Does IE7 support it all the way? Or is it like IE6?
I have no clue the developers have never mentioned it on the IE Blog, so I sorta-kinda assume that they havent done anything about it. They really have no incentive to fix it though, because then suddenly sites like this would be broken by an IE7 upgrade remember, Microsoft is crazy about backwards-compatibility.
Are you saying that it's an issue with language packs? Because the sharfess(?) 'ss' German character is a standard inclusion in MS Word, just like a cedilla 'ç'; They are basic characters. Surely browsers can recognise this?!
No no, thats not the issue at all.
GTP has a declared character set encoding of ISO-8859-1 (Latin 1). That tells the web browser that this webpage should
only have the characters shown
on this chart. As a parallel example, if I encoded a web page with a Japanese encoding, that would tell the web browser that there should only be Japanese characters on the web page.
So, these web pages are encoded for Latin only, yet they contain non-Latin characters. For instance, notice in my quote above that Ducks name shows up as Du?k thats because the c-with-a-dot-above-it is not a Latin character (notice how its not on that chart I linked to). On the other hand, your cedilla shows up fine in my quote, because it
is a Latin character.
One might observe that I obviously have no problem viewing somebodys non-Latin username, and that the problem only surfaces when typing it in a text box. In a nutshell, this is whats happening: I can see this
properly, because a browser is allowed to try to
render a character outside of the pages encoding. You could have a Japanese character on this page, and Id be able to see it.
However, the issue is when the text gets
processed. As Im typing up this response, I can insert whatever character I want for example, right now Im going to type the ohm symbol > ?. Heres a screenshot of what it looked like when I typed it:
But what do you see now in this post? A question mark. Thats because Safari encoded it as Unicode (as it should) when it reached the GTP server, the server went Oops, doesnt match the rest of the page! (since it wasnt encoded as Latin 1), and thus replaced it with a question mark.
On the other hand, if you were to type that into IE, it would probably go through just fine. Thats because IE would pretend that the ohm is Latin 1encoded (which it clearly isnt), thus the server would receive it as Latin 1, and it would have no problem spitting it back out onto the web page.
Everything is way more complicated than that (enough that I dont understand most of it), but thats the gist of the problem.
).
