Re: source for hz algorithm

[Keiron Liddle]

On Wednesday, January 29, 2003, at 03:20  PM, Rhett Aultman wrote:

This might be semantic nitpicking more than anything, but how can finding a worse break prove you have the best break? Wouldn't you have to find all possible breaks and verify that they're "worse"? Also, just for personal enlightenment, what principles govern "betterness" or "worseness"?

I didn't say it properly.
Okay, if we are going down a page the first break found after the first line will be the first best break. The next line will then become the best break as it is closer to the optimum distance and has no keeps. As we go down the page it will keep find better breaks for the current page. Going in only forwards a better break is one that has a lower keep value or an equal keep but is closer to the optimum. Once it goes past the optimum then if we find a break with an equal keep to the current best and it is further away then the current best is the best.

Then if for example we want to find the optimum break. There is also the possiblity
to get the next break within a context (which invalidates all further breaks) or
previous break.
<<

Could you please expound on this idea a little further? I don't think I'm quite following.

Sorry don't have time right now. There is a bit of info in the other message.

The only drawback is that it constantly needs to find the child layout manager that
applies to a given break and that finding the BPD distance could be time consuming
in some circumstances. Optimisations should help a bit.
<<

Offhand, I would think that this won't represent a reall performance bottleneck, and it would seem quite necessary given my somewhat green understanding of what you're proposing.

I am hoping that making the breaks simple and easy to find certain properties from
any position will help us to explore what to do next.
<<

I'd really like to see this feature in motion. Being able to find this seems imperative for handing conflicting constraints and other anomalous situations.

Take a look at the code in attached to the other message.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]