[Laszlo-dev] LzRichInputText Proposal

[Sarah Allen]

Cort,

Thank you for sharing your proposal with the community.  A link has been 
added to the enhancement proposals page 
(http://www.openlaszlo.org/wiki/Enhancement_Proposals) under "Pending 
Proposals".  Please move your proposal to the wiki where it can be kept 
as a living document during the development of this feature.

This is a feature that received many requests and I hope that this will 
provide an opportunity for further public comment.  

Sarah

Cortlandt Winters wrote:

>Hi Folk,
>
>This is a proposal to add a RichInputText class to the laszlo foundation
>classes. Please feel free to comment as bluntly as you'd like. I've thought
>about this way more than I should have, but there are still things I'm not
>quite sure about. In particular I need to create some test files that
>demonstrate the kinds of document style it would take as inputs and outputs
>and a simple grammar to describe it, but this is already large enough that
>it should probably be sent on it's own. I'll also send another email with
>some interesting notes that I came up with when doing background research
>into what this should be like.
>
>Proposal to add a richinputtext tag to the laszlo foundation classes
>
>Rational
>Currently, though lzx has a lot of support for styleing static text, there
>is little support for styleing input text. That is, you can style an entire
>textbox, but not individual sections of text. This is a vestigial effect of
>having used the Flash5 player until recently. In order to support things
>like content management systems, a finer grained text styling method is
>needed for input text.
>
>There are two different ways that this could be done. We could edit the
>existing LzText6 (the text tag) and LzInput6 (the inputtext tag) classes to
>support additional calls to support rich lnput text. Or we could create a
>new Laszlo Foundation Class to cover rich input text.
>
>After attempting the first approach for a couple of weeks it seems to me to
>be a good idea to try the second approach. Text is so key to everything
>throughout the platform and it has just gone through some major changes.
>It's tough to make changes to the text classes without creating ripple
>effects and since editable html text hasn't been officially supported
>before, it seems like a good place to set up a clean break from the past in
>order to keep things simple.
>
>I think it will probably also reduce confusion to seperate this into it's
>own class. Currently a <text> field won't be an html text field but will use
>html tags to style it's appearance at the implementation level. An input
>text field similarly uses it's parent' text field's html tags to style it,
>but is not an html input text field. The underlying html tags (formattting)
>are applied to the entire field at once to style it after any change. Thus
>the underlying html and input properties of an underlying textfield can both
>be true and yet it's still not supposed to be an html input textfield. Thus
>far, that's not really very confusing, it's just an implementation detail
>But adding, say, an htmlinput property or a richinput property to flag that
>this field should now be rich input text, seems to me to invite a little
>confusion.
>
>This proposal is for a new class LzRichEditText (with a richedittext tag)
>that should at least allow the styleing of text with as many traditional
>rich text syles as possible as a build in feature, with the possible
>additional capability of allowing folk to create their own tags and
>attributes if time allows or in the future.
>
>So what do we need?
>We need a field with an api that we can use to style individual sections of
>text independently from one another, and to allow selecting blocks of text
>based on mouse movements,  the cursor position or an api call.
>
>A set of styles that we can support out of the box would be
>
>=================================
>LzTextStyle Class
>=================================
>	Attributes
>	-------------------
>	bold (bool)
>	underline (bool)
>	italic (bool)
>	fontface (string)
>	textcolor (color)
>	textsize (int)
>	listitem (bool)
>	leftmargin (int)
>	align (string: "left" | "right" | "center")
>	href (string: "a url")
>	target (string: "_blank", "_self",  "_parent", "_top")
>
>
>The RichInputField might inherit from LzText or LzInputText or it might not
>there are some questions to consider like will it  make the LFC
>unnecessarily bigger to have it on it's own and yet duplicating some
>functionality. But "first do no harm", so it will be created first as
>independent and later on it might be good for someone with a better brain
>and more experience with LFC than me to use them.
>=================================
>LzRichInputText Class
>=================================
>	Attributes
>	-------------------
>	resizable
>	embedfonts
>	label
>	maxlength
>	multiline
>	selectable
>	richtext
>	plaintext
>
>	-------------------
>	Methods
>	-------------------
>	parseRichText(rich text document)
>	textstyle <=  getDefaultTextStyle()
>	setDefaultTextStyle( textstyle )
>	textstyle <=  getTextStyle(int)
>	setTextStyle(int, int, textstyle)
>	setSelection(int, int)
>	insertImg(path to swf or jpg, int)
>	int <= getSelectionPosition()
>	int <= getSelectionSize()
>	int <= getRangeStart(int, a list of one or more textstyle properties)*
>	int <= getRangeEnd( int, a list of one or more textstyle properties)*
>	richtext <= updateData()
>
>	Question: Is there a way for updateData to update both richtext and
>plaintext simultaneously so different folk can bind to different purposes?
>Does that even make sense?
>
>And what would be nice?
>It would be nice if a user could create his or her own custom tags and
>attributes to their text. We are still limited to what we can display in a
>textbox visually by what the flash player can display, but we can greatly
>enhance the usefullness of the field by also allowing tags to be added that
>are not visual, but semantic. So a user could essentially create his or her
>own tags and then just assign the style attribute of the tag to a textstyle
>objects as with css. To do this we would want a dictionary of textStyle
>objects, a dictionary of tag types that you can add and subtract from and
>add attributes to, and the ability to wrap a selection with a tag.
>
>*The idea here is for an abstract way to select an innermost area of
>interest. So if the user's selection position is in a block of text and they
>click on a button to align the block to the left or the right, the
>programmer can use this to select the innermost block that the align=right
>tag would affect. So if the text field contains the equivalent of...
>
><span class={align="right" bold="false"}>dogs frogs and <span
>bold="true">polyw|ogs</span></span>
>
>....and the caret is right after the 'w' of polywog, calling
>getRangeStart(getSelectionPosition,"bold") would give us the index of the p
>in polywogs.getRangeStart(getSelectionPosition,"align")would give us the
>index of the d of dogs. If there are two or more attributes listed, it gives
>us the innermost consistent selection.
>
>I still need to map out how that would get sent out as a document, would it
>be ...
><p align="right">dogs frogs and <b>polyw|ogs</b></p> ?
>
>Looking forward to hearing responses.
>
>Thanks for your help and thoughts,
>
>-Cort
>
>_______________________________________________
>Laszlo-dev mailing list
>Laszlo-dev at openlaszlo.org
>http://www.openlaszlo.org/mailman/listinfo/laszlo-dev
>  
>