout and behaviour 
of our supplied widgets.



We are probably going to have issues with i18n and l10n.
Dojo is only just starting to deal with this area, while CForms has 
always delt with it.

Dojo, performs i18n on the client instead of on the server as cforms does.
Very few of Dojo's widgets are i18n enabled yet.
Dojo uses a different format of i18n message dictionary than we do (all 
of this is kind of obvious).


Most of the work above will involve either extending existing dojo 
widgets or making new ones.

We ought to be adding i18n/l10n as we go along.

We need to decide whether we want to keep our message dictionary format 
(transforming it on the fly for dojo) or start using Dojo's format (for 
widget internals).



What did I miss out ?


I hope this whets your appetite !

Let's get on with the replacements first !!

WDYT ?

regards Jeremy






--
Alexander Klimetschek
http://www.mindquarry.com



Re: RFC: CForms Roadmap

2007-01-10 Thread Jason Johnston

Alexander Klimetschek wrote:

Hi Jeremy, hi all,

I have another feature request for Cforms: change the widget hierarchy 
separator from . to something else (_), eg. having a form called 
myform containing a widget named somefield would result in the fully 
qualified widget id myform_somefield.


The problem is that having a dot in the ids makes it very hard to do CSS 
styling for the final HTML version of the form. This is because CSS id 
selectors cannot contain a ., that is reserved for class selectors. 


Can't you just escape the .?

#myform\.somefield {...}

I remember CSS selectability was discussed back when the id naming rules 
were being proposed, and I thought I remembered this working correctly 
in the major browsers.


Here's some thread linkage: 
http://marc.theaimsgroup.com/?l=xml-cocoon-devw=2r=1s=css+escape+cforms+widget+idq=b




For example:

#myform.somefield {
  // styling...
}

is incorrect CSS or at least is interpreted as

#myform .somefield {
  // styling..
}

aka element with id myform and below any element with the class 
somefield.


A way to work around is either to define your special classes via 
fi:styling or to write complicated selectors that take the structure of 
the HTML document into account (div.someclass div.foobar input.forms 
...). Both are very limited: fi:styling is difficult if you are using 
custom styled elements (=custom XSL) where you cannot easily set a class 
and it does not work if you want to use it together with the standard 
classes like active or output. And it requires the CSS designer to 
change the form templates probably on each modification, which IMHO 
violates separation of concerns. The other solution I don't wanna talk 
about... ;-)


Jeremy you said that the dot separator is basic assumption for all the 
cforms java code, so that it would be a massive change. Nevertheless I 
find it quite a major flaw in the cforms design (the only one I know of 
;-).


WDYT?

Alex


Re: RFC: CForms Roadmap

2007-01-09 Thread Rice Yeh

Hi,
 Since namespace is employed now. Widgets' naming might start not to
include namespace-purpose prefix. For example, CFormsSuggest is better to be
just named Suggest.

Rice

On 1/7/07, Jeremy Quinn [EMAIL PROTECTED] wrote:


Hi All

Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid
platform to complete the modernisation of CForms client-side code.

I hope the main outcomes of this will be :
Leaner: only the resources that are used will be loaded by cforms
Richer: more interactive widgets with wider x-platform support
Cleaner: eliminate most of the messy script tags scattered
through
our forms
Simpler: use more html templates to simplify our xslt (and simpler
to adapt the widgets)

Here is a list of specific goals I would love us to achieve for the
next release.
I hope to be working on this stuff, you would be very welcome if
you'd like to join in 
If you would like to take on something from this list (or something I
missed out) please discuss it on the dev list so no-one is
duplicating effort.


Replacements :

Date/Time widget : replace MattKruse stuff with Dojo's implementation.
Help  validation popups: replace MattKruse stuff with a new Dojo
implementation.
Tabs: replace with Dojo Tabs
RichText: replace htmlarea with Dojo Editor, using a new fi:styling,
so htmlarea can still be used
MultiValue Editor: re-implement as a Dojo widget
MultiList (OptionTransfer) Selector: replace with a new Dojo widget
Maps: not even sure our current one is working, replace with Dojo
(Yahoo and Google)


Possible Additions :

Easy graphically rich buttons, dialogs, menus etc.
Charts to plot user data
Colour picker
Re-sizable textarea
Sliders: graphical selector for number ranges etc.
Spinner: adjust values up and down with ± buttons
Validation: plug in client-side validation where common datatypes
exist between CForms and Dojo, make new ones

Issues:

We need to do this work in such a way that has the absolute minimum
impact on existing cforms projects.
eg. adapting Dojo widgets to work the CForms way and not the other
way around.

We need to make it easier to customise the style, layout and
behaviour of our supplied widgets.


