WDPG Final Report

The Project

First published in book form in 1891, Oscar Wilde’s The Picture of Dorian Gray is one of the iconic novels of the nineteenth century. It is one of those novels that everyone should read.

Unfortunately, the general style of Victorian literature is no longer accessible for a large part of the contemporary audience, even if Dorian Gray is on the better side of that spectrum. Our mandate at Wilde Dorian Publishing Group—named after this novel—is to change that. We want to bring new audiences to appreciate the classics of Victorian literature.

The Victorian era has seen a recent return to popularity, due in part to the growth of the steampunk sub-genre of science fiction. Steampunk is inspired by the steam-powered machinery of the nineteenth century, and gears, pistons and the like constitute a significant part of that aesthetic.

That gave us our lead. To modernize The Picture of Dorian Gray, we decided to give it a steampunk twist. We’ve left the text largely untouched, preserving Wilde’s plot and characters; but we’ve made very specific changes to the nature of the deterioration visible in Dorian Gray’s portrait. Instead of becoming old and decrepit, Dorian’s image becomes more and more mechanical.

Aside from that, we’ve also updated the spelling of various words and place names. For example, we’ve removed the hyphen from “to-day” and related words to facilitate the reading for unfamiliar audiences.

The Print Project

Our print product is more a gift item than a traditional book. To fully play into the steampunk aesthetic, our book design is inspired by the style of the Franklin Library. Though it is not directly Victorian, it incorporates a lot of the features that have become associated with a high-end Victorian era book: a leather-bound hardcover with gold-foiled decorations. We’ve also added a mechanical clasp that reinforces the steampunk style.

On the inside, we used fonts that are reminiscent of the nineteenth century, and added gear graphics as chapter markers to reinforce the steampunk idea even when the deteriorating portrait of Dorian Gray isn’t mentioned. Finally, although we did not have a chance to explore this further, we feel the book would strongly benefit from illustrations, probably between chapters. In particular, one illustration should be used as a frontispiece, showing the mechanical portrait in its later stages. That way, any potential buyer who opens the book would understand right away that this is a steampunk twist to the novel.

The ePub Project

Two ePub versions of the book are planned; the first, to complement and promote the print version and the second, to be used as a promotional tool by the Steampunk fashion industry.

To promote the print copy, the ePub version re-uses the same styles and graphics, and a similar cover. The last page of the ePub directs the reader to the publisher’s website.

To promote the Steampunk fashion industry, the ePub presents the quality of images expected of a catalogue.  These images or photographs are used in a manner to blend in with the book  so not to be seen as embedded advertisement.  The last page of the ePub directs the reader to the various sponsors of the fashionably dressed characters represented at the beginning of each chapter.

The Tool: WordPress+Ickmull

The WordPress+Ickmull system was developed in-house at SFU’s Publishing Department during a previous MPub Technology Project. It combines the versatility of WordPress for editing, content management, and so on, with an XHTML export that can be used both for print and for e-book.

The main advantages to WordPress are that it is user-friendly, versatile, and backed by a strong support community. Its functionality is also highly expandable with plugins—there’s probably a plugin out there that does what you want it to.

By default, WP offers the ability to keep content private, which means only those involved in the project are able to see it; to compare revisions and revert to an earlier version; and to keep up a library of images and other media, among many other features. The most recent WP update also locks a page or post that is being edited—that way, you can’t have multiple users editing it at once, which would cause version conflicts.

Exporting with Ickmull

In order to optimize the Ickmull output, an editor needs to mark up the WordPress content with various designations that can be styled later. HTML tags like <h1>, <p>, or <blockquote>, or CSS classes, are the best ways to do this. CSS IDs do not survive the process. The tags and classes can be inserted by hand using WordPress’s Text Editor, which allows users to see their page’s HTML. Alternatively, by using a plugin called Ultimate TinyMCE, classes can be quickly applied to selected text on WP’s Visual Editor, a WYSIWYG editor. (Unfortunately we learned of this plugin too late in the process to make much use of it. However, we feel it could be a very useful and versatile tool for both the editorial and markup stages, due to its wide variety of features.)

To get the Ickmull XHTML export, we need the Ickmull plugin installed, activated and placed as a widget. This adds a “Save as Clean XHTML” link to the widget area of your page. The plugin itself and other necessary files are all available from the Ickmull Google code page.

Using Ickmull for Print

In order for InDesign, the industry standard software for print design, to understand it, the XHTML export needs to be converted to the ICML format. ICML is an XML file format used by Adobe InCopy, a software mostly used by magazines and newspapers, and that can be placed into an InDesign file and designed. When you place the ICML into InDesign, all the class styles it uses get imported as Paragraph Styles; and when the ICML file gets updated, it only takes a click to update it in InDesign, without affecting any of the design that has already been done. That way, the design process can begin before the content is finished without much hassle.

