Bugs And Limitations

I, Pete Spicer, author of this (wonderful) website knows that I'm not perfect. I know I make mistakes. Maybe they're typos. Maybe they make the site fail to function. Maybe something doesn't work as expected. These things happen.

If you should find something unexpected, please let me know via my contact page.

So, here is the list of known bugs. (Last updated 22 August 2007)

• FIXED: Page anchors (#anchor bits of links) not processed correctly.
• FIXED: Infinite looping/server timeout, masses of strpos errors in url.php
• FIXED: Some - very badly messed up - URLS are not being parsed correctly.
• FIXED: Some same-level links were not being turned into correct absolute links.
• FIXED: Any generated errors (404 Page Not Found, 403 Forbidden) not handled by site, but by server itself.
• FIXED: Instances of unquoted parameters in tags caused the engine to malfunction.
• FIXED: Some URLS with : characters were incorrectly used, expecting the : to define the transport protocol (e.g. HTTP, FTP etc)
• FIXED: Some non-standard tags were not parsed correctly (EMBED specifically)
• FIXED: Some external links weren't handled by the tags not being recognised (due to the type of whitespace)
• FIXED: Some forms of Javascript were screwing around with the paths.
• FIXED: @import now handled correctly.
• FIXED: Massive issues with non-entity non-ASCII symbols, now handled intelligently.

There are some limitations that we are aware of in the current implementation of The Voices of Many... here are the ones we are aware of. (Last updated 22 August 2007)

• Forms aren't handled by The Voices of Many at all. (If you have to click on a button to go to another page, not a regular link, and/or select tickboxes or drop-down menues, or anything like that, that's what is called a form.) What happens is that the webpage system translates the reference in such a way that if you fill in a form on a translated page, it will end up on the original website. For example, if I translated google.com through The Voices of Many, when I hit the Search button, I would find myself back at google.com rather than a translated search results page.
• If the originating page is not in English, results are guaranteed to be unpredictable. There is no language/character set checking involved, non-English pages will be subjected to the same translations as English pages would, so anything could happen!
• Multibyte strings will not be processed correctly, so consequently some accented characters and other non-ASCII characters may not be parsed correctly, if at all - especially messing up Chinese! (This is because the system is not set up for multibyte translation, and even if it could handle multibyte character sets, there would be no point in doing it at all as the translations are geared to English constructions!)
• The server does check to see if the requested webpage is actually a webpage. However, the server only checks the HTTP header sent with the file before downloading - if the server reports it as "text/plain" or "text/html", it will download and process it.
• The server cannot evaluate JavaScript on any pages which use JavaScript. Consequently, if any links are added by way of JavaScript (writing HTML as an output of a script), the link may be broken, unless the browser correctly interprets the BASE tag implementation, to avoid this. It's not foolproof, but it's better than nothing.
• The site is particularly slow on handling pages of which HTML makes up over 50% of its content (this site, for instance, doesn't have that much HTML, it's more text)