Le 04/04/2012 03:41, Bruce Dubbs a écrit : > Jean-Philippe MENGUAL wrote: >> Hi, >>> I don't understand what layout you mean. There was no discussion about >>> changing the layout. >> Here's a very short example. Instead of: >> <para>To test, issue:</para> >> >> We have: >> <para> >> To test, issue >> </para> > Oh, yes. I do that a lot for readability of the source. For xml, > newlines are meaningless except in specific cases, e.g.<screen>. It's > quite a bit like html (both xml and html were derived from sgml). I > thought you meant the xml files had embedded newlines like a dos file. > >>>> We note also that the english version of the book is not consistent >>>> about this stuff: some commits keep the former layout; some commits >>>> change the layout (with \n after tags), and also some commits fix a >>>> layout change. > Yes, that's just the editor. > >>>> For example, in the file commited in r9802 >>>> (general/prog/python-modules.xml), the author remove the >>>> lines<segtitle>Installed Program</segtitle> and<seg>None</seg> when no >>>> program is installed . In all other pages of thebook, these lines are >>>> present. >>> I wondered why Andy did that, but I didn't think it was an issue that >>> really mattered. >> Actually the problem mentioned by Denis is that this evolution is > not in the whole book, but only in some commits (and some commits > revert such change). And you're right, theorically it doesn't matter, > but the problem here is that svn diff become not relevant. Instead of > displaying differences in commands, words, descriptions, diff displays > as if all the paragraph was changed, although nothing has changed except > this evolution. As we do our commits from the commit mail on blfs-book, > the fact diff isn't relevant anymore is a bit fastidious. Example: > >> <para>No test are available.</para> >> <para> >> No test are available. >> </para> >> >> diff displays such change, whereas actually, nothing is changed in >> the text. So don't care such change. > I guess I don't understand why a source format change gives you such a > problem. It doesn't seem reasonable for us to not change the source. > Lets say we want to add a word to a paragraph and then reformat the > paragraph. If we don't reformat, multiple changes can make for ugly > source and make it more difficult to proff read. It would only show up > as a single line diff for you, but a reformat can change the whole > paragraph. > > Many times the reformat is in the dependencies where there is little for > you to change. Other times it's in the<screen> section where aligning > things is in the output and can make instructions much clearer to the > reader. I'd think in those cases, a cut/paste from the English version > to the French version would be fairly straight forward.
No problem for all changes that be more readable the book. My comment was just for changes in formatting xml, with no changes in the html or pdf version of the book. In fact, the real problem for us is to read a very long diff with some changes in many lines of change which can be ignored. > What you are asking for is for us to forgo restructuring the xml source. > I'm very reluctant to do that. > >> Same problem with<seg>, in dependencies descriptions. >> >> Denis adds to that the fact that sometimes we write "Installed >> Programs" then None; sometimes we don't write the entry at all. So not >> consistent. > That's true right now. I'd prefer to see the<seg>None</seg> left in > myself. It is the point that i note "no consistent". I do not mean if it's better or not to remove these lines, but maybe it would be good to choose one way to present the book. And, as says Jean-Philippe, it's a pleasure for us to translate your work... > -- Bruce > Denis -- http://linuxfromscratch.org/mailman/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
