Hi Arved, I actually have bits of code using BreakPoss lying around that I am going to try to commit rather quickly. It doesn't work yet, but it will give people some ideas. Not sure it will be this week though, since I'm away on business the whole week. Depends on if the hotel room has a modem port :-)
I've seen that you are making strides on xslfo-proc, but haven't taken the time to look at it at all. Not that I wouldn't like to, but my resources are limited right now. I'm not convinced that the tree building approach is that much different than the SAX one either. In fact, the layout managers are really only looking at one child at a time, so that they could be running in parallel with the FO tree building. Since tree building is done by SAX, they could also be pipelined with that. The FO reader/builder is pushing and the LM are pulling. Does that make sense, or am I just blathering on? Regards, Karen Arved Sandstrom wrote: > > Hi, Karen > > Well, _I_ am interested. :-) Already scanned it once, but I'll definitely > give it a thorough read quite soon. > > I think I might be able to resume contributing to FOP layout stuff sooner > than I thought. I don't mean the image stuff which is still on my plate but > more the guts that you and Keiron have already been tackling. The design > ethos that I see expressed in the new FOP so far conforms to how I would do > things anyway - we had lots of discussion about it last year and I don't > recall that anyone expressed dissatisfaction with the idea of managers. I > certainly didn't. :-) > > In fact I am using managers in my xslfo-proc prototype. They are really > rough around the edges at the moment but I expect to have fairly clean code > and good design docs for a first prototype release in about 2 weeks. I > already had "First PDF" yesterday, albeit on a very simple example, and not > without errors. I am hoping that much of the layout design in xslfo-proc can > conform rather closely to New FOP, so that if in future the production C/C++ > codebase gets folded back into Apache XML FOP, that there will _not_ be > distinctly different designs. > > The main difference I believe is that the tree-based FOP approach lends > itself to more layout sophistication at less cost (in understanding and > complexity) than does a SAX-based approach like xslfo-proc. But because a > SAX-based XSL formatter must also maintain a working subset of FOs, raw > layout, managers, and areas, at some point the approaches blur into each > other. The main difference is other requirements - in the case of xslfo-proc > one of those is memory, so some sophistication will be sacrificed in order > to minimize storage. Ironically, the existing FOP, for all the advantage > that it takes of having the entire tree available to it, may as well have > been written using SAX directly. But I think another advantage of having > xslfo-proc in the wings is that it may provide pointers as to how the new > FOP can be moved to SAX, with little actual impact, hence addressing memory > concerns. We'll see. > > In fact I am switching to FOP for the rest of the day, to start looking at > the image layout, so I'll be mixing stuff up a fair bit from now on. > > Regards, > Arved > > -----Original Message----- > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of > Karen Lease > Sent: March 17, 2002 12:19 PM > To: [EMAIL PROTECTED] > Subject: Break Possibility scheme in new layout > > Hi, > I wrote a fairly detailed description of the Break Possibility idea and > put it in the design docs section, for anyone who's interested. > > Could someone please tell me the procedure for getting the html version > of the design docs regenerated on the web site? Or does this just > magically happen? > > Thanks, > Karen > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, email: [EMAIL PROTECTED] > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, email: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]
