h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

First, a general HTML query.

If, on the page http://www.example.com/something/else you had the code Look at BBC 7!, it would create a link to http://www.example.com/bbc7. In other words, the initial "/" in the link causes it to return to the root directory and look from there.

Conversely had the link been to "bbc7", not "/bbc7", you would be brought to http://www.example.com/something/bbc7. Am I right?



Now, I thought that this worked on h2g2. I'm certain I could create links to other sections of the BBC site by enclosing the path, starting with a soldus, in the <<./>./</.>> and <<./>/.</.>> tags.

Thus, <<./>./</.>><./>/bbc7</.><<./>/.</.>> would produce a link to http://www.bbc.co.uk/bbc7. But that doesn't seem to work any more.

The other method is to commence the link with ../../../, going up a directory each time. But I'm not sure how many directories to go up. I can experiment till it works, but if you view the page without the skin reference in the URL (i.e., with a lesser number of directories), will the link still work? I know they're pseudo-directories, but the browser will see them as directories. I should check, I suppose, how the parser interprets it by viewing the HTML code. I'll try that next.

The thing is, I'm sure that the <./>/bbc</.> method used to work. Now you seem to ignore that initial soldus, and interpret it as an internal h2g2 link, like <./>RF1</.> or <./>NamedEntries</.>. Have you changed the coding, or is my memory at fault?

TRiG.

Creating a backspace link, so I can check the source and stuff.

Testing, testing: <./>../../../bbc7</.>.

That works in brunel preview. Let's see if it also works in the conversation, when I remove the skin from the URL.

TRiG.

It does. Looks messy, though, doesn't it?

The source code looks even messier:

../../../bbc7

So it does need all those ../ thingies, even with no skin in the URL.

Hey, but if you're signed out, the DNA server doesn't stick the skinname in the URL. I'll sign out and check again.

See you!

TRiG.

../../../bbc7

That's what you get when signed out and visiting a page with no skinname in the URL. That's an overkill, but it still works, because you can't go higher than the root directory (or should that be lower?).

<./>/dna/hub/communities</.> - Does this still work?

TRiG.

Yes it does. What's the difference?

Bye now!

TRiG.

This all got broken quite a few software updates back. There's even been at least one other thread reporting / complaining about the problem.

So it's not my imagination that it used to work? Good.

<./>/bbc7/</.>

How about that?

Well that form of the link looks to be working from here - assuming the destination page is meant to be that yukky!

Another memory has surfaced that perhaps the previous thread was recalling was mostly about including links in articles, ie GuideML/HTML, but I had also already noted the extra advertising/tracking stuff had been added for both situations. However, I had thought it was a little odd for Trig to be having trouble with an internal BBC link because generally those *don't* have all the extra junk added, but I couldn't easily find that previous thread to check what had been tested then.

I've looked at code from 2002 (just a random sample) and in that code, it wouldn't work either with or without the trailing slash.

Basically, we found that people were putting in links in GuideML with leading slashes. This was mostly in pages created before the BBC took over, when that link would be correct. In order to keep those links working there is explicit code which assumed a link with just a leading slash was still an h2g2 (or DNA) link. When it was noticed that it was now difficult to link to different parts of the BBC, we added an extra check where we didn't mess with links containing more than one slash, which is why /bbc7/ works and /bbc7 doesn't.

Thanks for the workaround, Jim. Is it only on the dna boards that a final slash gives you the frontpage instead of what you're looking for? So it's safe enough to add a slash to the end of everything else?

I suppose the simplest thing is to preview and check that it works.



Interestingly, you can bypass the BBC's go server thus:

<./>//www.example.com</.>

Nifty trick, if you want it. I'm not sure why you'd bother, though.

TRiG.

Ah. Scrap that. It just bypasses it in preview. And all links bypass it in preview. I hadn't noticed that before.

TRiG.

Hmm intresting,

I've just been on ../../../bbc7 in plain strange why does it light up in blue for you but not me http://www.bbc.co.uk/dna/bbc7
I'll keep this thread might come in handy lol.

The <./>NamedEntries</.> page will explain how to do links in plain text. Y'see, plain text isn't *entierly* plain.

TRiG.

http://example.com

Links no longer bypass the <./>/go/</.> server in Preview mode.

TRiG.