Bowerbird wrote:

p.g. introduces its own linebreaks into its plain-ascii e-texts, all without ever entering the markup arena.

and i put in my own line-breaks, right here in these posts to this listserve, again without using any markup at all, just the return key.

i'll bring this thread back to relevance starting tomorrow...

Yes, if one doesn't care about internal word breaks in the original, then the markup approach is *equivalent* to the plain text break approach. Using your example above, let's suppose we want to preserve internal word breaks, then we might have (ignore starting spaces, simply used to shift the left margin inward so it's easier to see in this message): and i put in my own line-breaks, right here in these posts to this list- serve, again without using any markup at all, just the return key. So the question is, is the word which is broken "listserve" or is it "list-serve"? Does it matter? Yes, for word searching purposes, and a few other purposes. I surmise you don't think that preserving the actual internal word break in the original is important, just shift the break to the nearest intra-word break. Well, fine, but some people might want to have the information preserved. The markup approach I presented gives the optional capability to mark this up. (I'm evaluating Marcello's feedback, so there are different markup approaches that may be taken.) Btw, if we don't care about internal word breaks, and place the break at the nearest intra-word break, your original example in markup becomes (using the or: namespace): <p>and i put in my<or:lb/> own line-breaks,<or:lb/> right here in these<or:lb/> posts to this listserve,<or:lb/> again without using<or:lb/> any markup at all,<or:lb/> just the return key.</p> If the above is rendered in a web browser, and the end-user does not care about where the line breaks occur and takes no action, the web browser ignores the <or:lb/> tags, and the text is displayed nicely to fit the browser window parameters. But an ebook reading system, as well as simple CSS, can be used to "activate" the <or:lb/> at user demand. Or if there's a conversion script of the markup to plain text, such as to ZML, then we know where the breaks are. (One advantage to using <or:lb/> rather than <br/> is that browsers will ignore the <or:lb/> tag by default -- it would take CSS or a special user agent to activate them on demand.) Another advantage with using <or:lb/> is that the markup document is not restricted to exact plain text formatting. This allows a lot of latitude for document authors to do what they want in their text editor editing the XML document. For example, the above markup could be expressed in the document as: <p>and i put in my<or:lb/> own line-breaks,<or:lb/> right here in these<or:lb/> posts to this listserve,<or:lb/> again without using<or:lb/> any markup at all,<or:lb/> just the return key.</p> Or as: <p>and i put in my<or:lb/> own line-breaks,<or:lb/> right here in these<or:lb/> posts to this listserve,<or:lb/> again without using<or:lb/> any markup at all,<or:lb/> just the return key.</p> Same thing... XML parsing user agents normalize all three to the same thing. But in plain text, if someone happens to edit your text, such as to and i put in my own line-breaks, right here in these posts to this listserve, again without using any markup at all, just the return key. The line breaks are changed and the original line breaks lost forever. What if someone takes a PG text formatted in ZML, and didn't understand it was ZML (see note below), did some line length reformatting, and then redistributed that -- especially if it's Bowerbird poetry? Jon Noring (Note: How would the user know the plain text they are working with is ZML? And how would they know in a particular instance that text line breaks *are* important? Is there going to be machine-readable metadata to say that the document is ZML or that text breaks are important? I recommended that a plain text document which conforms to ZML should have some message or processing-instruction-like thing at the beginning saying it is ZML, and which version, and possibly that line breaks are important in this particular document and why. That's the purpose for <?xml version="1.0"?> at the beginning of an XML document. It identifies it and even assists with determination of the text encoding. Will ZML require UTF-8 or UTF-16? Or will it stick to ASCII? Or will it allow ISO 8859-1? Or will it allow all of them? Will it allow any text encoding? How would a user agent know the text encoding of the ZML document, especially without having to process the whole thing?)