From koffice  Fri Dec 13 14:18:37 2002
From: Nicolas Goutte <nicog () snafu ! de>
Date: Fri, 13 Dec 2002 14:18:37 +0000
To: koffice
Subject: Re: Formatting large docs in Kword
X-MARC-Message: https://marc.info/?l=koffice&m=103978920608461

On Friday 13 December 2002 06:54, Clarence Dang wrote:
> On Tue, 10 Dec 2002 07:04 pm, Jonathan Drews wrote:
> > On Tuesday 10 December 2002 12:25 am, Clarence Dang wrote:
> > > Hi Jonathan,
> > >
> > > Try loading your text file with the other two options in the "End o=
f
> > > Paragraph" group i.e. "Sentence" and "Old method" (not "As is") in
> > > the "KWord's Plain Text Import Filter" dialog.  They are not perfec=
t
> > > but I think they will do what you want.
> >
> >  Splendid!  Worked like a charm Clarence.
> >
> :)
>
> I would be careful about those modes though (esp. "Old method") because
> they heuristically determine when it's time to end the paragraph (so
> sentences could actually be appended to headings).

Yes, sometimes "Old Method" is too smart. That is also why there is no re=
al=20
name for this method. However it works well with paragraphs separated by=20
empty lines.

>
> > > > BTW Kword now loads and edits these big docs pretty quickly.
> > >
> > > Really?  The filter reads one character at a time, last time I
> > > checked... You should try loading the 20MB DNA sequence associated
> > > with bug #45973 :-) (actually the real problem with that bug is the
> > > amount of memory that KWord takes up but anyway I'm getting OT agai=
n
> >
> >  Yes, but by using your advice it loaded Volume 1 of Decline and Fall=
 of
> > The Roman Empire pretty quickly.
>
> How many gigahertz is your machine? :)
> On my Pentium II, it took 8 minutes CPU time before I stopped loading
> it.... I'm actually going to be working on this filter in a few weeks
> (colour text via ANSI escape codes, table support...).

Do you plan to do it for the import only or for the export too?

>
> > I then saved that as a Kword doc (with
> > the new full page format) and reopened it.  Memory consumption was
> > negligible!
>
> I don't think KWord internally uses XML (haven't checked kwdoc.cc, but)=
=2E
>
> > When I loaded the original text file using "As is: At the
> > end of the line", the memory consumption was enormous.
>
> Surely, "Sentence" mode would have have taken a similar amount of memor=
y
> (it would be slightly less but not significantly less)?

Not necessary, end of lines are not commonly end of sentences, so you wil=
l=20
commonly get much less paragraphs than with the "as is" mode.

That is why I had made this mode, as it can make quite good paragraphs ou=
t of=20
monolithic texts (for example those converted from HTML with clueless=20
scripts.)

>
> > It used  ~500 Mb.
> > So the page formating makes a big difference in the memory
> > consumed?
>
> Well there isn't really a difference in Page Formatting at all with tho=
se
> modes, just how often "paragraphs" get written.  Dunno about memory tho=
ugh
> but I would imagine fewer paragraphs (not "As is" mode) would be better=
=2E
>
> > It took about half an hour to paste the 722 additional pages, to make
> > the 1444 page *kwd.

>
> IMHO, that's much too long!  But, I'm too busy to fix that at the momen=
t.

It is perhaps too long but it is probably the time KWord currently needs =
to=20
paginate. (Do not forget: you are making tons of double precision floatin=
g=20
point operations, you cannot have them in no time!)

>
> Clarence

Have a nice day/evening/night!

>
>
> ____________________________________
> koffice mailing list
> koffice@mail.kde.org
> To unsubscribe please visit:
> http://mail.kde.org/mailman/listinfo/koffice

____________________________________
koffice mailing list
koffice@mail.kde.org
To unsubscribe please visit:
http://mail.kde.org/mailman/listinfo/koffice