Re: Flickr and alt

2008/8/19 Henri Sivonen <hsivonen@iki.fi>

> The notion that a syntax specification should require software conforming
> to the specification to produce syntactically non-
> conforming output under some circumstances is patently bizarre.
>

I don't agree, and here's an example that illustrates why I don't.

I use the Oxygen XML editor to edit XHTML documents. By means of a coloured
box in the corner of the editing pane, which is green if the document is
valid and red otherwise, it can tell me at a glance whether or not the
document I am editing is valid XHTML.

This feedback also provides me, as an author, with a useful reminder that
validity is important (if it wasn't, why bother to check it?); it serves as
a moral reminder of my duty to the web to produce valid mark up. It can't
check everything - it can't check whether my alt text is helpful or not -
but it reminds me of my responsibility to do so, and it goes as far as it
can, by - for instance - checking whether I even bothered to insert alt
attributes at all.

Something Oxygen does not do, and, I believe, should not do, is prevent me
from publishing my mark up if it is invalid. I'm happy for it to notify me
that my mark up isn't valid (if it isn't), and to remind me that it should
normally be valid. But if I do want to output invalid mark up for some
reason (for instance, because I'm in the middle of a work in progress, or
because I want to publish a demonstration of what invalid mark up looks like
and why not to use it), then I really should be able to.

In summary, it may be desirable under some circumstances for authoring tools
to output invalid mark up, but they should warn the user before doing so,
and should go some way toward explaining why the author should think twice
before publishing such mark up.

Sam