[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: sgml tools help



On Wed, Apr 05, 2000 at 09:38:38AM -0400, Greg Ferguson wrote:

[Regarding generating html files from sgml files]
> In the wrappers,
> we create (using jade) a singular HTML file and individual/chunked
> HTML files 
[snip]
> The individual files ("chapter" or "sections") are what are used
> to view/browse in the LDP. The singular/large HTML file is used
> solely in the production of the plain text, PDF and PostScript
> variants. 

The singular files should also be used for the HTML files available
for download.  Hopefully these will also be supplied on CDs.  The
advantage for the chunked ones (non-singular) used for on-line viewing
is that you don't have to fetch the entire document (and transmit it
over the Internet) just to read a few pages of it.  But if you have it
on your PC's disk, then the transmission costs (and delays) from the
hard disk to memory are very small.  If it's singular then one can
search it while using a browser such as Lynx.  If it's in chunks
searching is not as easy.  One can search with chunks with grep but
it's not convenient since one has to then use a different program to
view the document.

But for a book-size document, one may still want it in chunks on an
old PC.  Thus I think that the only major uses for chunked documents
are for online viewing and possibly book-sized documents.  The ones
people keep on their disks should usually be singular (except for
images).

A problem with singular files generated from the linuxdoc type of sgml
is that they don't have tables of contents.  But there's a fix for
this.  Catenate all the chunks together with the "cat" command.  Then
fix the links to make them all relative to the resulting singular
document.  Remove the header section, etc.  A shell script could be
written to do all this using the sed editor to fix the links.  I
started writing this but haven't finished it.

-- 			David Lawyer


--  
To UNSUBSCRIBE, email to ldp-discuss-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


mirror server hosted at Truenetwork, Russian Federation.