I agree with Simon in that where the image usage is not under our control (like t:graphicImage), we should not output an alt tag unless the end-user specifies an alt tag.
On Mon, Jun 30, 2008 at 4:50 AM, [EMAIL PROTECTED] <[EMAIL PROTECTED]> wrote: > On Mon, Jun 30, 2008 at 12:15 AM, Hazem Saleh <[EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]>> wrote: >> >> Hi Team, >> >> Simon and me made a discussion about making the (t:graphicImage) >> component XHTML complaint. >> >> Here is the thread discussion : >> https://issues.apache.org/jira/browse/TOMAHAWK-1143 >> >> We need to take your opinion about that, >> Have we have to make the components XHTML complaint or leave this >> to the user's usage with warnings ? >> > Hazem Saleh schrieb: >> >> Sorry the thread discussion is here : >> https://issues.apache.org/jira/browse/TOMAHAWK-1291 > > Manfred commented on the jira issue: > [I moved this to the email thread, so we don't have half the discussion here > and half on the issue] > >> +1 for a strict (but sweet-tempered) behaviour > >> that means: >> - log a nag warning >> - render a non-empty alt attribute with a "meaningful" default text if the >> developer >> omits the attribute (or provides an empty one) > > The thing is that for h:graphicImage and t:graphicImage we have **no idea** > what a "meaningful" text would be. This is some arbitrary image that the > user has chosen. For what purpose? We don't know - unless we embed AI > software and do image recognition on the referenced file. So for > h:graphicImage and t:graphicImage we have **only** these choices: > (a) don't output ALT. This screws all blind users, but in an obvious way so > that QA departments can easily detect it and tell their developers to add > the needed alt attributes. And it is not our code that is at fault. > (b) output empty ALT. This screws all blind users, but it cannot be detected > by validation. And it is our code that is at fault as well as the user code. > (c) output ALT with "ha ha no description". See (b). > > For cases where myfaces components are generating the image references for > their own purposes, they *know* what that purpose is. Always. So they are > always capable of attaching a valid ALT description. > > > Regards, > Simon > >