Various methods can be used to encapsulate graphics or other page elements on a page that are transparent to the user. If design includes the use of frames, then provision should be made for the user community to choose a no-frame implementation of the same content. This should be considered in the maintenance plan as well. Frames shall not be used to mislead the user about the source, ownership or other aspects of frame contents. Frame presentation of third-party content shall only be done when full consideration is given to the copyright, presentation, appropriate commercial use, permissions and other legal and ethical aspects of such encapsulation.
Links can be expected by the user to lead to other sites and as such do not require these same ethical considerations (see clause 7.7). The _blank target, or other means of creating new windows, shall not interfere with the user’s ability to return to their page history (this relates to requirement 22(o) of 36 CFR 1194, commonly called Section 508. Portions of 36 CFR 1194 are provided in Annex I).
All graphic elements shall contain declared height/width display size, permitting the immediate allocation of page layout for these and concurrent rendering. The use of consistent style sheets can reduce page size, and provide for reuse of style for subsequent pages. Reuse of images, as opposed to use of new images, can reduce download time by taking advantage of local caching. Multiple graphic images at the server should be considered, providing for lower bandwidth connections, and/or user choice.
A potential convention is to have a “thumbnail” graphic delivered, which is also a link to a higher resolution graphic as an option for the user community. Where a server may deliver images in multiple formats, image URIs should not include a specific format name structure (e.g., xxx.gif). To allow for content negotiation with users and to minimize overhead in response, a diverse set of image formats should be provided. Images should not be used to bypass HTML limitations or provide “style” control.
Where available, CSS should be used. Images shall not be used to present text in an alternative style. This is disruptive to text-only browsers, it limits accessibility and global applicability, and it has a negative impact on performance. Graphic presentation of written materials for certain languages, cultures, or disciplines may be necessary. Sites should support image formats for JPEG, PNG, and GIF for compatibility, and seek to deliver the least overhead image acceptable to the client. For animated images, Network Motion Graphics (NMG) should be supported, and scripting or client-side executable languages may be more efficient means of providing the required functionality. The alt attribute is required by 4.2.7 to facilitate access by persons who are not displaying graphics with their browsers. This also facilitates indexing.
Alt attribute descriptions should start with unique information, for example, “home button” rather than “button for home page,” and use functional descriptions where applicable. Longdesc can be used to provide detailed information about graphical content where it is warranted. To facilitate access by older browsers that do not support longdesc, also provide an anchor link to that same data (longdesc takes a URI as it’s value). Unfortunately, firewalls and gateways can convert data types. Hence, the client may not receive the expected graphic.