How to create a screen-format PDF book

[Michael Hopcroft]

I have a technical question: I was psoting on the future of the PDF format on another board (I was not optimistic) and I got this comment.

I'm quite happy to buy pdfs of stuff that I want, but it really ticks me off when I get something which has been poorly produced. I expect to get a properly linked TOC at the very least, and if the document is intended primarily for on-screen display I expect the pages to be formatted so that a page will fit happily on screen without having to scroll up and down to read columns of text. If it's primarily intended for print, then I don't expect to get something filled with low-resolution pixelated jpegs.

Also, and this is a biggie, I expect the product to be priced taking into account that I'm going to have to spend time and money to get the thing printed. I will absolutely never buy a pdf which is as expensive, or even close to being as expensive as a printed book.

In my experience, most pdfs are just dumped straight from whatever layout program the author used for the print layout, without any consideration for the benefits and possibilities of the pdf format. Come on people, it's the 21st century for crying out loud!

Given all this, it caused me to think. I've never published a PDF laid out for the screen, but if I'm not going tobe printing these books, it could be a viable option. the question is -- hwo technically do I do it to make it work? What software do I need to have -- is Acrobat essential, or can I use features in Quark XPress or Pagemkaer to perform the needed functions? Which dimensions should I use to get a page to fit on a CRT screen? Is full color an option if we're not going to be printing the book?

So many questions, so few answers, and such horrid typing.

[quozl]

The best format for the screen is HTML.  It's also the easiest to do.

[Eero Tuovinen]

Given all this, it caused me to think. I've never published a PDF laid out for the screen, but if I'm not going tobe printing these books, it could be a viable option. the question is -- hwo technically do I do it to make it work?

Optimally you'll be needing ye olde layout program that can output EPS or something else Acrobat reads. Then you're going to need Acrobat itself with Distiller included. That's the easy route. There are various ways of doing PDFs without Acrobat or one of the other compilators, but it's generally a real hassle to get layers/links/indexing/any of the other kewl features. Luckily you have the Forge, and you can simply ask someone who has Acrobat to do the final stages fo you...

What software do I need to have -- is Acrobat essential, or can I use features in Quark XPress or Pagemkaer to perform the needed functions? Which dimensions should I use to get a page to fit on a CRT screen? Is full color an option if we're not going to be printing the book?

You can do with the layout programs anything you could do to a paper version, including a kewl page dimension suited to PDF. Any other features will require Acrobat or some other program that does PDF (which you'll conceivably need anyway to make the PDF, assuming you won't use an online service or something like that).

The best dimensions depend on whether you're planning to have any extraneous toolbars (like an index of bookmarks) on the screen when reading, and how many low-end screen resolutions you're willing to chuck. For example, if you're doing 1280x1024 or higher, you have to lay out the work to be a comfortable read in a window of that resolution with your planned toolbars eating a part of the window. It's just math, really, after figuring out what conditions you're going to optimize for.

If you're not going to do a printable (I'd make both a screen and print layout version of any product intended to be the flag of my company, if not of every product) version, you can basicly do anything you wish as far as colors and such go. Knock yourself crazy, put in animations and sound in if you wish. It's important to decide beforehand and have it absolutely clear what you're doing - is it a print version for home printers, a print version for color printers, a screen version for this or that resolution or an amalgam (multiple layers of layout in the same file; larger size, easier to use) or a crossbreed (not optimized for print or screen, but instead made as suitable as possible to both). Otherwise you'll have incoherent design ;)

I've never actually made complex PDFs (anything meant for screen instead of going straight to press) myself, so it's definitely an interesting project. If you cannot solve the problem in any other way, PM me and maybe we can work out something. I'm not interested in monetary payment, but we'd have to work around my itinerary and you'd have to take the risk of getting a suboptimal work if I'd botch it (obviously no guarantees). I have access to Acrobat and the other stuff, though.