We are probably going to have issues with i18n and l10n.
Dojo is only just starting to deal with this area, while CForms has
always delt with it.
Dojo, performs i18n on the client instead of on the server as cforms
does.
Very few of Dojo's widgets are i18n enabled yet.
Dojo uses a different format of i18n message dictionary than we do
(all of this is kind of obvious).

Most of the work above will involve either extending existing dojo
widgets or making new ones.
We ought to be adding i18n/l10n as we go along.

We need to decide whether we want to keep our message dictionary
format (transforming it on the fly for dojo) or start using Dojo's
format (for widget internals).


What did I miss out ?


I hope this whets your appetite !

Let's get on with the replacements first !!

WDYT ?

regards Jeremy








Re: RFC: CForms Roadmap

2007-01-09 Thread Bruno Dumon
Hi Jeremy,

On Sun, 2007-01-07 at 13:01 +, Jeremy Quinn wrote:
 Hi All
 
 Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid  
 platform to complete the modernisation of CForms client-side code.

snip/

 
 Replacements :
 
 Date/Time widget : replace MattKruse stuff with Dojo's implementation.

I'm going to need this in the next couple of weeks, so it's very likely
I'll work on this.

 Help  validation popups: replace MattKruse stuff with a new Dojo  
 implementation.

I've got something like this in Daisy already, it seems to work quite
well, so I could move it into CForms.

 MultiValue Editor: re-implement as a Dojo widget

This is also one I'm motivated to work on, since I wrote it originally.

I still need to get up-to-date with dojo 0.4 and the current cforms, but
if all goes well I hope to be able to do something about these 3.

-- 
Bruno Dumon http://outerthought.org/
Outerthought - Open Source, Java  XML Competence Support Center
[EMAIL PROTECTED]  [EMAIL PROTECTED]



Re: RFC: CForms Roadmap

2007-01-09 Thread Gabriele Columbro

Hi Jeremy!

First of all thx for the great job you did on the modernization of CForms,
I didn't have the proper time to investigate it or better to use it
(hopefully I will do that this or the next week), but reading from ML
threads seems to be quite cool stuff!

On 1/7/07, Jeremy Quinn [EMAIL PROTECTED] wrote:


Hi All

Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid
platform to complete the modernisation of CForms client-side code.



snip/

I hope to be working on this stuff, you would be very welcome if

you'd like to join in 



I  think I can be active on many of the improvements you mentioned, but
first I have to investigate a bit more, as I was saying, the dojo
0.4integration with CForms. Maybe you can pin-point me ( again ;-) )
to a
particular thread or some docs, or even a crucial code fragment you wrote,
to ease the understanding. Or better I have just to dig into code...

If you would like to take on something from this list (or something I

missed out) please discuss it on the dev list so no-one is
duplicating effort.



snip/

Maps: not even sure our current one is working, replace with Dojo

(Yahoo and Google)



I've been recently working quite hard with GMaps and Dojo, I can surely help
in this task, and I'm pretty motivated to work on that.  Expect me coming
with questions as soon as I studied your code.


Possible Additions :


snip/

Validation: plug in client-side validation where common datatypes

exist between CForms and Dojo, make new ones



This is also very interesting, and there's been interest in the community on
that in the past. I may be able to contribute also on this point. But as you
correctly say, let's go for the replacements first.


Issues:


We need to do this work in such a way that has the absolute minimum
impact on existing cforms projects.
eg. adapting Dojo widgets to work the CForms way and not the other
way around.

We need to make it easier to customise the style, layout and
behaviour of our supplied widgets.



+1000

In addition to being pretty configurable in the layout, I would really see
CForms solve another tedious problem I have to face while working
especially in government projects with high requirements of accessibility:
some design is still table-based in the forms styling xsl, as well as
propietary attribute are sometimes coming out in the html (say ajax = true,
or dojoType = CFormsForm, maybe the latter is changed), preventing any site
to be:

1. completely W3C compliant
2. completely accessible (WCAG 1.0 and 2.0) and easily customizable (tables
are somewhere used for design purposes, and the absence of div in those
place limits the customization power of CSS, and prevents from having a
highly configurable default using this [1]) .  See, I'm not looking at
CForms deeply since a 2/3 months, so maybe everything is already changed,
but, according to me, a key point in the success of CForms would to have a
sensible default behaviour that automagically assing CSS (multiple) classes
so that the forms-styling.xsl can be almost ignored during development,
and widget Id's can be used as the only, ultimate, contract between
developers and designers.
WDYT? Can it be a valid improvement in this refurbishment roadmap?

snip/



I hope this whets your appetite !




Sure it does ;-)


Let's get on with the replacements first !!


