[Too Soon Old Too Late Smart Epub
Virginie Fayad
The SMART tool is a free Web-based complement to the Ace accessibility checker. It helps guide evaluators through the manualchecks necessary to ensure conformance with WCAG and the EPUB accessibility specification.
After logging in to SMART, a new evaluation is started similarly to the Ace tool. Instead ofdragging an EPUB file into the application, however, the report.json file output byAce can either be dragged and dropped into the blue box on the page or the link in the box canbe used to select the file to upload.
This dialog indicates how the SMART tool has been optimized for the publication. For example, itlists any tests that have been disabled for media types not found in the publication (e.g.,there is no need to list audio checks when there is no audio).
The Publication Information tab is the default view in the SMART evaluation interface. Itallows general information about the publication, such as its title and authors, to bereviewed and modified. This metadata is only used in the final report that the SMART toolgenerates, however.
One key field to note is the publication format. The SMART tool will automatically set thisfield to EPUB 3 or 2 based on the Ace report, but the value specified here is also used todetermine the type of tagging to generate in later evaluation steps (i.e., for the discovery and result metadata).If the tool outputs the wrong tags, make sure the format has been set correctly.
The Additional Metadata field at the bottom of the form is used to include extra informationin the final report. The data input here does not have to conform to a metadata standard.The field takes a human-readable label separated from a value by a colon and space (e.g.,"Region: North America").
Due to the translations to make the metadata human-readable in the final reports, theSMART tool does not currently provide the option to generate new metadata tags whenmodifications are made. The source publication must be manually updated in suchcases.
The Conformance tab contains all the tests that need to be carried out to verify thepublication is accessible. Each test is a self-contained unit with instructions on how toevaluate for conformance and a field to indicate if it passes or fails.
The instructions provided in the SMART tool aim to cover everything that needs to be checkedto pass a test, but this is possible to provide in all cases. Some WCAG criteria requiredetailed checking of the markup. For these tests, a list of the most common problems isincluded to look for.
In addition to the guidance provided in the tool, links to additional documentation in thisknowledge base, in WCAG supporting documents, and in the EPUB accessibility techniques arealso provided. These links are initially hidden but can be viewed by expanding the relevantheading.
The SMART tool will automatically set tests to N/A if it determines from the Ace report thatthey are not likely to apply (e.g., all image, audio and video tests will be set to notapplicable if there is only text content).
The Discovery tab is where the accessibility metadata for the publication is displayed. Thetab contains input fields for each of the properties required or recommended by the EPUB Accessibilityspecification.
The SMART tool will auto-populate these fields during the loading of a new evaluation if itdetermines the values are likely to apply. Make sure to review the values it sets, however,as evaluating the publication may change some fields (e.g., if all images are determined tobe presentational, any image-related metadata set would likely no longer apply).
If accessibility metadata is added or modified, the SMART tool can create a new set of tagsto use in the publication's package document. Click the Generate button at the bottom of thepage to open a dialog with the new markup.
The distribution metadata fields are like the discovery fields, although there are somedifferences between the two standards. When metadata set in the Discovery tab has anequivalent ONIX property, that field will also automatically be set in the Distribution tab.For example, the summary input into the Discovery tab is automatically copied to ONIX00 field.
Like the discovery metadata, it is also possible to output the distribution metadata in itsnative format by clicking the Generate button at the bottom of the tab. The ONIX tags arerendered in a new dialog window.
The result tab indicates whether the publication successfully meets the requirements of theEPUB Accessibility specification. This field is not editable, as it is automatically updatedas an evaluation is carried out.
If any tests are not complete, or any required metadata missing, the result will be"Incomplete". See the section on validating evaluations formore information on how to locate the incomplete sections.
The final two fields are optional and only need to be completed if a final report is beinggenerated. The first, the evaluator, is used to declare the name of the organization orindividual who carried out the evaluation. The second, the link to report, is for the URLwhere the final report will be hosted, if it will be made publicly available.
If any notes have been included in the evaluation the radio buttons at the top of the tab allowcontrol over which are output. By default, all notes are included in the report, but the outputcan be limited to only failure descriptions, only notes, or no notes of any kind.