By the by: quozl is right about HTML being the better choice for screen work, if you have a skilled coder to do it (I'm talking at least CSS and rudimentary server/client scripting, and preferably the rest, too). It being easier is not true, though, in my experience; if you already know how to lay out work in Quark or Pagemaker the PDF conversion is only an afterthought to that. I'd say, after mastering layout this year and webwork before, that the bother of learning layout and pdf combined is about the same as learning HTML and CSS. If you have to do scripts, the web option becomes the harder one (again depending on if you already know a programming language). It's cheap, though, and gives options not available in PDF.

[Andrew Martin]

I'm an amateur at producing PDFs, so take what I say carefully and with a grain of salt.

I've tried various free ways of producing PDFs but the best I've found so far is to use the very latest version of Open Office (OO), layout the pages and text so that each page is in landscape orientation (the default is portrait mode), and have two columns per page with about a 1 cm (a bit less that 1/2 inch for Americans using the Imperial system) margin between the columns. After exporting to PDF (this is a menu option in Open Office), this allows the entire page to be shown on the screen in Adobe reader and produces a nice printed version. The export process in OO allows optimization for either printing or screen layout.

Otherwise as quozl indicates, the best format for the screen is HTML, XHTML or XML. OO allows saving as HTML so this can be an alternative?

[jdagna]

HTML may be the best screen format, but it's lousy for real layout work.  Never mind the lack fine-tuning features (like changing tracking, leading, horizontal/vertical scaling, word-wrapping on curved edges and more), it just fails miserably for RPG documents.  How do you tell someone to look at page 27 when every computer will paginate differently?  

Anyway... Quark will do most of what you want, Michael.  Just export as a PDF straight from the program.  The file you get will not be as small as Acrobat can get it, but it will work*.  Quark (version 6 anyway) will also create indexes, tables of contents and such with automatic links to the PDF pages so that you don't have to do that by hand.

In terms of size, I'd recommend playing around with it yourself and getting your playtesters' opinion.  Assume a screen size of 800x600 and see how it shows up.  Ideally, a screen-format page will put one page of text up on the screen at a time, and still have the text at a readable size.  I saw someone recommend a size of about half a standard page, laid down portrait size (about 6" wide by 4.5" tall).  Generally, anything with a 4:3 ratio will fit the screen, but that's ignoring tool bars and the side menu for bookmarks or page previews.

You can always try getting some screen-optimized PDFs.  Donjon, for example, has both a screen version and a print version included when you buy it, and Clinton did a good job with both.

*The size savings on my files tends to be about 50% when I run them through Acrobat's PDF optimizer, which may well make Acrobat worthwhile.

[Peter Hollinghurst]

Speaking as someone who works with both html and pdf formats, I would suggest that html is in no way superior for screen work. It is certainly easier for people with html writing experience and no dtp knowledge to use, but it is far far inferior a tool for control of the page layout. A good DTP prog like quark or in desgin gives you control of everything. You cannot do professional page layout in html, even with css. The original point about scrolling is also an issue for any method, and imo pdf is actually better for handling that as well, since the acrobat reader actually gives better control of viewing for page sizes than a browser normally does. Use html for web pages certainly-but I would think twice before trying to use it to produce anything else, and if I could get web pages that had all the layout advantages of proper DTP software I would ditch html for web design like a shot.
The scrolling issue has been dealt with particularly well by the pdf publishers that use a landscape format btw-a few are easily available via rpgnow.com, so I would suggest checking them out and seeing how they handle it.

[philreed]

Landscape is definitely the way to go for screen PDFs. They're very easy to read on almost any screen (I use 12 point Times for my screen PDFs and I find that's easy to read on any monitor).

There are, in fact, hundreds of PDFs on RPGNow in this format. It's a simple matter to track down an example and I'm surprised someone would even ask here before searching for samples.

[Eero Tuovinen]

Wasn't this PDF versus HTML thing discussed already just a while ago? Anyway, what proponents of PDF invariably disregard is that HTML/CSS is not a layout tool in the old sense. Nobody said that you can get the same exact layout in HTML, or that you could control font features nilly-willy. Instead, what you get is a dynamically client-rendered view from a program whose sole purpose is to view data in as unagravating way as possible. Features of web pages include things like automatic scaling for screen resolution and color, client controlled font sizes, easy updating of data, dynamic reflowing of text, real-time control of the amount and nature of graphics - all very hard to implement for PDFs. The philosophy of each is so different that it's laughable to try to compare them as ways of representation; if you have to control the placement of every letter yourself, obviously the tool meant for the job is what you need. That'd be PDF. On the other hand, if you want maximal adaptability and instant updates, web pages are the solution.