Yep, hopefully I can come again with more appropriate questions of proposals
in a 10 days time.



regards Jeremy



Ciao,
Gab


[1]
http://mail-archives.apache.org/mod_mbox/cocoon-dev/200605.mbox/[EMAIL 
PROTECTED]



-
Eng. Gabriele Columbro
Consultant at Sourcesense Italy
-
work: [EMAIL PROTECTED]
private: [EMAIL PROTECTED]
mobile: (0039)3201612846

yahoo: g.columbro
gtalk: [EMAIL PROTECTED]
AIM:   gabrielecolumbro

-
Keyboard not found.
Press F1 to continue
-


Re: RFC: CForms Roadmap

2007-01-09 Thread Jeremy Quinn

Hi Bruno

Thanks for the heads-up.

On 9 Jan 2007, at 09:06, Bruno Dumon wrote:


Hi Jeremy,

On Sun, 2007-01-07 at 13:01 +, Jeremy Quinn wrote:

Hi All

Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid
platform to complete the modernisation of CForms client-side code.


snip/



Replacements :

Date/Time widget : replace MattKruse stuff with Dojo's  
implementation.


I'm going to need this in the next couple of weeks, so it's very  
likely

I'll work on this.


Excellent.
I hacked one up recently . not well enough to commit . one  
possible issue could be problems with l10n of the dates output by the  
widget. It was outputting US-style dates, I was probably doing  
something wrong.



Help  validation popups: replace MattKruse stuff with a new Dojo
implementation.


I've got something like this in Daisy already, it seems to work quite
well, so I could move it into CForms.


Great !
I was thinking of using a Dialog class as this would let users turn  
on cool lightbox effects etc.



MultiValue Editor: re-implement as a Dojo widget


This is also one I'm motivated to work on, since I wrote it  
originally.


Cool again ;)

I still need to get up-to-date with dojo 0.4 and the current  
cforms, but

if all goes well I hope to be able to do something about these 3.


Great news!

I went through all of the implementations of the Cocoon-supplied Dojo  
Widgets, trying to clean up their usage of the Widget API. So  
hopefully they provide a reasonably set of clean examples for you.



Have fun ;)

regards jeremy



smime.p7s
Description: S/MIME cryptographic signature


Re: RFC: CForms Roadmap

2007-01-09 Thread Max Pfingsthorn

Hi everyone,

This is very interesting! I would love to help, but I have like 5 
projects going on in the university and otherwise... However, two of 
them are geared at the web, so this might be interesting to investigate.


By the way, has anything improved authentication-wise in Cocoon? I 
haven't quite followed the development in the last few months, but with 
the real blocks and so on, it's getting more interesting for the kind of 
web application I have in mind. (for which I would use struts2 otherwise 
*shiver*)


Anyway, I can't commit much time right now, but I'll definitely have a 
look when I get home!


Bye, and take care!
max

Jeremy Quinn wrote:

Hi All

Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid 
platform to complete the modernisation of CForms client-side code.


I hope the main outcomes of this will be :
Leaner: only the resources that are used will be loaded by cforms
Richer: more interactive widgets with wider x-platform support
Cleaner: eliminate most of the messy script tags scattered through 
our forms
Simpler: use more html templates to simplify our xslt (and simpler 
to adapt the widgets)


Here is a list of specific goals I would love us to achieve for the next 
release.
I hope to be working on this stuff, you would be very welcome if you'd 
like to join in 
If you would like to take on something from this list (or something I 
missed out) please discuss it on the dev list so no-one is duplicating 
effort.



Replacements :

Date/Time widget : replace MattKruse stuff with Dojo's implementation.
Help  validation popups: replace MattKruse stuff with a new Dojo 
implementation.

Tabs: replace with Dojo Tabs
RichText: replace htmlarea with Dojo Editor, using a new fi:styling, so 
htmlarea can still be used

MultiValue Editor: re-implement as a Dojo widget
MultiList (OptionTransfer) Selector: replace with a new Dojo widget
Maps: not even sure our current one is working, replace with Dojo (Yahoo 
and Google)



Possible Additions :

Easy graphically rich buttons, dialogs, menus etc.
Charts to plot user data
Colour picker
Re-sizable textarea
Sliders: graphical selector for number ranges etc.
Spinner: adjust values up and down with ± buttons
Validation: plug in client-side validation where common datatypes exist 
between CForms and Dojo, make new ones


Issues:

We need to do this work in such a way that has the absolute minimum 
impact on existing cforms projects.
eg. adapting Dojo widgets to work the CForms way and not the other way 
around.


We need to make it easier to customise the style, layout and behaviour 
of our supplied widgets.