To convert the Ickmull XHTML export to ICML, you need to use the XSLT transform included in the Ickmull downloads. On a Mac, you run a simple script on Terminal to make the conversion. The script can also be found on the Google code page.

Problems with Ickmull for Print

Most of the complications with the Ickmull system arise at this conversion stage. As mentioned earlier, the conversion ignores CSS IDs, as well as images. We had to reinsert our graphic chapter markers in InDesign, although we left them in the WordPress page for the e-book flow.

The Terminal script may also encounter errors that prevent the conversion. Thankfully, Terminal gives line references with its error messages, so they are easy to find and correct. All of the errors we came across were either tag mismatches or special characters. WordPress understand some tags implicitly, so a simple <p> tag, for example, doesn’t need to be included when you write and mark up the content. However, it gets added to the HTML output and the XHTML export. Depending on how you place any other tags you may be using, such as <div>, the open/close tags can get mismatched and cause the error.

As for special characters, the conversion understands many of them. Among the ones that are not, the most prominent is a non-breaking space. For example, Chicago and some house styles recommend ellipses be set with spaces between the dots. However, if the ellipsis falls at the end of a line, this could cause the dots to be separated into different lines. The best way around this is to use a non-breaking space character, which keeps the ellipsis whole. The most common way to insert non-breaking spaces is the &nbsp; entity, but this is not understood and causes an error. One way around this is to use the numeric code   instead, which does get understood. It just means having to look up the correct numeric codes for the special characters you need.

In theory, the Ickmull process could be done with the entire book in a single page, or with each chapter or section separated into its own page. Since each page needs to be exported and converted individually, this can be a rather labour-intensive process if using multiple pages—the more chapters you have, the more work it is. For this reason we decided to use a single page. The conversion goes faster, but the ICML file is rather heavy and can slow down InDesign (possibly even make it crash, for longer books). It is also unwieldy for editing, clean-up and markup; if we were using multiple pages, we could have different editors working on different chapters, and the process would be quicker.

Finally, a relatively minor issue is that the widget makes the export link publicly visible. That is part of the reason we kept our working pages private; however, the link is still available on our marketing pages, which are public. This is not a big problem since we are working with public domain material, but it is definitely not desirable for someone working with copyrighted material.

(NOTE: We also came across one issue that is specific to this project because we don’t have FTP access to the server we are using. The problem is that WordPress automatically inserts <p> tags around images, which caused the ICML conversion to understand our chapter markers as the first paragraph of each chapter, which affected our print design. With FTP access to the server, we could have added a simple PHP function to our WP theme to prevent the <p> tags from being added.)

Using WordPress+Ickmull for ePub

ePub with inDesign

inDesign allows you to generate an ePub,  but —the ePub created is incomplete and requires modifications.

ePubs are xHTML-based which means that tools that create and style flowable content should be favoured.  InDesign is a tool created for styling fixed layout documents, in other words print publishing.

InDesign needed to prove that it could style flowable content just as well as fixed layout content and assemble this content in an ePub.

The ePub generated by inDesign was not of a sufficient quality as to be marketable.

inDesign CSS

inDesign does produce a CSS that preserve the character and block styles created for the print version.  We considered reusing the CSS in our Calibre template, but decided against it.  —The inDesign CSS was specifically targeting inDesign paragraphs classes; re-using it would have required modifications to the CSS or to the HTML.

Taking into consideration the limited CSS support offered by the eReaders, the generic CSS template created for the Calibre ePub was more than sufficient and did not require modifications.

ePub with Calibre

Calibre is a software that is used to assemble HTML, CSS and graphic files and package them into ePubs.

With Calibre, the metadata file is created using a dialog window.  All the files required for the ePub are listed in a textbox ; these files will be loaded and included in the ePub manifest.  The cover graphic is identified and loaded separately.

Calibre allows for a single HTML file to be loaded and chunked into chapters according to an XPath formula.  After all the parameters and preferences for the creation of an ePub have been entered, an ePub and its XML ePub template file are created.  The XML ePub template can be modified and used to create other ePubs.

Any additional modification required by the ePub can be done using an XML editor such as XML Oxygen.

The xHTML saved by Ickmull was used to generate the ePubs from Calibre.

Suggestions for Improvement

To make the Ickmull system more widely and simply usable, we recommend several improvements. Some of these are probably simpler than others to address. First, it should be able to deal better with mismatched tags, particularly since <p> tags aren’t required by WordPress. Second, the ICML conversion should be able to understand special character entities, since they are so common. Third (but probably least important), images should be able to survive the process, even if only as placeholders, since web images are not appropriate for print anyway.

Fourth, the plugin should be a back-end function, not a public front-end one, so that administrators can control access to it. Sixth, there should at least be an option to automatically run the ICML conversion to, trim a step for the user.

Finally, users should have a choice between exporting pages individually or in a batch; whether each page should be processed separately or complied into a single file; and whether the output should be optimized for print and/or e-book. This would allow the system to fit into several work-flow preferences and be as useful as it can be.

Leave a Reply