So IMO the original point stands: although web pages are currently harder to do, they would be the perfect tool for representing information on computer screen. It's not gutenbergian layout, but rather a dynamic of the future, with both the designer and the reader participating in the creation of the final viewing experience - it's different, yes, but so is the whole idea of using computers for reading. Currently coding a high-end web page is pure hell with all the browser incompatibilities and unfulfilled standards, but if you disregard the effort angle, a good web page is certainly pound for pound a better and more flexible way of representation than a static PDF. The reason I recommend PDF for most is that the work involved is, like, ten times more with a good web page than with a good PDF.

[Keith Senkowski]

I guess the first question is what software are you using to create the PDF?  With InDesign, when you make the PDF you get the option of choosing the compression rate.  It just so happens that one is labeled Screen which will do you just fine.  If you are using Acrobat Professional there is a tab under Advanced called PDF Optimizer which allows you to dumb it down to screen resolutions.  If you are using Quark I would have to get back to you.  I haven't used it in about a year.

Hope this helps

Keith

[quozl]

So IMO the original point stands: although web pages are currently harder to do, they would be the perfect tool for representing information on computer screen.

I'm surprised to see that you think HTML is harder.  How so?  We're talking header and paragraph tags with linking.  How is that hard?

[quozl]

How do you tell someone to look at page 27 when every computer will paginate differently?  

A link.

[Eero Tuovinen]

So IMO the original point stands: although web pages are currently harder to do, they would be the perfect tool for representing information on computer screen.

I'm surprised to see that you think HTML is harder.  How so?  We're talking header and paragraph tags with linking.  How is that hard?

It depends on what you're looking for. I was thinking of the absolute high-end: big site (which has to be controlled through CSS or dynamic page generation), kewl effects (columns, all kinds of responding interface), absolute compliancy to both the standard and IE models (meaning all kinds of obscure hacks healthy people should never have to learn). I think that this is harder than for example Quark+PDF if you have no basics in either. Experience speaks here; I just now learned to do complex CSS, and you can go see what my site looks like after a couple of weeks of intensive fiddling. To compare, it took me three days to learn both Quark and Acrobat to a degree sufficient for print publication. In the latter case I started from knowing Pagemaker already, but in the former I knew HTML, so it balances.

Then again, you are perfectly correct in that most content will be perfectly happy with a tastefully done basic HTML+CSS representation. It's only with post-96 technology that problems start to appear. In that sense it's true that it's easier to do HTML than a PDF: it all depends on what you want, really.

Based on the discussion I'd maybe scale different publication formats in the following way based on the effort required from nill to produce them:

Text in a file
Text editor files (most have Word)
Basic HTML (tags, no nested tables)
Basic HTML+CSS (no controversial selectors/properties)
Print layout (small book, basic layout)
Basic PDF (optimized for screen, nothing fancier than links)
Advanced HTML (nested tables, common hacks, basic scripting)
Advanced PDF (everything from scripting to layers to sound)
Advanced print layout (mucho graphics, hundreds of pages)
HTML+scripting (script, dynamically generated pages)
Full HTML+CSS/XHML (everything currently doable)

As you can see, different HTML implementations reside on my list in both extremes. It should be however noted that if we have to factor in the cost of tools HTML wins almost every time - while you can only produce an average PDF with free tools, you certainly can do the absolute best in HTML representation with only a text editor and a bit of balls.

[Alex Johnson]

It should be however noted that if we have to factor in the cost of tools HTML wins almost every time - while you can only produce an average PDF with free tools, you certainly can do the absolute best in HTML representation with only a text editor and a bit of balls.

I must have fallen asleep at the wheel.  Giving this thread a little bump...

LaTex can be done with far more professional results than HTML and is 100% free and besides the compiler can be written entirely with your favorite text editor.  Of course you pay the price in complexity, and it isn't WYSIWYG.  But you get exactly what you ask for.  But that's a layout description language, and not necessarily the best suited for display on a screen (as opposed to a print).  I just noticed you neglected it in your extensive list of publishing methods.

