Douglas W Philips wrote: > On Fri, Mar 26, 2010 at 6:49 AM, Hussein Shafie <hussein at xmlmind.com > <mailto:hussein at xmlmind.com>> wrote: > > --> I now understand what you want to achieve and why you prefer to do > it this way. > > There are probably more issues with ditac than you imagine. In order to > fix *all* the problems, we need an example (topics, map, *several* > ditavals), as simple as possible, showing *all* the features you need. > > > I'm not sure how to reply to this. We are using several custom > conditional attributes. > As we adapt our document into DITA we will probably add more. There is > nothing that we have read about DITA to indicate that there are any > "hard coded" limits to this customization. We have five new conditionals > that capture the pre-existing conditionals that were used in the > FrameMaker versions of our documents. Any combination of those five > dimensions is possible. > > I imagine we'll be adding a few more as we continue to adapt our content > to DITA.
Understood. > > The only difference between our conditionals and "platform" or "product" > or "audience" or .. is that we have defined them instead of them being > defined by the DITA spec. They operate in exactly the same way. > > For example, you'll have to show us how you specify: include this > > element but only when customer="X" and situation="Y". > > > <p customer="x" interface="y" deliverable="z">...</p> is a paragraph > that only gets included when all those things are included. If we > exclude any one of them, the paragraph as a whole is exluded. This is, > as we understand it, the way conditionals work in DITA. That's right. > The only > difference here is that we're using custom/specialized conditionals > instead of the ones hard-coded in DITA. > > This is already working! at the DITA xml level, So you confirm that, using ditac (i.e. not the ODT), there is currently *no* *problem* performing any kind of filtering/flagging you want, as long as you specify the conditionals at the DITA xml level (e.g. a paragraph). > it is only at the map > level that it is not working, which is very confusing to us. How is it > that we can conditionalize a paragraph correctly, but not a topicref? I > know, I know, you explained how ditac processes map-file topicref's > already. What baffles us is that the fundamental "conditional > processing" would be so different. > > > > Ditac currently has a low priority on our development schedule. > Therefore I'm sorry to say that it may take weeks, if not months, before > we release a new version of ditac. > > > Understood. I am wondering if you can release a new version quickly > that just flags as an error any attributes on topicrefs which are > (currently) being ignored? That would at least give us a safety net that > we aren't going to release something marked as super-secret by accident > because the topicref filtering was being ignored. :) This would take almost as long as really fixing the problem. If you confirm the above status, may be the problem is not as severe as we first thought it was. The first step for us will be to learn more about the "props" attribute and how it may be used to specify extra conditional processing attributes. (I'm just guessing: despite the fact that I've read the DITA spec several times, I don't remember *anything* related to this.) > > > > Thank you for your patience and thank you for helping us to improve > XMLmind DITA Converter (ditac). > > > Overall I am very happy with ditac. Since it is in the open source part > of the XMLmind products, is there a way that we could contribute back to > it? I see from the archives that it was released last fall (2009), so > I'm not surprised that there are still some rough edges. (This is > important to us because there are some customizations we'd like to do > with XSL processing that DITAC currently does not expose, but that is a > larger topic for another thread). > We'll happily accept any contribution which truly improves ditac, especially if this improves the *looks* of what is generated by ditac. We are still far from the typographic quality of what was generated by FrameMaker.
