Emile wrote:
> Michiel wrote:
> 
> > I have no real clue how to communicate in some comprehensible way this
> > kind of information to the client, for javascript impelmentaiton of the
> > same or so.
> >
> 
> This is where I figured the client-side XML would earn it's keep.
> 
> > I figure that much of this stuff will for some time remain to be
> > customized on the JSP's/javascript. Or would you have hope to define
> > these kind of things genericly?
> >
> Yes, to some or other degree.  Simply starting with a relationship type or
> definition of sorts (user-define or standard) would go a long way in
> providing generic logic and variants thereof.  I'd suggest that javascript
> may earn a space in CSV before long.

Yes, I agree it may be a good idea to provide some javascript library to enforce
mmbase specific constraints on the client side, and which understand how
MMBase gives this meta-info (perhaps as an XML indeed).

> 
> > Perhaps it is easiest to talk back to the server with request for
> > validity only, which it then can report, without crashing an exception
> > to the poor user.
> >
> That's a simple validation exercise, 

Yes, simple, and therefore a nice start.. ;-)

> but the more scalable option would be
> the inclusion of generic (and specific) client-side logic, which could then
> also manage some degree of business rules and complex validation.
> 

We certainly have to keep that  possibility in mind. We includes also
you btw, so if you want to contribute something more concrete on this
area, please feel free :-) I myself am IIRC the leader of the
'fieldtypes' project, and field-value validation in the broadest
possible sense I consider in the scope of this project. I was about to
add a simple server side 'check' function for that, but I totally agree
that that would only be a start, and that client-side validation is
preferable as much as possible.

I think it never can be complete, btw. For example one of the already
possible constraints on a field-value is uniqueness, for which I see no
other way then server-side checking, because it is of course out of the
question to send all existing values to the client... We can of course
agree that the current 'check' is too lame (i.e. simply confronting the
client with a database exception...)

Michiel


-- 
Michiel Meeuwissen                  mihxil'
Mediacentrum 140 H'sum                [] ()
+31 (0)35 6772979         nl_NL eo_XX en_US



_______________________________________________
Developers mailing list
[email protected]
http://lists.mmbase.org/mailman/listinfo/developers

Reply via email to