Language, Charset, etc

Forum Moderators: open

Message Too Old, No Replies

Language, Charset, etc

What other assumptions are we making?

Dave McClure

5:37 am on Nov 9, 2004 (gmt 0)

Recent experience writing Arabic pages has made me question a few assumptions. I'd be interested in any thoughts on this. For an Arabic page to work at all, you must define the charset in the header, you must also define the language (lang="ar") because the same character set is rendered differently for different languages, and you must define the direction, right to left. Technically, I suppose for an English language page we should make similar definitions (and for validation charset is recommended), but most pages out there don't bother. My question is - are there any other areas where we should be future-proofing by rigorously stating things that we often just assume, that might fall in a heap in future browsers?

tedster

6:52 am on Nov 9, 2004 (gmt 0)

Good topic! - of course I'm biased because I am also working with some non-English sites at present, including several Asian languages and Hebrew.

One thing people assume is handled, but it may not be, is their server headers - specifically Content-Type and Content-Language.

Yes, your meta tags are "http-equiv" and can substitute or override the server header. But why not put the character set and language explicitly in your HTTP Header in the first place? Who knows what problems you might avoid by being thorough?

I've noticed that the lang attribute is often missing from the <html> element, even on non-English university and newspaper sites. I still agree it's a good idea to include it.

Dave McClure

10:12 am on Nov 15, 2004 (gmt 0)

Sorry for starting this topic then dropping it - we have this public holiday, Eid al Fitr, to mark the end of Ramadan. I've had no access to my office for a few days.

Anyway, thanks for reply. For years, my headers contained nothing but <title>, then I started adding description and keywords. It's only recently I've been going for validation and been pointed to doctypes and charset. But your reply tells me I don't really understand what http-equiv means. Can you suggest any threads that go into the meat of these header tags? I definitely need to brush up. As per previous message, the importance of getting it right really comes home in bilingual or RTL sites.

tedster

1:06 pm on Nov 15, 2004 (gmt 0)

Before a server sends out the packets with your HTML, CSS etc, it sends out an HTTP header. The http-equiv meta tags are a way of sending similar information in your HTML page that would otherwise be expected in the server header.

HTTP 1.1 has many more possibilities for header tags related to language.
See: W3C reference [w3.org]

So if you have enough access to your server to set this header information, I'm suggesting you work with these tools as well, and not just meta tags.

This is not really browser-side stuff, and it's a little out of the range of this forum. But I thought I might point you, and anyone reading the thread, to other issues that exist around languages and are worth digging into.