From: Paul V. <pa...@vi...> - 2004-03-22 15:43:37
|
Hi Darcy, Here are the issues I've found in the PDF you made with Ventura. I have no idea how easy or difficult it will be to fix them, but most of them *must* be fixed because they really "break" the doc. Just an idea: would it be easier if you imported the HTML instead of the XML? Because in the HTML, the rendering has already taken place. I can easily set up a target so that what needs to become one PDF file is also in one HTML file (unlike now, where it is chunked into top-level sections). OK, here goes: - "Chapter One" shouldn't be there (single-chapter doc). - Can URL/email links be made to work? It's not critical, but it would be nice. - Speaking of which: URL links wreak havoc all over the place: they are formatted as a block (should be inline) and they "outdent" from their surrounding blockquotes, listitems etc. instead of respecting the current indentation/margins. They also tend to break away from the listitems they're supposed to belong to. - <sect1>: Leave more vertical space before it (say 1.5 times as much as now). Within an <article> (as opposed to a <book>) it would be nice if <sect1>s started on a new page (but not very important). - Blockquotes: bit more space before them (one line's worth). - <quote> tags should not translate to italics. Italics emphasize, whereas quotes often relativize. Best translate them as surrounding double-quotes. - Sometimes spaces get eaten after a tag, e.g.: "primerif" and "primersif" under "DocBook XML - An Introduction". - Internal links should work. - <programlisting>s should honour whitespace, especially newlines and start-of-line spaces/tabs. A fixed-width font is also highly desirable here. OTOH, no extra whitespace or newlines must be added. This goes wrong under "Typing text" where the text demonstrates that an "unreadable" layout has the same validity as a pretty one. - If a programlisting occurs within an indented block, it should not "outdent" from the block. Same goes for <example>s. - The element <sgmltag class="starttag"> should be translated to '<' + element content + '>', like in the HTML. If class = "endtag", it should be rendered as '</' + constant + '>'. For "emptytag", it's '<' + content + '/>' Now you see in the PDF: "italics is a start tag, italics is an end tag, ..." - you don't see the difference between the two. - The element <sgmltag class="genentity"> should be translated to '&' + element content + ';', like in the HTML. In the PDF it now says: ...you type this: lt That's an ampersand, followed by the letters l and t (for less than), followed by a semicolon. The "lt" should be "<" of course, otherwise the whole point of the text is lost. - Content of <literal> elements should stand out typographically. The usual solution is to use a fixed-width font with the same size as the surrounding text. No italics or boldface. And keep it inline. - <emphasis> should stand out, preferably as italics. But: <emphasis role="bold"> should always be rendered in boldface. - Admonitions (<caution>, <important>, <note>, <tip>, and <warning>) should be recognizable as such. They should be formatted as a block, indented or with a frame around it, possibly in a smaller font, and with some kind of caption denoting what it is: a note, a warning, a tip, ... Of course this can also be a graphic (light bulb for a tip, excl. mark for a warning etc.) - Paragraphs within a listitem shouldn't get bullets of their own - this makes them look like extra listitems. I see this happen in itemizedlists and variablelists. - listitems in an <orderedlist> should not have bullets, but (by default) Arabic numerals. Unless otherwise specified by the "numeration" attribute. - <literallayout> and <screen> MUST honour the source layout! <screen> must have fixed-width font. - Tables: <tfoot> elements should wind up at the foot of the table, not near the top. <thead> and <tfoot> content should be printed bold or italic (preferably bold) to distinguish them from ordinary rows. "align" attributes should be honoured if possible. - command, application, superscript and subscript tags should be honoured. Same for all those other tags discussed under "Various inline elements". It's OK if you render them all the same as <literal> (see above). - <computeroutput> is an inline element and shouldn't be formatted as a block. - Spaces before a smiley :-) are eaten in the PDF. - A nested list should have its own indentation level and preferably also another type of bullet. This goes wrong under "Style". Greetings, Paul Vinkenoog |
From: Darcy O'N. <ds...@sk...> - 2004-03-22 20:09:12
|
Hello, I believe 90% of these are easily fixed in Ventura, the other 10% shouldn't be much harder since Ventura supports scripting which will allow me to correct any issues. For example the ampersand issue would use a script to fix the import problem. I'll see what I can fix up in the next few days. Darcy |
From: Darcy O'N. <ds...@sk...> - 2004-03-23 20:59:50
|
Hello, Well things are moving along pretty well. Here are a few remaining issues, some discussion might help resolve them, plus it will give you more background on the process. 1. Chapter 1 - Basically, with Ventura a chapter system allows linking from the Table of Contents, also we can create one large book from multiple XML files. Even though we may have one big book, we could still have individual docs. 2. URL Links - Can be fixed, some links will have to me manually identified and created but it's a small job that I will do. 3. URL Wreaking Havoc - Mostly fix, a few small adjustments but looks good. 4a. <sect1> Vertical Space - this is a preference issue, design wise it is infinitely adjustable 4b. <sect1> New Page - Looks good for websites, absolutely horrible for books, to much white space and actually that would be a chapter if you did a page break after every <sect1> 5. Blockquotes - Fixed 6. <quote> Double Quotes - Fixed 7. Space Eater (primerif) - Fixed 8. Internal Links - Will work again might require manual intervention 9. <programlisting> White Space - This is the big issue that needs some more experimenting. Manual intervention worst case 10. <sgmltag class="starttag"> - Fixed 11. <sgmltag class="genentity"> - Fixed 12. <literal> - Fixed 13. <emphasis> - Fixed 14. Admonitions - Fixed - Graphics will be added later 15. Paragraphs within a listitem shouldn't get bullets - this will be an interesting issue as the <para> is so common but Ventura has a hard time working with nested tags. Should be able to fix but needs more research 16. listitems in an <orderedlist> - Fixed 17. <literallayout> - partially fixed 18. Tables - Future (can be fixed) 19. command, application, subscript etc. Fixed 20. <computeroutput> - Fixed 21. Smiley's - In professional docs??? 22. Nested Lists - Fixed I'll post a followup message detailing some future issues that may come up. Darcy |
