Re: Validity of System and User Names


RadSolution
 

FWIW I agree that _changing_ something which the user doesn't / can't know about is a _bad_ thing...

IMHO it is a much better user-experience if the 'faulty' value is rejected - as long as the user knows / is told WHY it was faulty - and the user learns and then enters a correct value.

This doesn't remove any requirement for validation; it just removes silently changing things.

In the above, WHY shouldn't be limited to "invalid input" - it should, for example, say "a name may not start or end with a space character"...

Regards, Dave Radcliffe

On Thursday, 23 May 2019, 06:02:14 BST, Bob Jacobsen <rgj1927@...> wrote:


The following is why the entire normalization paradigm is broken.

Ken, if you really believe this is a problem, then you’ve just advocated for removing normalization (the _changing_ of what was entered), training and trusting users to enter proper-format names, and rejecting anything that’s not in proper format.

Let’s go back to that and end all the normalization madness (but not the syntax error checking).

Bob

> On May 22, 2019, at 7:18 AM, Ken Cameron <kcameron@...> wrote:
>
>
> I'm thinking if you do the create, you have a handle on the bean. You use
> that to ask what the system or user name is for that object. You compare
> that with what you asked for, or just update your side with what it said it
> created, and now you would know if or what changed.
>

--
Bob Jacobsen
rgj1927@...





Join jmri@jmri-developers.groups.io to automatically receive all group messages.