[viktor_haag]

Given all this, it caused me to think. I've never published a PDF laid out for the screen, but if I'm not going tobe printing these books, it could be a viable option. the question is -- hwo technically do I do it to make it work? What software do I need to have -- is Acrobat essential, or can I use features in Quark XPress or Pagemkaer to perform the needed functions? Which dimensions should I use to get a page to fit on a CRT screen? Is full color an option if we're not going to be printing the book?

Getting back to the original questions:

(a) How do you make it work? If you work on a Macintosh with OSX, you don't need any extra software to make PDFs; any document you can print, you can make a PDF from since OSX's display and print engine is built on PDF. If you work on another platform, then you need a printer driver that can produce PostScript output; then you need software to turn that PS into PDF. You can use Acrobat for this, but there are a number of free alternatives to Acrobat; they are generally of inferior quality to Acrobat, but they are serviceable.

(b) Using Quark/InDesign/PM. I believe that Quark and InDesign on all platforms can produce PDF directly for you. PageMaker I'm not so sure about. But in any event, with some elbow grease you can add a PS-to-PDF filter for free (see point 'a'), or plump for Acrobat.

(c) What dimensions should I use for landscape pages? The trick with PDF is that source size is not nearly so relevant as page orientation. Acrobat Reader (and most PDF readers) can scale the PDF page as needed/desired by the viewer. Accordingly, a landscape page that uses the same display ratio as the reader's monitor is going to make optimal use of original size (that is, it can be maginified up in its original H/W ratio better than a page with a H/W ratio that doesn't match the monitor). Since most folks have a 4:3 monitor, if you're really designing for the screen your page size should have a 4:3 ratio. Keep in mind that smaller font sizes (even "apparently small") are not necessarily harder to read, as long as they have sufficient leading and white space around them; however, also keep in mind that anti-aliasing (font smoothing) tends to "smudge" fonts at smaller font sizes, so you want the "apparent" font size in the end result to be somewhere between 10 and 14 point for most people (12 point seems the happy medium). I would do product testing on 800x600 and 1024x768 screen resolutions; although most people seem to use their desktop monitors as grossly large resolutions, you might also want to do product testing on a "standard laptop" resolution, since laptop users may find the page not terribly useful for them if you design it for the 800x600 monitor user.

(d) Full colour is of course an option, and is also an option for printing as well. Most modern 1200 dpi lasers can convert colour into grayscale reasonably well, and without too much extra work. If you're thinking about a dual purpose book (screen and landscape paper) keep in mind that you want the majority of the page to be white, as colour of any sort on the page will get grayscaled by the print engine and thus use toner/ink. However, if the PDF is solely for use on-screen, then background colour also becomes an option. If you want to go this route, then you'll also want to think very carefully about your colour combinations, as slight mis-steps there can cause eye fatigue very quickly. In my past projects, I've tended to use spot-colour only (coloured text headings with the odd coloured chotchkie here and there) to punch out appropriate page elements. I've also found that a very pale corn yellow or light grey can often make a soothing page background, but then you have to be very careful about what you lay on top of it.

General comment: if you're going to be producing two versions of the same work (screen plus a print version) then I urge you to carefully select your tools. If at all possible you want to select tools that will let you produce both versions of the work with a minimum amount of hassle, and from the same source text. Otherwise, the cost to make changes to the work in progress, or later on, becomes quite high.

If you're working on Windows, I would highly recommend FrameMaker as the tool of choice for works that need a modicum of layout, but not strenuous layout. As a guide, I would say GURPS, HERO, Call of Cthulhu, HeroQuest, Pendragon, Traveller (any version), and Sorcerer, are all examples of the kinds of books that FrameMaker is an excellent tool at handling. It's long book support (tables, indices, contents, variables, conditional text) is second to none as a package; other products provide slightly better support in some areas, but as a package, Frame is THE best tool for 80% of the work produced in the RPG industry. Oddly enough, I'd say that only a fraction of the RPG market actually uses Frame (and this may explain several things about their problems with layout glitches, and production time).