From: Paul V. <pa...@vi...> - 2004-03-24 15:15:03
|
Hi Darcy, A lot of things fixed already, great! > 1. Chapter 1 - Basically, with Ventura a chapter system allows > linking from the Table of Contents, also we can create one large > book from multiple XML files. Even though we may have one big book, > we could still have individual docs. But is it mandatory that a doc (a single PDF file) have at least one chapter? It's not the end of the world, but it looks a little silly to have a document with just a "Chapter One" and no other chapters, even though it is in itself complete. Oh well, if we can't change this, we can put the Docbuilding + Docwriting Howtos in one PDF (they already form one DocBook <book> together). > 4a. <sect1> Vertical Space - this is a preference issue, design wise > it is infinitely adjustable. > 4b. <sect1> New Page - Looks good for websites, absolutely horrible > for books, to much white space and actually that would be a chapter > if you did a page break after every <sect1> OK, agreed. The point is that the Docwriting Howto turned out a lot bigger than I had originally envisaged. Some of the sections look more like chapters. No problem as long as we have that extra vertical space before <sect1>'s (and a little less before <sect2>'s, etc.) > 10. <sgmltag class="starttag"> - Fixed > 11. <sgmltag class="genentity"> - Fixed Class endtag too, I suppose? Anyway it's the same procedure. > 21. Smiley's - In professional docs??? Sure, why not? :-) This was a doc for fellow (aspiring) docwriters. Tone is a little more familiar here and there. Maybe the space got eaten because Ventura thinks there should't be a space before a colon? This has to be fixed, because there are also other situations where a space precedes a colon, e.g. in a C code fragment or in a ratio. Greetings, Paul Vinkenoog |
From: dso <ds...@sk...> - 2004-03-24 17:35:42
|
Hello, >>1. Chapter 1 - Basically, with Ventura a chapter system allows > > But is it mandatory that a doc (a single PDF file) have at least one > chapter? It's not the end of the world, but it looks a little silly to > have a document with just a "Chapter One" and no other chapters, even > though it is in itself complete. Oh well, if we can't change this, we > can put the Docbuilding + Docwriting Howtos in one PDF (they already > form one DocBook <book> together). Actually the Chapter can be removed for single doc's no problem. But if we do go to a 'complete' book then the chapters are beneficial. In the separated Chapters I just need to create a 'chapterless' template. Easy enough. >>4a. <sect1> Vertical Space - this is a preference issue, design wise >>4b. <sect1> New Page - Looks good for websites, absolutely horrible > > OK, agreed. The point is that the Docwriting Howto turned out a lot > bigger than I had originally envisaged. Some of the sections look more > like chapters. No problem as long as we have that extra vertical space > before <sect1>'s (and a little less before <sect2>'s, etc.) Can be done, I'm going make some modifications to the layout etc., and then put a copy on my website for you to review. We'll do the final adjustments as we go. >>10. <sgmltag class="starttag"> - Fixed >>11. <sgmltag class="genentity"> - Fixed > Class endtag too, I suppose? Anyway it's the same procedure. Yep, those all work. >>21. Smiley's - In professional docs??? > Sure, why not? :-) > > This was a doc for fellow (aspiring) docwriters. Tone is a little more > familiar here and there. OK, now I just need to see if Smiley's are supported in Ventura. > Maybe the space got eaten because Ventura thinks there should't be a > space before a colon? This has to be fixed, because there are also > other situations where a space precedes a colon, e.g. in a C code > fragment or in a ratio. I think maintaining 'spaces' is going to be the hardest part of these doc's. Ventura supports formatting without problems i.e. cut and paste a text doc into Ventura and it will keep the tabs and spacing perfectly. Import XML and it doesn't see the spaces because it see the doc as a XML file and removes all spacing, then applies formating to the tags. Anyway, like I said, worst case scenario requires a little hand editing prior to final publishing. Darcy |
From: Paul V. <pa...@vi...> - 2004-03-25 14:11:34
|
Hi Darcy, >> Maybe the space got eaten because Ventura thinks there should't be >> a space before a colon? This has to be fixed, because there are >> also other situations where a space precedes a colon, e.g. in a C >> code fragment or in a ratio. > I think maintaining 'spaces' is going to be the hardest part of > these doc's. Ventura supports formatting without problems i.e. cut > and paste a text doc into Ventura and it will keep the tabs and > spacing perfectly. Import XML and it doesn't see the spaces because > it see the doc as a XML file and removes all spacing, then applies > formating to the tags. I don't understand this. Sure it can remove all spacing *between* elements, i.e. it can convert <section> <para> <emphasis> Don't forget this! </emphasis> </para> </section> to: <section><para><emphasis>Don't forget this!</emphasis></para></section> But here we're talking about spacing between words etc. *inside* elements. I suppose Ventura doesn't convert <para>This is some text</para> to <para>Thisissometext</para>, and neither should it convert <para>a = x == y ? b : c</para> to <para>a=x==y?b:c</para> or <para>Hi :-)</para> to <para>Hi:-)</para> Greetings, Paul Vinkenoog |