Re: Efficient linking between sections of different documents

Ankur Bhatnagar wrote:
> 
> I think fetching _entire_ document is undesirale for two
> reasons:
> 1. It wastes bandwidth.  Sometimes the relevent reference is
>    only of one or two lines while the file consists of 500 or
>    more lines.  It wastes time also esp. if the reference is
>    somewhere towards the end of the documents.

Depends. Suppose we have a page with 10 footnotes and we need 5 of them.
Fetching entire page once will save bandwidth in this case.

> 2. It distracts the attention of the reader towards other
>    unrelated topics.

Hmmm... Having a page with several unrelated topics is author's decision. 

> This mechanism will require server to do some processing but
> I think ultimately it will result in net savings to the Net 
> community.
 
I am very reluctant to support anything that requires HTML parsing on
server side. People tend to put all kind of garbage on their pages.
Last brilliant and inventive thing I saw was using --!> as a comment
terminator. Lusers are too inventive and servers can't keep up with
them.

> There is no need to add any additional tag.  Only a new
> convention needs to be defined.  I suggest that the referred
> content be enclosed as follows:
> 
> [Start of the document]
> ...
> <a name="xyz">Heading</a>...Content...<a name="/xyz"></a>
> ...
> [End of the document]
> 
> The server will search for "xyz" and then "/xyz". It will
> return the content starting from "xyz" till "/xyz".  If there
> is no "/xyz" (current practice), it will return the whole
> document starting from "xyz".

But how is server supposed to know that you want him to search from "xyz"
to "/xyz"? You'll need new HTTP header for this. HTTP 1.1 draft is currently
in last call. You have 2 or 3 weeks to propose your method on www-http list.

-- 
Life is a sexually transmitted disease.

dave@fly.cc.fer.hr
dave@zemris.fer.hr

Received on Wednesday, 26 June 1996 17:02:20 UTC