Re: Is Flickr an Edge Case? (was Re: HTML Action Item 54)

On May 23, 2008, at 6:00 PM, Karl Groves wrote:

> ----- "Maciej Stachowiak" <> wrote:
>> On May 23, 2008, at 3:56 PM, Matt Morgan-May wrote:
>>> On 5/22/08 10:33 AM, "Andrew Sidwell" <>
>> wrote:
>>>>> Well then, making @alt optional in the edge case of Flickr or an
>>>>> inkblot
>>>>> test is then moot.  Those edge cases will remain non-conformant,
>>>>> and @alt as
>>>>> a mandatory requirement is a sealed deal, as optimising for edge
>>>>> cases is
>>>>> not a reasonable thing to do.
>>>> Flickr is hardly an edge case.
>>> On the contrary: Flickr is the _ultimate_ edge case.
>> Flickr is an extremely popular site with millions of web pages. Look
>> at the Alexa Global Top 500:
>> I see the following obvious photo sharing sites:
>> #26 Photobucket
>> #34 ImageShack
>> #39 Flickr
>> #57 Fotolog
>> #59 ImageVenue
>> #84 Metroflog
>> #100
>> And the following social networks with photo galler features:
>> #6 Myspace
>> #8 FaceBook
>> #11 Orkut
>> #40 Friendster
>> #56 LiveJournal
>> Clearly photo sharing is one of the most popular activities on the
>> web. It is not an edge case but rather a major use case that should  
>> be
>> given significant consideration.
>> Regards,
>> Maciej
> According to a NetCraft Survey, there are 165,719,150 sites out  
> there [1].  Your list of major photo-sharing websites is hardly a  
> drop in the WorldWide bucket, as it were.

Word Wide Web traffic is widely believed to follow a power law  
distribution (looking at the "reach" statistics for the top 100 sites  
is consistent with this hypothesis). This means that the top few sites  
likely get as much traffic as the remaining 165 million combined. Keep  
in mind also that flickr is a site with billions of pages. It is clear  
that by any measure, photo sharing is one of the most popular  
activities on the Web. Only search, video sharing, social networking  
(which very often includes an aspect of photo sharing), blogging (also  
sometimes including an aspect of photosharing) and online shopping  
appear to be more or equally popular, as far as one can tell from the  
most popular sites and their reach. If one of the most popular  
activities on the internet is an "edge case", then what would you  
consider to be a valid use case?

> Using the same Alexa data you cite, Flickr has a 3 mo. average reach  
> of 1.706% of web users[2].
> While 1.706% is certainly very respectable, given the context,  it  
> *is* still only 1.706%.

Given the hundreds of millions of global internet users, that is a  
huge number. If you add up the reach of all the sites I listed, you  
will get an even bigger number.

> This whole conversation about Flickr is a red herring anyway and all  
> of these back & forth messages about it are detracting from the real  
> issue: Failure to provide alternate text is a violation of every  
> accessibility standard in existence.

No, that's not the real issue. Accessibility standards are a means to  
an end, not an end in itself. The real issue is: what HTML document  
conformance requirements are likely to lead to the best accessibility  
outcomes, considering all the use cases and possible unintended  
consequences and second-order effects? Our goal should be to improve  
accessibility, and accessibility standards are only valuable to the  
degree that they help achieve this goal.

> If this WG intends to work out a way to mark an image in a way that  
> says, in essence, "alternate text isn't available or wouldn't be  
> useful", then by all means continue the debate in that context.

Several ways to mark in image in such a manner have been proposed. The  
current spec draft proposes:

<img src="photo.jpg">

Other alternatives that have been proposed:

<img src="photo.jpg" noalt>
<img src="photo.jpg" importantimage="importantimage">
<img src="photo.jpg" alt="">
<img src="photo.jpg" alt="*">
<img src="photo.jpg" alt="_">
<img src="photo.jpg" alt=" ">

> Red herring arguments about how popular a particular site is really  
> isn't very productive and certainly hasn't lead to any useful  
> solutions to the problem.

They are relevant to whether the use case of embedding images where  
alternative content is not available is important to consider. If you  
agree that it is, then I agree there is no need to debate the matter.


Received on Saturday, 24 May 2008 08:53:29 UTC