Wcag: Understanding doc contradicts SC2.4.5 Multiple ways

[guyhickling]

The second paragraph of the Understanding document says "For a three or four page site, with all pages linked from the home page, it may be sufficient simply to provide links from and to the home page where the links on the home page can also serve as a site map." The way I read this is that in this case each page will have only one way to locate it, which directly contradicts the SC, unless I'm reading the paragraph wrong. What does everybody think?

And yes, I agree this might seem a trivial case to raise, but it sets a precedent for ignoring the SC. Suppose you have a 20 page site, but it is the same simple structure as the example in the Understanding, i.e. all 20 pages can be reached from a link somewhere on the Home page and all have a link back to the Home page.

While seemingly trivial, this question takes on more meaning when you consider the tablet and mobile phone responsive layouts. Even a 4 option menu can become a drop down hamburger menu in the phone layout, and we all know how prone they are to not work for all kinds of users. (And the reason I am raising this query is I just had a case of exactly that. It was 4 pages and, yes, the phone version and 400% zoom layouts did not work for keyboard users. If there had been a second menu in the page footer that wrapped lines instead of turning into a drop down menu on the zoomed layout, the site would still have been usable by those people. As it was, it wasn't.) I agree the prime remedy in that case is to fix that failure, but it does highlight that this issue of multiple ways for small sites is not as trivial as it first looks.

So, should there be this apparent contradiction in the Understanding document?

[patrickhlauke]

The way I read this is that in this case each page will have only one way to locate it, which directly contradicts the SC, unless I'm reading the paragraph wrong. What does everybody think?

I'd agree that that particular paragraph does contradict the SC.

But it brings me to a different/related question: if the paragraph instead said "and on each page there is always link to any other page" (i.e. there's a standard navigation, and every page can be reached from any other page), does that pass? Because if you have N page, there are N-1 different ways to reach that page (from any of the other pages).

The way I've interpreted (loosely) the SC is to mainly be "user doesn't have to remember 'oh, i can only reach THAT page if i'm on THAT OTHER page and there's a link there to it". that basically they can always work out how to reach a page easily, and they don't have to start remembering complex paths/journeys that are the only way to get to a particular page they had in mind.

[detlevhfischer]

I think the inaccessibility of the menu button in your example is a different issue covered by 2.1.1 (possibly also 1.3.1 / 4.1.2) and should not enter the discussion of 2.4.5.
I personally think the exception for small sites makes sense. Even if someone links 20 pages directly from the start page and you can go back there and then go to another one of those 20 pages, is that really an issue? It may be unusual or awkward but the simplicity of the structure still seems to suggest you would not need multiple ways to get there?

[patrickhlauke]

Dredging this up again, but as I've just had a discussion on this topic...this is one of those SCs that seems slightly "handwavy" in one sense, but then also oddly restrictive.

IF we assume that the example with the hairdresser site only passes because it has the following two ways to navigate:

• sequential links from one page to the other (I mean, which site does that, these days?)
• homepage with links to all pages

then are we saying that if every page has a complete navigation on each page, it still must have at least either a search or a sitemap? those seem to be the only options possible. and if so, wouldn't the vast majority of the web really be failing this? i rarely find sites with a complete sitemap (that covers absolutely every page), and while it's easy to say that a site can just add a search function (and one that covers pages, not just products or similar), this is not always easy/possible to do on a technical level.

You could even go further and start dissecting what exactly a "set of web pages" is. Is a whole site, by virtue of being a single site (with same/similar look and feel, navigation, footer) a "set of web pages" necessarily? Or is a set more a self-contained section within that site (like an article split over two/three pages)?

Lastly, can this criterion even be applied to something that is app-like? i'd argue that in most cases, it can't. you couldn't have a "map" that contains links/buttons to all possible screens within the app, nor could you create a search of sorts that surfaces all possible screens of an app.

In short, I admit that this is one of those SCs that is very difficult to judge. Either you take it very loosely as meaning "there must be ways for users to reach a page, rather than having to remember exactly one specific path they need to go down/pages they need to traverse in order to reach this page", in which case most times you'd pass. Or you take it very restrictively as essentially mandating a search function (that covers all pages, including static pages, not just a product search or similar on an ecommerce site) or a complete sitemap, which is a non-inconsiderable amount of work for a small-to-medium site.

[patrickhlauke]

while on the topic of 2.4.5 ... the advisory technique to use <link rel="..."> should probably be retired now too. no browser exposes this in any meaningful way, to my knowledge (yes, old Opera used to...10 years ago or so). not aware of any common/readily available extensions for browsers that leverage these links either.

[mraccess77]

Basically most just treat this SC that there are 2 ways to reach the page without specific pathing. That is there are two other pages that will link to the page - I also agree with you that how you define a set of of pages can make a difference.