Michiel Meeuwissen wrote:
Ernst Bunders wrote:
so, if i am correct, the mapping between field types and actual input
fields is undefined and arbitrary? that seems like a bit of a hole in
the fied types specification. has it never been considered to define
this aspect as well?
I do not understand that conclusion.
The field- or datatype only defines the field, and tries to provide all necesary
information for creating input fields.
So, it does not say 'create a text-aray' or 'create a <input
type="string"'. I figured that the correctnes of "\n" would provide that
information sufficiently, and in a logical, semantic way.
well, that is arbitrary, as wel as logical. it would also be logical to
take a line input type for every string that can only be 40 or 50 chars
long. But that's not really the point.
The point is that the specification dous not allow you to define the
interface type for a given presentation medium (html, swing, is there an
other..?). So the result is that the mmeditors, the my_editors and the
wizards might produce different interface elements for the same
fieldtype. Which is not a problem, becouse they still have to follow the
rules that define the field type, but still it is not consistent.
Not a real problem, but i am just a bit surprised. I have not fully
delved into the fieldtype code and specification yet, but i guess i had
more or less expected something like this.
The funny thing is that now using the 'guitype' element in the builder
configuration, you can define a lot about a field, except the actual gui
type :-)
ernst
So, the precise mapping is undefined, indeed, because naturaly it
depends on the presentation medium, about which the datatype must know
nothing. But it is not 'arbritrary', because there must be a sensical
relation between the datatype's restrictions and properties and the way
to present value-editors for that datatype.
Michiel
_______________________________________________
Developers mailing list
[email protected]
http://lists.mmbase.org/mailman/listinfo/developers