We are probably going to have issues with i18n and l10n.
Dojo is only just starting to deal with this area, while CForms has 
always delt with it.

Dojo, performs i18n on the client instead of on the server as cforms does.
Very few of Dojo's widgets are i18n enabled yet.
Dojo uses a different format of i18n message dictionary than we do (all 
of this is kind of obvious).


Most of the work above will involve either extending existing dojo 
widgets or making new ones.

We ought to be adding i18n/l10n as we go along.

We need to decide whether we want to keep our message dictionary format 
(transforming it on the fly for dojo) or start using Dojo's format (for 
widget internals).



What did I miss out ?


I hope this whets your appetite !

Let's get on with the replacements first !!

WDYT ?

regards Jeremy





Re: RFC: CForms Roadmap

2007-01-09 Thread hepabolu

Jeremy Quinn said the following on 9/1/07 13:12:

The next level of complexity, is all of the groups and layout stuff in 
the cforms xslt.
In hindsight, it seems this could have been done cleaner in a separate 
namespace, but we do not have that option now, unless we want to force 
everyone to completely re-work their form templates etc.


However, I do find lots of inconsistencies with the layouting code . 
for instance, I have not found a way to combine the layouting tags with 
stuff like repeaters and as you say, there is possibly too much usage of 
tables.


I would love to see more of the layout structure use divs and css, but 
this was not done originally I suspect as these types of layouts are 
more difficult to achieve.


I once started (way back when 2.1.7 or 2.1.8 was released) to convert 
the XSLT to divs and CSS, replacing the tables. I ran into problems with 
the autoformat settings like columns and rows, because there is no 
way to get that flexibly working without tables when you have no clue 
what type of widget it is and how long the label is.


Also: fieldset is rendered differently in different browsers, that would 
mean you have to apply CSS hacks for all possible browsers thus moving 
the complication from XSLT to CSS.


Just my 2ct.

Although I'd love to tinker with this some more.

Bye, Helma



RFC: CForms Roadmap

2007-01-07 Thread Jeremy Quinn

Hi All

Now that Cocoon 2.1.11-dev runs Dojo 0.4.1, I think we have a solid  
platform to complete the modernisation of CForms client-side code.


I hope the main outcomes of this will be :
Leaner: only the resources that are used will be loaded by cforms
Richer: more interactive widgets with wider x-platform support
	Cleaner: eliminate most of the messy script tags scattered through  
our forms
	Simpler: use more html templates to simplify our xslt (and simpler  
to adapt the widgets)


Here is a list of specific goals I would love us to achieve for the  
next release.
I hope to be working on this stuff, you would be very welcome if  
you'd like to join in 
If you would like to take on something from this list (or something I  
missed out) please discuss it on the dev list so no-one is  
duplicating effort.



Replacements :

Date/Time widget : replace MattKruse stuff with Dojo's implementation.
Help  validation popups: replace MattKruse stuff with a new Dojo  
implementation.

Tabs: replace with Dojo Tabs
RichText: replace htmlarea with Dojo Editor, using a new fi:styling,  
so htmlarea can still be used

MultiValue Editor: re-implement as a Dojo widget
MultiList (OptionTransfer) Selector: replace with a new Dojo widget
Maps: not even sure our current one is working, replace with Dojo  
(Yahoo and Google)



Possible Additions :

Easy graphically rich buttons, dialogs, menus etc.
Charts to plot user data
Colour picker
Re-sizable textarea
Sliders: graphical selector for number ranges etc.
Spinner: adjust values up and down with ± buttons
Validation: plug in client-side validation where common datatypes  
exist between CForms and Dojo, make new ones


Issues:

We need to do this work in such a way that has the absolute minimum  
impact on existing cforms projects.
eg. adapting Dojo widgets to work the CForms way and not the other  
way around.


We need to make it easier to customise the style, layout and  
behaviour of our supplied widgets.



We are probably going to have issues with i18n and l10n.
Dojo is only just starting to deal with this area, while CForms has  
always delt with it.
Dojo, performs i18n on the client instead of on the server as cforms  
does.

Very few of Dojo's widgets are i18n enabled yet.
Dojo uses a different format of i18n message dictionary than we do  
(all of this is kind of obvious).


Most of the work above will involve either extending existing dojo  
widgets or making new ones.

We ought to be adding i18n/l10n as we go along.

We need to decide whether we want to keep our message dictionary  
format (transforming it on the fly for dojo) or start using Dojo's  
format (for widget internals).



What did I miss out ?


I hope this whets your appetite !

Let's get on with the replacements first !!

WDYT ?

regards Jeremy





smime.p7s
Description: S/MIME cryptographic signature