If you're on the Macintosh, and you're not afraid of using an end-of-lifed product in Classic mode, Frame is still the best choice. However, Adobe won't support it any more, and you may find that, with an OS upgrade one day, it may suddenly stop working (i.e. I'm crossing my fingers to see what the Panther upgrade does to Frame). However, I personally still use Frame almost exclusively for the work I do; it is mostly solid, and its feature set is still strong. I have not had any more stability issues with Frame on OSX than I have had with any other mainstream product (like Word, for example).

If your layout needs tend to the more flashy (for example, the kinds of layouts used by WotC or White Wolf) where there is a lot "going on" on the page, and you need a great variety of graphic appeal and layout variation, then you'll need a tool dedicated to the layout task, and not necessarily the text manipulating task. There are several low-end page layout tools that are also low-cost: Ragtime and Pagestream leap immediately to mind. PageMaker is also available at slightly higher cost, but support for it won't last much longer. At the high end you have Quark and InDesign. Currently InDesign's typographic features cannot be matched; however, Quark has a larger installed support base (i.e. when it comes time to talk with printers and so on). If you're at all serious about your page layout needs, then I would recommend spending the extra money and buying the big-guns. You will be able to drive them much farther into the future than any other choice; however, they are costly. If you're on a budget, then I would take a serious look at PageStream. Although not many people know of it, it's been around for a long time (it's built on a mature engine), and it's much less expensive than the heavyweights on the market.

As for HTML, it may be useful as an output format, but I would select a tool based on its use as a writing and layout environment for print/PDF first and formost; then, you can look at its ability to produce HTML output. You do NOT want to be trying to write your stuff directly in HTML, especially if you have serious layout needs. The precise kinds of things that the RPG market requires, most direct HTML production environments are not particularly good at (multi-column pages, rich tables, cross-referencing).

Finally, as for LaTeX, while I have a great deal of respect for what TeX and LaTeX can do, I really don't think it's the right tool for the job here. Once again, while what LaTeX does it does supremely well (serve as a writing environment for scholarly or research-oriented work), it falls down in areas that are really considered "de rigeur" in the RPG workspace (rich use of tables, multi-column pages, sidebars, graphics, variety of fonts).

If price were no object, and you work on Windows, I would say you should decide whether you want to produce books that look like GURPS/HERO or that look like White Wolf. If the former, then Frame is the best tool for the job. If the latter, then InDesign or Quark. I lean toward InDesign currently because of their superlative typography tools; however, if you're more familiar with Quark, then it's no slouch.

If price *is* an object, or you work on a Mac, then I reservedly recommend Frame (classic app, no longer supported, however, still a mature, capable product). The word processor 'Mellel' is showing signs that it wants to move into some of the space left vacant by Frame, and their development staff is quite keen (and the price is very, very low); however, as a professional-level tool it currently falls a bit short of the mark. Might be worth looking at it, or Nisus, as an investment in those company's futures. For DTP on a Mac, if you can't afford Quark/InDesign, then I say look at PageStream or RagTime. Also Stone Create might be a useful alternative (it's been around for ages, although most people don't know it, since Stone was originally a NeXT development house).

Hopefully, this long screed is of some use. My apologies for the lack of brevity.[/b]

[viktor_haag]

I mentioned "Panther upgrade" in my preceeding, lengthy post; what I meant was "Tiger upgrade" (i.e. the next version of OSX due in about six months or so). Frame runs quite nicely, thank-you under Panther. It has had intermittent crashing problems, but it's nowhere near as problematic as MS Word in that regard. If using Frame in Classic, it's quite sensible to set it to auto-save every few minutes (I have mine set to five min); Frame is typically quite good about (a) not corrupting its files when crashing, and (b) attempting to save out to recovery files before dying. Much MUCH better than Word in that regard.

When Frame does crash on me, I have never to my recollection lost more than five minutes worth of work.

One final caveat with Frame -- because it's a classic application, it can't make use of OSX's OpenType support, nor can it at all handle "shadowed" fonts. Other than that, it's still very serviceable on OSX 10.3 (Panther).