Text Encoding (Big 5) not being auto-detected

Question

Level 1

106 points

Text Encoding (Big 5) not being auto-detected

I just had to reinstall OS X (10.4, the upgraded to 10.4.9) due to a bad crash. In Safari (2.0.4), I tried to view a particular page.

http://www.ettoday.com/

This is a Big-5 (chinese) encoded page; I have my Default Text Encoding set to Western (ISO Latin 1), and the page looks garbled when I try to view it, unless I go to View->Text Encoding->Traditional Chinese (Big 5), and then it displays correctly. My question is, why isn't this encoding being auto-detected and display correctly without manually setting it? Looking at the source, about 22 lines down there is this:
<meta http-equiv="ªF´Ë·s»D³ø" content="text/html; charset=big5">

which looks like the appropriate specification of the encoding. Can someone who has chinese fonts installed please verify this? Is this something Safari is doing wrong, or possibly something got lost in my system reinstall (to tell the truth, I can't recall specifically if this was a problem prior to my crash).

Thanks,
David

Posted on Apr 6, 2007 3:57 PM

Reply

Answer 1

Tom Gewecke

Level 10

115,882 points

Apr 6, 2007 4:33 PM in response to David Yeh

Looking at the source, about 22 lines
down there is this:
<meta http-equiv="ªF´Ë·s»D³ø" content="text/html;
charset=big5">

I think it has to say <meta http-equiv="Content-Type" ...> etc. The browser can't read this tag if it is not like that, so it uses your default browser setting instead. Try it in the validator and it says this too:

http://validator.w3.org/check?uri=http%3A%2F%2Fwww.ettoday.com%2F

Reply

Answer 2

David Yeh Author

Level 1

106 points

Apr 6, 2007 5:16 PM in response to Tom Gewecke

Ok, thank you. I gave up trying to understand html right around when the "blink" tag came out, and I just get paranoid every time I do a reinstall that I'm missing some critical part. I like the validator-- never knew about that.

Now, if you do a search on the source code for that ETToday page, somewhere far down you'll come to another section that reads:

<head>
<title>Untitled Document</title>
<meta http-equiv="Content-Type" content="text/html; charset=big5">
blah blah blah

This does appear to be properly formatted, and maybe it's there specficially for English-default browsers. However, perhaps neither Safari nor the validator are picking up this, or maybe it's referring to something else altogether?

thanks,
dave

Reply

Answer 3

David Yeh Author

Level 1

106 points

Apr 6, 2007 5:26 PM in response to Tom Gewecke

And, oh, not to bag on Safari, but I just tried this in FireFox which I don't use that often, and it did come up automatically with the correct encoding. Maybe this is just a matter Safari not being tolerant of some errant coding.

Reply

Answer 4

Tom Gewecke

Level 10

115,882 points

Apr 6, 2007 5:58 PM in response to David Yeh

Maybe this is just a matter Safari not being tolerant
of some errant coding.

The coding is just totally wrong -- you should never have two sets of html headings like that on one page, and only the one at the top, which they botched, counts. But they probably get few complaints because most people who want to view it have their browser default set to Big5 already. Or if they use FireFox, that browser has a special ability to automatically detect Chinese and certain other encodings even when the required charset header is missing.

Reply

Answer 5

David Yeh Author

Level 1

106 points

Apr 6, 2007 6:05 PM in response to Tom Gewecke

Great-I'll let the webmaster know, hopefully it will be fixed.

Reply