Re: [CSS21] [CSS3 Text] Should 'text-transform: uppercase' apply to input type="text" (text entered, typed in text field by user)?

(12/06/26 23:06), Sylvain Galineau wrote:
> [Kang-Hao (Kenny) Lu:]
>> (12/06/26 1:45), Sylvain Galineau wrote:
>>> [Christoph Päper:]
>>>> With 'text-transform' some UAs ignore the code/style distinction already:
>>>> When you copy text that was case-changed through CSS from a browser
>>>> into a plain-text environment, it will often be pasted with the
>>>> casing displayed in the browser instead of the one stored in the
>>>> source code. I strongly believe this is just as wrong as not copying
>> "display: none" parts to the clipboard.
>>>
>>> I'm not sure why that would be wrong, especially from the point of
>>> view of an end user. If someone copies/paste something from a web page
>>> into their email client and the case changes they are imo far more
>>> likely to be surprised and consider it a bug than to think 'oh thank
>>> God the browser preserved the state of the markup instead'.
>>
>> Speaking for end users, wouldn't it be very confusing if what's sent over
>> the wire is different from what's shown to the user (when script is
>> disabled)? Since we are likely to have more 'text-transform' features in the
>> future, I kind of think we should not allow authors to trick users like
>> this.
>>
> I'm not sure I follow how disabling script would impact the application of 
> text-transform? I don't know when, how or whether it would be confusing. It all 
> depends on how the feature is used; anything can be confusing if it's misused 
> and there is a multitude of ways to 'trick' users. 

I was trying to claim that a user should be pretty safe if she/he is
browsing the Web with script disabled. Apparently I have likely missed
lots of 'tricks' that can be carried with CSS+HTML alone, like
@font-face + 'font-family' + <input> as suggested by Jonathan.

> If the author applies text-transform to form input but no longer
> applies it when the value is displayed back to the user, is the
> problem with text-transform?

I am not sure I understand this...

> Conversely, should we require text-transform to affect the value that is submitted
> then I suspect it won't be very reliable in practice e.g. because a stylesheet didn't
> load, a number of existing UAs may not apply this property to the content of form 
> controls but applying it to existing forms could break applications that did not 
> expect it to apply to form controls or affect the field value etc.

I agree.

> Overall I think of this as a display-only feature pending use-cases that prove otherwise.
> And even then I would be concerned about the compat impact.

I can't quite believe there's a use case of applying 'text-transform' to
an input control. I guess browsers just implement this for consistency
and I don't have a strong argument otherwise, but this behavior is still
weird to me.


Cheers,
Kenny

Received on Tuesday, 26 June 2012 15:35:31 UTC