[Date Prev][Date Next][Thread Prev][Thread Next][Author Index][Date Index][Thread Index]
Re: [zzdev] Vob linebreaking
- To: b.fallenstein@xxxxxx
- Subject: Re: [zzdev] Vob linebreaking
- From: Tuomas Lukka <lukka@xxxxxxxxxx>
- Date: Mon, 7 May 2001 21:50:15 +0300
- Cc: zzdev@xxxxxxxxxx
- In-reply-to: <5047.989245180@xxxxxxxxxxxxx>; from b.fallenstein@xxxxxx on Mon, May 07, 2001 at 04:19:40PM +0200
- Mail-followup-to: b.fallenstein@xxxxxx, zzdev@xxxxxxxxxx
- References: <20010507104032.X23764@xxxxxxxxxxxxxx> <5047.989245180@xxxxxxxxxxxxx>
On Mon, May 07, 2001 at 04:19:40PM +0200, b.fallenstein@xxxxxx wrote:
> > Finally, found a satisfactory solution: at first, simply insist that all
> > spaces are between vobs. That way, we can just use TeX's glue model
> > for the boxes. Hyphenation can be done later and as in TeX (IIRC), only
> > after trying to break with the spaces. That way, there's no need for the
> > endIsBreak etc. stuff.
> Huh?!? How can you call this satisfactory? It has virtually *all* horrible
> linebreaking problems I
> can think of:
> - It creates FAR too many vobs, because each word is a vob...
For edited text consisting of spans, this will be true anyway.
> - ...and each space is a vob, too, if you don't want to introduce extra
> complexity with these.
It's not too much complexity to name the extra width of the space as
the end of the vob preceding it. The spaces are a problem in any model.
> - There is no way to specify that a break should not be between two vobs,
> thus breaking at
> arbitrary positions inside vobs (a mistake we've made already!).
Actually, there is. The system is based on TeX glue and if there is no
glue between two boxes then no break. That's what I best like about
this: explicit control.