I agree with you.  I wasn't suggesting using HTML as a model. Any good 
serialisation language like JSON would make sense.  YAML perhaps. Lisp not 
really a fan but wouldn't  loose sleep over. 





________________________________
From: Daniel Danilatos <[email protected]>
To: [email protected]
Sent: Thu, 23 June, 2011 9:39:55
Subject: Re: Wave richtext editor

Hi Paul,

There's a few things to mention here. Firstly, wave doesn't really use
XML - it uses nested lists with attributes and standoff annotations.
The transform logic is all about list/text insertion/deletion
behaviour. It just so happens that you can represent that as an XML
string if you like (though you need to resort to some non-standard
tricks with processing instructions), which appeases people that like
XML. Perhaps in our debug output we shouldn't be rendering it as XML
but as some other format that looks a bit more like json or lisp, to
avoid the confusion people keep having in trying to think of it as
HTML.

The advantages of HTML5 over XML as far as the W3C is concerned are
not really relevant here. Wave's "XML" data model is NOT a rendering
format like HTML, it is an arbitrary data model, that happens to be
used a lot for rich text, but it is also used for many other forms of
structured data. It can be rendered however you like, and often we
happen to render it to HTML. In fact, the mapping is non-trivial. HTML
is also not amenable to an OT environment, details on this are
plentiful in the mailing list.

Regards,
Dan

Στις 23 Ιουνίου 2011 6:28 μ.μ., ο χρήστης Paul Thomas
<[email protected]> έγραψε:
> This has been covered before XML is too clunky and verbose. Really I don't see
> the issue once documentation is there. You are talking about a wave document 
>not
> a layout. XML was caused wasted years at W3C IMO, till they finally had some
> sense and embraced HTML5. I don't have a problem with XML in principle, but it
> became too much of an ideology and never lived up to it. the solution to
> everything was more complexity.  In fact the XMLness of wave document, if
> anything made things much more complicated. There was a reason why they didn't
> use pure XML and XSLT for transformations. It would have been a waste of time,
> it is not fit for live editing, the algorithms are complicated enough.
>
>
>
>
>
> ________________________________
> From: Anton Starcev <[email protected]>
> To: [email protected]; [email protected]
> Sent: Thu, 23 June, 2011 8:45:18
> Subject: Re: Wave richtext editor
>
> P.S.
> XML also be approached, but without namespaces
>
> On Thu, Jun 23, 2011 at 11:42, Anton Starcev <[email protected]> wrote:
>> On Thu, Jun 23, 2011 at 11:12, David Hearnden <[email protected]> wrote:
>>> When you say "users", do you mean developers?  Or actual end users?
>>
>> I mean front end developers of the site. They make html layout and, in
>> some cases, they might want to look in source.
>>
>>> Can you give some more detail on the scenario you're describing?  I might be
>>> able to provide a better answer if I understood it a bit more.
>>
>> Main scenario is html layout and editorial changes in content management
>>system.
>>
>>> On Thu, Jun 23, 2011 at 4:52 PM, Anton Starcev
> <[email protected]>wrote:
>>>
>>>> Hi all,
>>>>
>>>> Keep doing standalone editor based on wave editor.
>>>> Have a problem with source code of editor's text, sample output is:
>>>>
>>>> <doc>
>>>>  <body>
>>>>    <line/>
>>>>    <l:p>
>>>>      <text>text</text>
>>>>    </l:p>
>>>>    <line/>
>>>>    <l:p>
>>>>      <text>bo</text>
>>>>      <l:s fontWeight='bold'>
>>>>        <text>l</text>
>>>>      </l:s>
>>>>      <l:s fontWeight='bold' fontStyle='italic'>
>>>>        <text>d</text>
>>>>      </l:s>
>>>>    </l:p>
>>>>    <line t='h3'/>
>>>>    <l:p t='h3'>
>>>>      <text>headlin</text><text>e</text>
>>>>    </l:p>
>>>>    <line/>
>>>>    <l:p>
>>>>      <l:s fontFamily='tahoma,sans-serif'>
>>>>        <text>text</text>
>>>>      </l:s>
>>>>    </l:p>
>>>>  </body>
>>>> </doc>
>>>>
>>>> Tags like "l:p", "l:s" etc are too difficult for potential users (they
>>>> expect something like plain html).
>>>> Is in this some hidden logic, or I can just replace names in Pretty<N>
>>>> and parse class?
>>>>
>>>>
>>>> Anton Startsev
>>>> ______________________________________________________________________
>>>> [email protected]| www.artlebedev.com
>>>>
>>>
>>
>>
>> Anton Startsev
>> ______________________________________________________________________
>> [email protected]| www.artlebedev.com
>>
>
> Anton Startsev
> ______________________________________________________________________
> [email protected]| www.artlebedev.com
>

Reply via email to