- From: Carl Morris <msftrncs@htcnet.com>
- Date: Tue, 17 Sep 1996 16:40:29 -0500
- To: "David Perrell" <davidp@earthlink.net>
- Cc: "WWW Style List" <www-style@w3.org>, "WWW HTML List" <www-html@w3.org>
| > are meaning less, should not require ending #, only a ##( should be | > allowed to open this new specification, it is a new specification you | > know... | | I think you misunderstand. Making this syntax legal doesn't invalidate | any existing URL. The only thing I'm saying here is that '#' explicitly | terminates a fragment id. The end of the string implicitly closes a | fragment id. There is nothing here to invalidate existing URL syntax. A | UA enhanced to recognize the extended syntax will recognize the old. Exactly, and I don't see a need for the "trailing" #. The fragment only ends at the end of the string, your fragment extension ends at a matching ) or what ever... I don't see why one would require a #)... A normal URL requires escaping the parenthesis, your extension SHOULD NOT! All additional pounds should be escaped, again, yours should not be. This is exactly what reserved characters are reserved for, new extensions.... also, remember, your fragments here are exactly that, yours, you'll probably never see them seent back to the server... | Older browsers do a sloppy parse and only parse off the text from the | rightmost '#' - the URL still works in my testing but the valid | fragment id is not recognized and is sent on to the server, which could | conceivably cause problems on old server software that does not | prudently parse off possible fragments. Also, Netscape, being an | exceptionally sloppy piece of software, IMHO, searches through a The problem is non frame browsers have no use for any of this information... the closest this extension gets them is the initial frameset page, which if it contains NOFRAME's tags will be okay, but, does not give this extension full backwards compatibility.... they will have to manually get the the page that was of interrests... I sent an example to the www-style list... but again: http://hello.com/index.html##(main=products.html##(product=nwsamd.html)) might be a url I might give out to potential customers... but this gets a unframed browser no closer than the main index file... maybe: http://hello.com/nwsamd.html##(index.html##(main=products.html##(product =nwsamd.html)) ) MSIE currently handles that just fine... actually, using the perenthesis is giving me troubles... "can't find site http" ... but a bracket [] doesn't give me anytrouble, and still, even with an html file that contains <A NAME="#[index.html"> (whacking the rest of the string of the url too) will not make MSIE go to that fragment... meaning MSIE knows its incorrect... Now if MSIE supported such a fragment, it would easily know that the double # marks frame based state data, and should load the first string as the base frameset... now I tested this: http://hello.com/index.html##index.html and placed a named anchor in the file with "#index.html" and again MSIE would not look for the anchor... so this could be used: http://hello.com/nwsamd.html##index.html[main=products.html[product=nwsa md.html]] ... however I am finding that MSIE does not escape any of the perenthesis/bracket/brace characters... its quite simple though, this extension requires they be escaped... just an idea to maintain complete backwards compatibility though... I am sure many or more understanding with some of the does and don'ts around those characters, but I must stress that, some characters were marked reserved solely for this type of stuff... its time they be used... -|- Carl Morris (N0YUV) -|- 1:285/302 -|- msftrncs@htcnet.com -|- -|- Hooper Connections BBS -|- 1-(402)-654-2102 -|- 28.8kbps V.34 -|- -|- http://199.120.83.179/~moreese/ -|-
Received on Tuesday, 17 September 1996 17:45:54 UTC