---Oleg Kobchenko wrote: > > From: Sherlock, Ric > > > > ---Oleg Kobchenko wrote: > > > This already looks much better. > > > A few things about egpage2a.html though: > > > > > > * white space is important, it must be symmetrical, > > > in fact I much prefer the same white space > > > treatment as in the original dictionary HTML > > > > Are you referring to the vertical whitespace around the > > names? There shouldn't be any problem adjusting the CSS > > to provide the desired white space, but of course personal > > preferences will come in to play. I have made a formatting > > change that I think improves the visual clarity of the > > page. > > http://www.massey.ac.nz/~rsherloc/jdict/egpage2a.html > > I actually meant white space in the source HTML, > ie indentation, etc. Sorry for confusion.
I added the indentation to try and show the underlying structure of the source better, but have no problem with removing it. I don't particularly mind how the raw HTML is stored, that's probably best left to the maintainer. There are plenty of tools that will indent HTML if desired. I have updated the page source to try and emulate the white space conventions of the original HTML. > > > * I gather, <!--top jump start--> and > > <divid="Header"> > > > should be mutually exclusive > > > > > * I am not sure <divid="Examples"> > > has much merit, > > > with <P> addition, the rest of the page can > > be just > > > thought as part of the body. > > > > > > * The same refers to > > <divid="Content">. > > > It could also be just part of body. > > > > By removing all these divs it would no longer be possible > > to specifically target tags (<a> <li> <p> > > etc) in the Header vs the Content, and distinguishing > > different tag formatting for pdefn vs Examples would be > > more awkward. I would argue that the Header, Content, > > Examples, Exercises and Footer divs add enough benefit, for > > a small increase in complexity, to be retained. > > Why if we keep the Header and Footer DIVs? > While others would be top-level, but without enclosure > inside Content, Examples etc. Sorry, I had misunderstood from your previous post that you were suggesting that given <!--top jump start-->, there was no need for <div id="Header">. Nevertheless, even if Header & Footer DIVs were kept, I think it is still probable for example that you would want to target and format an ordered list in Exercises differently, from an ordered list in the top-level content of a page like http://www.jsoftware.com/help/dictionary/intro09.htm . An Exercises/Examples div makes that easy. The Content div would then provide a simple way to target all content that isn't a header or footer including Excercises/Examples. > > > * separation of symbol and ranks is too fussy. Either > > putting > > > them all in one TD with a class or id, or at most > > separating > > > just symbol and ranks together in spans: > > > <span > > id="sym">^</span><span > > id="rank">0 0 0</span> > > > but this is too much, as they could be easily split > > by space. > > > > To me they are logically different elements and it would > > therefore be better to separate them using markup even if > > they are not separated visually. If you were going to > > create an index for the pages, should the index entry be > > "^" or "^ 0 0 0" ? > > XPath: substring-before(//[EMAIL PROTECTED]'symrank'], ' ') This shows it is possible to pull out the symbol for an index, but it still wouldn't be possible to target them separately using CSS (if that were desired at some point)? > I believe it is better to keep it simple and return to > the original two tables: > > <table id="SymHead" width="100%"><tr> > <th id="MonadName" align="left" width="33%">Example</th> > <th id="SymRank" align="center" width="34%">^ 0 0 0</th> > <th id="DyadName" align="right" width="33%">Example</th> > </table> > > <table id="SymDef" width="100%"><tr> > <td id="MonadDef" width="49%">Description of <tt>x</tt> ...</td> > <td id="Spacer" width="1%"> </td> > <td id="DyadDef" width="50%">Description of <tt>y</tt> ...</td> > </table> > > (Possibly widths and alignment placed in CSS). I agree the use of IDs to label the cells and CSS to format them is preferable to the current markup, and I imagine it would be easier to automate a move to this style of markup than to the others proposed. However (as I've probably already belaboured enough) I would prefer to see the HTML markup represent the logical structure of the document elements rather than confuse that structure by the use of tables as a simple page layout mechanism without regard to the logical structure. If the consensus is that that would require too much effort, then so be it. If a two table structure were to be retained, I'd suggest that the spacer column in the 2nd table be removed. It is pure formatting rather than content and the desired spacing between the two columns can be added using CSS. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
