[prev in list] [next in list] [prev in thread] [next in thread] 

List:       koffice
Subject:    Re: koffice file formats
From:       weis () stud ! uni-frankfurt ! de
Date:       1999-07-01 18:03:02
[Download RAW message or body]

Hi,

On Thu, 1 Jul 1999, Leon Widdershoven wrote:

> John Zedlewski wrote:
> > 
> > >I think such "sections" belong in different files and the document should
> > >essentially be a directory. this can be put into a tar-file for transport.
> > 
> > I have to say that I really agree with this approach.  Then the documents
> > can still be valid XML, just containing an XML tag that points to the
> > content to be included.  This should confuse parsers a LOT less than a
> > series of XML sections separated by binary chunks.  It will also make it
> > much easier and more efficient if the user wants to, say, look at or tranfer
> > a preview file that doesn't include all of the final graphics.  The program
> > could just display an empty, correctly sized box where the binary section
> > would go.
> > --JZ
> 
> I do not agree. I think it is not the way to go, having multiple files
> for one saved file. Especially not when it is up to the user to tar
> or whatever the files together.

KOffice tars the stuff automatically during saving. The user is not bothered.

> A better idea would be to put the binary chunks at the end of the file.
> 
> Or alternatively to make a tarred file the basis of the save file. But
> this would mean the savefile is not XML, since it has to be preprocessed
> before use.
> 
> In the end, I think a good parser should just ignore binary stuff. This
> is not so difficult if on top of the binary chunk the size of the chunk
> (in bytes) is added. Something like (I don't speak XML):
> <BINARY 10000>
> hfsj ... hdasiu
> </BINARY> 
> for a 10000 byte binary.
> A simple fwrite( tmp_buf, nbytes, 1, inputfile ) would
> do the job. Or a fseek( inputfile, nbytes, SEEK_CUR ), if you don't want 
> the data. 
> 
> Regards,
> Leon
> 
> -- 
> In a world without walls and fences,  l.widdershoven@fz-juelich.de
> who needs Windows and Gates?          Institute of Plasma physics (IPP)
>                                       Forschungszentrum Juelich GmbH
> 
>
Bye
Torben
 

[prev in list] [next in list] [prev in thread] [next in thread]