Sorry, I've spoken too quickly. Hussein Shafie wrote: > Steve Sarette wrote: >> I also discovered that copying the XInclude and then pasting it into a >> different location in the document causes XmlMind to paste just the >> text that the XInclude resolves to -- and not the XInclude itself. >> This is unfortunate because if that copy/paste operation would work >> then I could avoid a lot of time in the menus. Also, if I copy a block >> of text that has an XInclude and then paste it elsewhere in the >> document, the XInclude is lost. This is a bug, IMO. > > > This is not a bug but rather a serious limitation. In practice, this > limitation is so serious that we'll try to remove it in next release.
We didn't find a clean, simple and reliable way to implement this. >> For the moment I'm just putting special character strings into my >> documents as I write and then once I get done with any given file, I >> close it and then replace those character strings with the appropriate >> XInclude statement using vi. >> >> You guys have an otherwise pretty good product, and this is the only >> major area where I find that it falls down. I hope you can forward >> this email on to your engineering team for consideration, or file some >> feature requests or something. >> >> Why is it that your engineering team doesn't believe in shared text >> references anyway? I was quite surprised to find you eliminating them >> on document open. > > > Our engineering team clearly doesn't believe in shared *text* > references, nor in character references, nor in the necessity of > creating beautifully indented XML. > > But even if we are not convinced by this way of creating/editing > structured content, our customers and Standard Edition users have > complained too loudly and for a too long time. > > Therefore, we'll do a major effort in these areas: > * Character references. > * *pure text* references (i.e. not a mix of text and element). > * beautifully indented XML. > > After that, you'll no longer need to use vi and Alexander Dupuy will no > longer need to use unxxe.pl. Here too, we didn't find a clean, simple and reliable way to implement references to pure text entities. We can just give you our word that in the next release we'll have a good support of character entities (i.e. such as " ") and that the generated XML files will look much better.
