- From: Ernest Cline <ernestcline@mindspring.com>
- Date: Sun, 22 Feb 2004 11:20:19 -0500
- To: "Boris Zbarsky" <bzbarsky@MIT.EDU>
- Cc: "WWW Style" <www-style@w3.org>
> [Original Message] > From: Boris Zbarsky <bzbarsky@MIT.EDU> > To: <ernestcline@mindspring.com> > > > A possibility, but I would think that in those two cases, a UA would be > > likely to possess out-of-band knowledge about such sheets > > Why? Some typical UI for user sheets that I can think of (and actually want to > implement) is just a filepicker/textfield that lets the user pick a file or > paste a URI for the sheet.... As I said, it might well need to ask the user the first time it loads the file, and it should in my opinion, if it is unable to determine the charset by the other rules. > > Any other cases? > > Tools that are loading a sheet for the sake of loading a sheet. > Validators come to mind. So do certain types of GUI CSS editors. I'd consider CSS editors to be much the same as a browser in this regard. Since it will be working with such a file on a regular basis, I would hope that when saving the file it would include sufficient info, to enable the sheet's character encoding to be determined by the other rules instead of by assuming. As for reading in CSS files the first time, I'd rather have it ask what charset than just simply assume. As for validators, if it ever has to assume, I definitely would want it to warn the user by default of such an assumption, no matter what charset it assumed. Still this is now getting into the realm of SHOULD and not MUST, and since not all such examiners of CSS will, I can see the need for a default assumption, altho I would like a statement added so that rule 4 should read: 4) If all else fails, assume UTF-8, but warn the user that assumption is being made.
Received on Sunday, 22 February 2004 11:20:21 UTC