[
Lists Home |
Date Index |
Thread Index
]
Didier PH Martin wrote:
>
> As usual, in this list there is a lot of talk and very little walk. Here is
> a challenge I am proposing to the member of this list and particularly to
> the people who think great application can be done with a browser technology
> and W3C frameworks.
>
>
>
> Here it is:
>
> The goal is to get an XML document to be edited on the client, all
> validation performed on the client and the filled XML document returned to
> the server. Let's start with a modest document as below. What is to be
> filled are the data content for each element. Simple no?
>
> So to recap:
>
> a) The XML document is associated with a stylesheet for rendition.
> All the element's data content are empty. The XML document is rendered in
> the browser using an XSLT stylesheet (you design and implement the
> stylesheet - rendition is opened)
>
> b) The user enters the data and the XML data content is filled. The
> element Version has to be validated. Only 1.0 is allowed. Validation
> obviously is performed on the client side.
>
> c) The XML document is returned to the server with all its data
> content filled
>
I have done this in several prototype and demo projects two years ago
and more. Here is the method I used in one -
a) Styled an xml file in the browser with an xslt stylesheet, creating
an HTML form, and used javascript to build data structures and to
populate list boxes that the user used to select allowable data values.
b) Used javascript to validate data selections.
c) When the user asked to send the form back to the server, I walked
form and built up an xml document using javascript, then POSTed it to
the server.
On another one of the projects, we had a workflow situation where, at
each station, the user modified or added data in an HTML form. Under
the hood, the form had been created by styling an xml file. Each
station recreated the xml file with the new data before passing it along
to the next station. Each station had a completely different style
sheet, so each operator saw a very different view of the same data.
We did not actually send a modified file directly to the next station.
Instead, it was stored locally, and the workstation notified the next
one in line that there was a document waiting - a very poor man's
message queue-like system. It did that by posting a message to a server
on the next workstation, which then made an HTTP request to get the file
from the previous one. Very RESTful.
- References:
- Challenge
- From: "Didier PH Martin" <martind@netfolder.com>
|
