The brute force method (still used) is to save the whole Arabic text as .gif file (or .jpg) – that is in graphics format, so any graphics enabled browser (most today are, apart from Lynx) will display the text either as scanned or made transparent (our original contribution to this backward technology was to capture the screens and apply giftrans.exe to them). This method is still used by a number of Arabic newspapers online either as the only option or an alternative. Needles to say, you can’t edit the text (without using OCR – optical character recognition) and it takes too long to load, plus if you turn images off, you never know what is going on. You can’t also search that graphical text.
You can also output files in .pdf format and require viewers to use Adobe Acrobat Reader.
Here are the details on How To Read Arabic Text on the W3
Writing Web Pages in Arabic:
The choices are also limited here and the issue of coding sets arises. You can, in principle, use any Arabic text editor to do that (by inserting the html tags) but those can easily get mixed up with the text and result in funny behavior. Some Arabic wordprocessors on different platforms started adding HTML filters. Examples include MS Office 97 Arabic edition for Arabic Windows 95 (and the infamous MS Internet Assistant for MS Word 95 Arabic), Nisus Writer for Mac and HTML editors bundled with Arabic enabled browsers. MSIE 4.01 MiddleEast has a customized version of FrontPage Express which has limited Arabic support.
Arabic pages on this site are done in MS Word 97 Arabic then saved as HTML followed by a lot of source doctoring using WordPad.