Hebrew on the Web

The main text flows correctly - if you try to read this using Safari under OS X, you won't at first notice any problems. Then, you'll note that the ambiguous characters (glyphs such as punctuation that could be placed differently depending on the language the browser things is the base for the current paragraph). Bidirectionality is a messy subject. The problem is that when you tell a browser that you are using UTF-8, for instance, it is easy for the browser not to be sure where to put punctuation marks: if the base language were English, say, then periods go on one side of the sentence display. If the base language is Hebrew, the opposite is true. On the pages I did for "We are the future" (see, for instance, www.wearethefuture.com/he/concert_about.html, everything looks fine on a PC using IE or Mozilla. But, fire up OS X and take a look with Safari and you see hyphens and periods that are placed wrong. Not fatal - at least the text flows nicely from right to left as it should - but not what I want. Here is what I used:

• In the head of each document, I noted utf-8: <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
• Paragraphs got styled thus: <p align="right" lang="HE" dir="RTL">. (Normally I'd put this in a style sheet, but this was a drop-in and I didn't want to isolate the pages I worked on from the general style sheet and any global changes. And the job was too fast to comfortably isolate the items in a local style sheet.)
• The Hebrew characters were encoded using some Microsoft? characterset, using the range (aleph to taf) &1488; through &1514;. I got this by saving a Word document containing Hebrew to HTML. I used Word 2000, under Windows NT, with Hebrew resources installed. (I don't think that what was saved is Unicode, which I recall having a different hex offset that is much larger - but I'm new to this, I could be entirely wrong.)

Microsoft puts charset=Windows-1255 when saving the word doc as HTML. This is the near-universal set used in Israel. The thing is, I do feel that I don't want to be relying on browsers on non-Windows platforms simply figuring out how to accomodate Windows. I'd like something universal. The goal is to stop having to worry/think about platform. It doesn't matter in this case. Putting "charset=windows-1255" at the top of the file changes nothing compared to utf-8. Additional notes: I got very curious and tried to view this using IE 5 and Mozilla 1.2 under OS 9.x on the Mac. IE displayed nothing but question marks, regardless of whether the meta tag indicated utf-8 or windows-1255. Mozilla looked perfect, just as on the PC. So, I did the obvious and downloaded the OS X version of Mozilla. Still perfection. The problem is Safari, not my encoding. I'm willing to bet to bet that what I did is still not the optimal way to do this. It worked well enough for this project (pending some final fussing to see if I can figure out how to encode things so that Safari groks the language and base direction correctly - of course, but that depends on finding a workaround Safari's bugginess), and it feels great to have some Hebrew web pages up. Next try: do it better. First path to explore? Convert these character encoding to actual Unicode and see if that works better.