What characters are allowed in filenames for HTML files on ALL servers (*nix, Windows, etc.) ? I'm looking for the "lowest common denominator" that will work on all servers. USE: I'm naming a file to be served up publicly (Mysite.com/My-Page.htm)
E.g., space? _ - , etc.
E.g., can I have File-Name.htm, File_Name.htm File Name.htm?
Obviously, this needs to work with all servers and browsers. (IIRC, the name is limited by the server not the browser, but I could be wrong).
There isn't such a thing as an html filename.
Certain characters have to be encoded in html (eg if used in links) but the allowed characters in the document names will depend on the web server (and possibly the file system on the server).
Any file name will be URL-encoded so you should be fine. And for the record all three of your file names would work just fine.
If you don't want your filenames to be encoded by the server, you should avoid reserved characters: $&+,/:;=?@ and unsafe characters: space, quotation marks, <>#%{}|\^~[]`
But as the previous answers stated, the web servers should cope with whatever you want to use by encoding the chars.
Be sure to eliminate
* . " / \ [ ] : ; | = ,
which are never allowed, due to inconsistencies in file naming conventions standard practice is to use a-z and 0-9 and the underscore character. Space is needful for most users but if you can get away from using it there are parsing issues that improve reliability, you can read rfc's on mime ( multi-part internet mail extensions ) to get a taste of what is involved.
No matter what you do, something somewhere is likely to make life difficult - so much so that I now use cryptographic methods to generate random a-z lowercase strings and use those as filenames, embedding the useful info in the file source code.
Avoid the ampersand at any cost, ...
What characters are allowed in filenames for HTML files on servers?
That totally depends on the server. HTTP itself allows any character at all, including control characters and non-ASCII characters, as long as they are suitably %-encoded when requested in a URL.
On a Unix server you cannot use ‘/’ or the zero byte. (If you could use them, they'd appear in the URL as ‘%2F’ and ‘%00’ respectively.) You also can't have the specific filenames ‘.’ or ‘..’, or the empty string.
On a Windows server you have all the limitations of a Unix server, plus you also can't use any of \/:*?"<>| or control characters 1-31 and you can't have leading or trailing dot or spaces, and you'll have difficulty using any of the legacy device filenames (CON, PRN, COM1 and many more).
This is nothing to do with HTTP; just how filenames work on Windows, which is complicated.
can I have File-Name.htm, File_Name.htm File Name.htm?
Certainly. But in the last case you should link to it by URL-encoding the space:
<a href="File%20Name.htm">thingy</a>
Browsers will usually let you get away with leaving the space in, but it's not really valid. If you want to avoid having to think about URL-escaping, HTML-escaping and case-sensitive issues, stick to a–z, 0–9 and underscore.