For non-special-character textbox, should I use inline validating message or block users from inputing invalid characters?

Question

I have one textbox that means to allow no special characters. Now, I have to choose below approaches:

1. Display inline validating message say something like "char * is not allowed" (A) or "Your input cannot contain %$*&" (B) when user input chars in black-list.
2. Block user from enter invalid chars. Meaning that if user hit chars that are in black-list, nothing would happend.

Which of the approaches should I take?

Answer 1

With approach 2 (block invalid characters), the user might type something quickly and end up with a result which looks nothing like what they wanted.

Let's say a user-name field should only contain standard unaccented characters. A user could enter Juergen without a problem; but attempting to use ü instead of ue would be ignored and the user would end up with Jrgen.

Attempting to create a username like $4nta Cl4u$ would be similarly stymied: ntaClu.

This is bad: the user would have to work out what had been omitted and why and then correct it. Approach 2 should be avoided.

Method 1 (validation message) is therefore to be preferred — and this should be inline during entry rather than returned after the form is submitted.

I would prefer adding the user's invalid characters to the message as they are entered. Thus the message You cannot use ü would appear as soon as that character was typed; and in the second example it would end up as You cannot use $4[space]. (Actually using [space], not a blank character.)

In Method 1, it would be useful to make some help available as well in order that the user can find out exactly what is valid. That could appear as a tooltip, or open a separate page.

^{download bmml source – Wireframes created with Balsamiq Mockups}