Re: [3] [kDev] All your Wiki, Blog, Forum, Poll and Email List are belong to us...

["Arup Bhanja" <webmaster@xxxxxxxxxxxxxx>]

Hi Daniel,

Well for semanticising you are already using TBL's cwm API [as evident in
the wiki]. You might also want to take a look at :
1. Dr. Martin Porter's algorithm
2. Marciej Cieglowski's CNG [Contextual Network Graph] a paper he read in
July 2003 regarding the foundations of a contextual web + rules for making
it work.
3. Also Phil Wendley's website + http://www.lockergnome.com

[search 1 and 2  in google]

As for the technobabble [RSS/RDF et al] we have to deal with it a little as
we have to delve into quite technical info. The Contextual web you are
talking about is about to happen and I think Kendra is well positioned to
use it as it comes.

But we have to first set out the rules that we will use and in what context
[or maybe TBL and others will??] I remember the old ad copy in UK "Nothing
sucks like Exlectrolux" that no adman would ever use in US due the different
context of the word 'sucks' in US.

Both Dr. Porter as well Mr. Cieglowski have pointed out that the contextual
web is about to happen but they are yet to define in what context will it
run.

As for 'giving you content without any jot of XML' I am sure folks besides
kendra are working OVERTIME so I ain;t got any homework [he he  :))

Best
Arup
btw TBL is Tim Berners Lee. ;-)

----- Original Message -----
From: "Daniel Harris" <daniel@xxxxxxxxxxxxx>
To: <k-developers@xxxxxxxxxxxxxxxxxxx>
Sent: Thursday, November 06, 2003 4:12 PM
Subject: Re: [2] [kDev] All your Wiki, Blog, Forum, Poll and Email List are
belong to us...


> Hi Arup,
>
> In short, I want detailed answers... and please avoid technical terms
> (like RDF or hash) that cover a wide set of ideas when you're trying to
> explain a specific context.
>
> > This is very much similar to PageLet engine I developed...
>
> What it seems you've done is copy a whole site and then change all the
> urls to go through a php script that then gives you back the webpages.
> Sorry, but what's the big deal? How does it relate to what I'm talking
> about?
>
> > What we can now do is as you say 'Avoid the dross' is to search...
>
> Well, OK, but that's not really what I'm talking about. You are still
> talking about pages; and doing textual (not semantic) searches. All
> your text is fixed in the page format. You can't reorder the text based
> on a query. You're just talking about highlighting it - that's not
> enough for what I want to do.
>
> For example, on a simple level, I may want page type BenefitList to
> list all the benefits of Kendra to various stereotypes: content owners,
> content consumers, ISPs, etc. Then I'll want a page for each stereotype
> (StereotypeISP, etc) that lists the benefits and then other specific
> stereotypical information. So, 2 types of pages but the benefit text in
> each type of page is from the same source text. Edit the source text
> and you change the text on both BenefitList and StereotypeISP. Yeah?
>
> > The current offering will be to search a given RSS/RDF url...
>
> What does that mean?
>
> > ... we will deliver a hash of search results...
>
> What does that mean?
>
> > You can already 'read' or 'feed' any RSS/RDF file to your website
> > using...
>
> OK. Here it is. Don't talk to me about RDF or RSS or hashes. Give me
> tools/demonstrations where I don't see one jot of XML, where all I see
> is simple tools to enable searches textual/contextual/semantic and give
> me the results in the format that I want. Hey, now you've got some
> homework (ain't we all)! ;-)
>
> Cheers Daniel
>