On Tue, Jul 5, 2011 at 5:54 AM, Russel Winder <rus...@russel.org.uk> wrote:
> On Mon, 2011-07-04 at 11:56 -0600, David Gileadi wrote: > [ . . . ] > > One concern I have is that the D site has a lot of pages, and the > > navigation can be a bit hairy. I'd like to see a mockup of a sub-page, > > a Phobos or Language Spec page for instance, that shows what > > sub-navigation would look like with your design. > > Although my familiarity with the experimental studies is now 12+ years > old, and so arguably a little out of date, the headline from then was > that there is a very complex interaction between "immediate visual > impact", "navigability", and "need to use". If the "immediate visual > impact" of a page is poor then there has to be a very high "need to use" > for "navigability" to be a factor. Similarly if the "navigability" is > poor then "need to use" has to be high for good "immediate visual > impact" to offset the problems. In the middle ground there is a lot of > individual preference. > > Of course where "immediate visual impact" and "navigability" are good > "need to use" falls away as a factor, and indeed the web site enters the > realm of being a positive draw. > > So what is my point? Much of the debate recently has been about > "immediate visual impact" and from what I can tell, none has been about > "navigability". Without there being an easily inferable navigation > model, there are always going to be problems, and often they manifest as > grumbles about "immediate visual impact" whereas in fact the problem is > a lack of "navigability". > > Basically, I think there needs to be a discussion of the navigation > model and navigation structures as much and probably before discussing > the look. Hi Russel, you're right: most of the discussion has been about visual impact, yet navigation is absolutely crucial. By designing a homepage I've given the impression that I just want to design a "showpiece" site. This is wrong and I've been giving thought to navigation, but haven't had the confidence to properly tackle it yet. There are a few things I'm reasonably sure of: - Documentation (language and library reference) is 95% of the challenge. Navigation of the rest of the site, which is either just single top-level pages or lists of pages (e.g. Howtos), pales in comparison. - Documentation should not be squished into the rest of the navigation. I much prefer the approach taken by languages that host docs in a sub-site. This works because I visit their documentation all the time, but the rest of the site only rarely. Pop quiz: how many of you visit http://www.d-programming-language.org/ every day to read the quote from Andrei? Not many I'd bet -- you go there as a waypoint in getting to the docs. - Navigation design should be based heavily on experience. I'd like everyone's input on what online documentation exists that really works, and what aspects of it work. (For instance, I like it when all exported symbols are listed at the top of the page, like this <http://golang.org/pkg/io/> .)