Using Google Imode search to investigate your page

(imode is software on many japanese phones)

[Brett_Tabke]

There has been alot of talk public and private around here about buying links based on PR. It is very tricky business, because even if you are the only link on the page, there is no garantee that any pr is going to be passed to an external site.

Ever look at your site from Googles Imode search?
[google.com...]

Notice how it id's your sites navigation links? That's done on the fly. Imagine what they can do with it in an algo.

[le_gber]

Not quite sure I understood everything ...

I went on my site but there wasn't all the links or all the content from my pages ... does anybody know why?

Leo

[TheDave]

I see what he means, if you click on the SERP link there *may* be a "NAV..Home..Index.." link which takes you to a simple row of text links. It's even grabbed the alt text from my image based nav bar. Also, it's got a page I updated only a couple of hours ago. I just about fell of my chair, I thought the freshbot had lost interest in my site. Fantastic :) [edit] just realised that its a proxy.. never mind :P [/edit]

[edited by: TheDave at 2:11 pm (utc) on April 3, 2003]

[skipfactor]

Powerful. Why the Japanese slant?

[danny]

You know, I reckon the web looks better that way than it does normally... but I've always thought it went downhill when they added support for images :-)

(My main site is still 100% image-free.)

[adamas]

Very impressive. It must be doing more than just looking for consecutive links.

I have a nav bar - which it picks up as navigation.
I have a list of recent widgets and a random selection of widgets which it picks up as navigation.

Then I have text that includes lists of different types of widget with each type a link to a page. Out of context the types themselves would just be comma separated links. In context they have text either side as part of a sentence. This is listed as is rather than declaring it navagation.

Then it picks up the bottom menu bar as nav. RHS links as nav. Footer links as nav.

I'd say it was spot on on every call.

[heini]

>I'd say it was spot on on every call
Yep, almost scary.
>japanese
Because Imode is originally japanese, slowly coming to Europe, don't know about the States.

[Brett_Tabke]

>be doing more than just looking for consecutive links.

Much more, and much easier.

Break out your secret decoder ring. As you probably know, in an ASCII character set [robelle.com], we assign a number for every letter of the alphabet. Here is the same but simplified idea:

abcdefghij
0123456789

Now imagine how a scrap code looks:

<a href="abc">abc.com</a>

And if we converted it to numbers, it might look (example string)

13112154546879301274

Now think of that template on every page. To id the template, we just have to do a sliding match looking where character strings match.

13112154546879301274
13112154546879301274

Do that sliding match onward down through the page looking for identicle fragments of code and we have identified part or all of the site template. We can then use the template like a cutout pattern to "sift" the rest of the pages and any new pages that come up.

Template scraps from the home page that have in site links to other valid pages that in turn have identifiable templates, have the highest likelihood of being a nav link.