CSS Styling of Namespaces in HTML | Hacker News Reader

CSS Styling of Namespaces in HTML

aaronsw.com

31 pointschenster11 comments

There are a control and 4 tests, but the results for the listed browsers there have only 3 letters, like GGR. Shouldn't there be five for each browser? Otherwise, which three of the five are you testing?

I don't mean for this to be a snide remark, but I'd like to let you know that, unfortunately, Aaron won't be able to respond to you.

Context: https://en.wikipedia.org/wiki/Aaron_Swartz#Death

it's sad and beautiful that we can have conversations with ghosts.

Don't they have to reply to make it a conversation?

They will eventually.

OP will deliver. We just have to wait...

I assume that for all browsers listed, the control passes.

It is my understanding that the correct sequence is GGBG. I believe that if < 4 are given for a browser, then it is the first few that are tested; assume black if it isn't listed. This abbreviation is because later tests cannot be expected to work if earlier ones don't, so it's not worth listing them (though it's still a little inconsistent which are listed).

Because the page is so old (even before Aaron died), the only one of the browsers mentioned that I can verify is

IE6/Win: GBBB

I believe that 3 is guaranteed to be black if either 1 or 2 are black (making the IE5/Mac and Safari results slightly redundant). I believe that 4 is guaranteed to be black if 2 is black or 3 is red, meaning that it's not worth showing for any of them (... except for Opera 7b2 after refresh. Dang, I thought I had it figured out).

Not wanting to disrupt piety, and I'm not sure how situation was back in 2002, but I'm slightly confused and I think there are several mistakes in that test:

1) source pretends to be XHTML, but since document is served as text/html, it is treated as HTML. And there are no namespaces in HTML. I assume there never been any in 2002, but cannot swear on that.

2) CSS selector "ns\:elem" has nothing to do with namespaces, it is just selector looking for <ns:elem>s 'tagname'. It works in HTML, and that's what page actually shows. Try this:

    data:text/html,<style>a\:\:\:{color:red}</style><a:::>foo</a:::>

3) Even if it was served as application/xhtml+xml, it would be: 3a) invalid: ex3 namespace is not declared; 3b) if fixed, then without applied style: SGML comment around CSS content would hide it from parser (or at least it does something like this in my test).

4) @namespace declarations in CSS must be "on the top" (after @charset and @imports, IIRC), not after rules, so that one declaration in the test is ignored.

5) Test4 has a typo: element name in html is `text`, but CSS selector looks for `test`.

I'm not sure all those remarks would make sense back in '02 context, but if it was as misleading as it seems to me, it is quite sad how many people has been possibly confused by that, considering it is fourth verbatim Google result for "CSS namespaces": https://www.google.com/search?q=css+namespaces&tbs=li:1

Looks like Aaron will always continue to live in the Internet. RIP Aaron.

Reply on news.ycombinator.com