Slow Javascript Navigation Links

[TheLittleGuy]

Hi there. I'm going to ask my first question, but I think I already know the answer!

I'm in the process of putting together my Actinic Business v7 website, and I've implemented a DHTML Javascript menu system on it. The catalog links are taken from "Act_section_tree_URLs.js". They're pretty much just the filename of the target page, and do not include any CGI calls. However, when I try to navigate to any page using the menus the URL is intercepted by the PERL scripts and changed. It applies the "NOLOGIN" wrapper to the scripted URL to detect whether the user is logged, and then generates an appropriate link accordingly. Unfortunately this is time consuming, adding a noticiable burden to navigation speed.

I'd really like to bypass the application of the NOLOGIN wrapper as I don't need it. I've read the forums here, and I've seen people with similar problems but no solution. There's nothing to help me in the Advanced User Guide either. I've checked my host, and it's performing well. I've tried turning off customer accounts (which I don't really need) and applied various filters to the script URLs, but the wrapper is always applied regardless.

Any advice?

[I-CC]

Any links to your sections and sub-sections that are produced using the menu systems pulled from the javascripts will all go via the cgi scripts, regardless of the nologin situation.

If this is really slow for you, then it might be the strain being placed on the server you are on as opposed to the links themselves.

The main issue with the menu created from the links is the search engines (in the past most noticable being Google) as this has caused some issues with the results pages.

[NormanRouxel]

Are you using one of my patches from www.drillpine.biz/actinicstuff ?

[TheLittleGuy]

No, I'm using my own script to tie into a different DHTML menu solution. Have you found a way around the problem?

As for the google issue, surely I just need to have a standard link to an HTML site map?

[Mike Hughes]

A site map will help Google find your pages.

But..

all the pages will be two steps away from your home page and will only have a single link to them. Compare this to a site that use HTML navigation where the pages are linked to from the home page and by all the other pages in the site/section.

Consider Google's ranking based on links into a page and the anchor text in those links and think about which one will do better.

Mike

[TheLittleGuy]

Thanks for the tip. I think I'll keep the HTML section links as well, but hide them in an invisible DIV. That way I can have my cake and eat it.

[Mike Hughes]

Just make sure you don't trip one of googles spam filters. Google doesn't like content that's put there for them but invisible to visitors.

Mike

[jont]

Just make sure you don't trip one of googles spam filters. Google doesn't like content that's put there for them but invisible to visitors.

Mike

Mike - how does Google know if you have content on the page but hidden with a CSS statement in another file? Have you tried this out? I have a client who wants me to do this but I have told them its a bad idea but have no proof as such.

[NormanRouxel]

No, I'm using my own script to tie into a different DHTML menu solution. Have you found a way around the problem?

Both my menu systems NorTree and NorCascade allow you to either go via the cgi-bin (for full Business compatibility) or via direct html page links if using Catalog.

Each also comes with a GUI admin program that allows you to alter the zillions of configuration parameters without editing any JavaScript.

[wjcampbe]

@jont #8
Run a page through a spider simulator such as http://tools.summitmedia.co.uk/spider/ to see what google would see.

[jont]

Cheers Bill - will have to create a dummy page with hidden <div> tags to see how it fairs....

[magicalwonders]

Thanks for sharing the link Bill.

Myles

[Mike Hughes]

Google doesn't say publically how they detect spamming pages so I'm not sure how that page will help. It looks like a fairly standard analysis page to me.

As Google don't say what the can / can't detect (hardy surprising really as if they did then everyone would know what they can get away with and what they can't) then the only real advice I can give is to avoid anything that breaks their guidelines or specific recommendations; in this case "Avoid hidden text or hidden links".

For more visit: http://www.google.com/intl/en/webmas...uidelines.html

Mike

[wjcampbe]

Mike, the page will simply demonstrate how much of what is 'hidden' is really still visible to any spider - as you say, Googlebot may have 'superpowers' that let it see even more.

I would repeat and endorse your advice - avoid hidden text or hidden links. Everything on the page should help the human user.

[TheLittleGuy]

Dear god, I've been so STUPID! I was working with local versions of the URL script that I'd "localised" myself!

My filter now works, and I've erradicated the problem! Hurrah!