Sunday, May 20, 2012

Firefox urlencode lapse, or ?

I was validating an HTML page at today when I hit a few URL errors in anchor links.

The URL's were added ad hoc for kanji from pages found in Firefox and the URL copied out of the address field.

Oops!  These URL's are for kanji stroke order diagrams and the URL includes a Japanese character.  The URL comes back from Firefox address bar "COPY" with the kanji character urlencoded. But: the URL includes a param following an ampersand and THAT was not encoded.

According to w3c the URL is not valid unless the ampersand is escaped (but the links were tested and worked fine ...)

Once the ampersand was escaped,  the HTML validator stopped objecting - but the "?" for the start  of parameters was still not escaped!

Here is what Firefox returns:

Here is what the HTML validator insists upon:

Notice that %26 is not what is required for this URL to work correctly.

Firefox reports the URL on my web page as

The HTML anchor attribute on the page is now set to


The kanji is

Friday, May 11, 2012

Kanji Recall

The "Henshall" kanji Curl applet at my kanji study pages now will remember where you were last in the set  -  at least as long as the Curl Surge RTE plugin is running ( a session cookie is being used.)  You can even change browsers from, say, Firefox, to Chrome, and Curl remembers which Kanji you were on.

If you have a request for different behavior or a different kanji set, then forward your request to where the "AT" is our @ mail symbol.

Note: the applet requires the MIT Curl browser plugin from or

Curl has a forum at