- From: Gavin Nicol <gtn@ebt.com>
- Date: Fri, 21 Jun 1996 05:16:10 GMT
- To: erik@netscape.com
- Cc: JMHX.DSKPO33C@dskbgw1.itg.ti.com, www-international@w3.org
>You need 2.02 or higher. Look for "httpAcceptLanguage" in the *.ad file. >That is the X resource that you need to set. We haven't created a GUI >for this yet. How about Windows? This alone would help the sniffing logic a great deal. Actually, the combination of Accept-Charset and Accept-Language would provide enough information so that you could get close to the results of correct labelling (I do not say they should replace it!!!). >We want to avoid cluttering the UI with options that the average user >doesn't understand. I suppose we could put it in a config file. I would recommend having language, encoding, and other such things in the option GUI. If they are expected to understand proxies, they should be able to understand this. >> What happens if you have, on a single site, many different forms in >> many different encodings? What happens if the forms are dynamically >> generated, where you do not know a priori what the encoding of the >> form is/was? > >I guess we haven't come across many cases like this. One reason is because it's hard to do now, as are truly multilingual sites. This *must* change. Such cases *will* increase. There is a definite trend toward managing web data in database *especially* in large corporate intranets. There's a business tip for you. >>Data sniffing would also be simplified if a single encoding was >>choosen for each language. > >The fact is that more than one encoding is used for some languages. It >would be nice if we could get the servers to use fewer encodings without >affecting the client users (customers). Sure. Initially, I was under the impression that Navigator sent ISO-2022-JP to the server (kind of makes sense..) and would have preferred that to the current scheme. The current scheme *can* be made to work, even to work fairly well, but as the number of possible encodings for data sent to the server increases, the less of a chance you have of getting meaningful results from the sniffing logic. Because the vendors almost uniformly do not do the correct thing, anyone attempting to create a truly multilingual site will have to implement thousands of lines of code more than they would otherwise have to, and it will *still* not always work. This is what I have wanted to have fixed since so long ago. One word: interoperability. >Extended_UNIX_Code_Fixed_Width_for_Japanese is not the same as EUC-JP. >That is the wchar_t form. What you're thinking of is >Extended_UNIX_Code_Packed_Format_for_Japanese. My mistake.
Received on Friday, 21 June 1996 01:18:15 UTC