Add functionality that will allow adding shared datasource to multiple components with individual filters.

This way we will be able to have one data source and multiple grids showing different part od data.

I would like to propose that GET handlers return values that can be returned to their callers.  This would allow for things like proxy objects that can return values from other objects, without having to override the get function.  It would also allow for modification, augmentation, or replacement of values.

For examples you might find this interesting: http://millionmunkeys.net/PiMunkey/1.5/Documentation/functions.cfm#filters

Currently the grid fires the change event if the user clicks on a row. I can use this event to do something, for example, update controls in a detail part of the page below the grid.

If the user groups or sorts the grid, the grid rebinds. It does fire an event for that, but since the grid also removes any selection, to be consistent, it should also fire the same event it fires when the user selects a row. After all...the selection is changing (to nothing).

The success() function in the DataSource should check the response status. If the status is 304 "Not Modified", that indicates that the data has not changed and so the DataSource should not modify its internal data structures nor fire the "change" event.

As it stands only during a read the aggregate are fetched from the server via the dataSource. This is incomplete behavior. A grid should, if requested by the programmer,  fetch the aggregates on each update of a row in the grid. That way the aggregates will always reflect the sum(), etc.. correctly.

I need way to revert all changes for observable view model

kendoui data source should support working with arrays or arrays rather than arrays of objects
this way the json being passed between the client and server can be much smaller (and network times reduced)

it will require positional information in order to get to each field instead of name
it will be especially important for grids

Currently the default model binding of Kendo assumes that enums will be serialized as numbers. While this is the case with the default MVC JSON serializer, there are a number of problems with it (including this poor handling of enums) that make people replace the default JSON MVC serializaer with other popular ones - such as the https://github.com/JamesNK/Newtonsoft.Json.

If would be really nice if Kenod datasources could be able to map enum values in both integer and string value name format and if they were binding back the same format returned in the request.

When working with related data and odata you can you something like
http://local/get/Contact//?$expand=ContactType

Where Contact is the primary record and ContactType is n-1 to Contact also need to support 1-n

So the schema needs to be able to handle complex relations right now it is flatted out so I cant specifiy ContactType as a related entity with its own fields in the current schema implementation.

Unless the hierarchical datasource can handle this and bind to more than just the treeview widget, but I can't tell from the docs on the hierarchical datasource.  But really this is more for using this in the model definition such as kendo.data.Model.define({
Thanks,

As discussed here:
http://www.kendoui.com/forums/framework/mvvm/min-max-bindings-for-datepicker-and-timepicker.aspx

Provide a way to bind the minimum and maximum dates to a model, so that changing the model changes the min and max dates. This would be useful in creating a date range selector where the start date must be prior to the end date.

Mapping xml with namespace prefixes is not currently possible.

Basic XPath predicates should be supported (no less than is already natively supported by jQuery).

Kendo UI should have first-class support for XML.  As it stands, Xml support is currently second class to JSON.

Currently when I choose a column filter and click on the filter button, a request is sent to the server and data is bound. However, if I work with large data sources and I know in advance that I want to filter by FirstName and LastName, for example, I would like to choose both the filters first, and when filter settings are finalized, I would like to rebind the grid with a button click, for example.

This would enable to bind only once the grid, instead of n times, if the user chooses lazy filtering which would be very useful, especially in the case of large data sources

It would be nice to have the IF and IFNOT bindings like KnockoutJS has.  It would help keep the DOM cleaner instead of having to use the Visible binding.

Currently when there are any errors in data-bind attributes you usually get some array out of bound indexing error (in parseBinding()) or some other obscure error.  

This suggestion is to trap these errors more aggresively and provide better error messages.

Trying to bind to the 'id' and 'for' properties gives an error that this is not supoprted. Both are basic bindings and should be suported.

Currently if you have combos that are bound via MVVM or knockout etc and have a kendo data source on them for doing suggest lookup they will query everything without a filter to get the display value for the bound value that is set to the value property of the combo.

I should be able to pre-populate an initial set of data on the datasource which it will use when the value property is set. If and only if it isn't set, it should then go and do an odata query with filtering for the specific item and only the specific item based on the field that the dataValueField is bound to, not everything.

Further the combobox and auto-complete should cache data and searches previously done so that it doesn't have to hit the data source again in these scenarios as the user types.

Instead of using JSON to create a pie chart in Dataviz, it would be so much easier to use an SQLDataSource, So that the pie chart can be dynamically updated.

The DataSource Model including validation should come from the server and not defined within javascript where possible.  There should be two primary ways to populate the model using the DataSource widget.

1.) Retrieved from the DataSource call itself as a JSON property off of the data call.  You should be able to define a schema: { parseModel: function(data) { return data.model } } to retrieve the model directly from the data call.

2.) You should be able to specify the schema: { model: new DataSource({ ... }) }.  The model itself should be DataSourcable so you can retrieve (with ETag/Last-Modified and Cache-Control) the model as a separate request alongside the data.

Given that you should be able to retrieve the model as a separate request,  you should be given the opportunity to augment/modify the model within the client side using something like schema: { changeModel: function(model) { // Supplement Model here
} }