If there are no issues that need fixing in the preview, the next step is to generate adownloadable version of the report. Click the Create button and a browser prompt to save thereport file will appear (some browsers may automatically download the file without promptingdepending on the current settings).
As it is not always possible to complete an evaluation in one sitting, the SMART tool providesthe option to save and resume an evaluation later. Clicking on the Save button in the upper-leftcorner of the application opens a dialog asking where to save the evaluation.
Evaluations can be saved as often as desired but must be saved at least once toresume later. The SMART tool does not save evaluations by default, so if the data is not saved,the evaluation cannot be reloaded from the history.
Each row in the evaluation history table represents one saved evaluation. There are columns forthe name of the publication, when the evaluation was started, when it was last saved, itscurrent status, as well as a set of buttons for maintaining the evaluation.
Evaluations that have been saved locally can also be resumed by dragging and dropping the JSONevaluation file (not the Ace report) into the drag-and drop area used to start an evaluation.The SMART application can differentiate the data files and will automatically load theevaluation if it matches a record in the history.
If an evaluation was stored on the SMART server, deleting the file permanently deletes all recordof it. If an evaluation was stored on the local file system, the data file will also need to bedeleted.
If an evaluation is accidentally deleted and the data is saved locally, the evaluation can stillbe retrieved. Start the evaluation over using the Ace report file and then save the data locallyagain, but do not overwrite the existing saved data. The original evaluation data can now bereloaded from the evaluation history.
The latter problem poses more of a challenge. The SMART tool tries to mitigate this issue byidentifying what to look for and providing reports that explain what the evaluator found.Regardless, it is possible for two evaluators to assess a publication differently.
Therefore, while the SMART tool is intended to be used by anyone to improve the accessibility oftheir publications, it should not be used to certify content as accessible except by personsfamiliar with WCAG and evaluating its success criteria.
This. Don't assume it's ID and the export causing the problem, when it could be anything from the basic nature of the PNG conversion/.resolution, or other factors including how the reader choses to process and display it.
Export the figure yourself to an acceptable 300ppi image, and import that. All of the "auto conversion" steps that the Adobe suite will do among its components are a useful luxury, but not necessarily best practice in every case. EPUB tends to be well outside "most cases."
You don't say anything about the specs from AI to PS to ID, but I'd cut through the steps and simply export from AI to either PDF or export to JPG at 300dpi. Then use PNG as your EPUB export mode if you like. I've generally had better luck with JPEG as source images in EPUB than PNG.
An image of vector elements, rasterized and resampled at least twice in a process, is going to lose some sharpness of edges and text. But try PDF to see if that preserves enough detail in the final conversion; if not, an optimal JPEG with good sharpness should come through acceptably in the reader.
Try another EPUB reader, such as Thorium or Calibre. If the image issue persists there, it might be fixable. If it's only on the Apple reader, it may be a quirk of how that (nonstandard) reader is interpreting the image.
We create a new document meant to be published to regular iPads as the highest quality digital platform. We can still use the produced epub on lower resolution devices, of course (but you have to realize that images may have double the resolution actually required for those displays).
When we import an image and place it we should check the effective PPI in the Links panel. If the placed image (after transforming to the required width, height, and cropping) is much lower than 72PPI we have a problem. Ideally it should be 72PPI or higher!
It is actually not necessary to go through these steps. While ideally we provide InDesign with a finely crafted bitmap, as long as the effective PPI of the placed bitmap is equal to or higher than an effective PPI of 72, we can always export the epub with an image resolution setting set to 72 PPI in the Conversion Settings.
Here I imported a PNG bitmap in an iPad document. The original PNG's resolution is 2008px by 1782px. Too high for the iPad Retina in portrait orientation, and that does not matter! The effective PPI is 99, and that is more than 72PPI. So when we export to an epub, we simply ensure that the conversion is set to PNG and 72PPI. This will convert our imported bitmap to and EXACT 1:1 version at the native resolution of the iPad retina (because it is placed aligned to the default margins in portrait mode, the final pixel resolution of the width will be 1536 - 36px - 36px = 1464px. The height is inconsequential in a reflowable Epub (up to an extent: depending it will be displayed scaled down to fit...)
795a8134c1