The authors are not responsible for the content of any external sites linked to from webdesignbook.org
©2006–2012 MacAvon Media.
Answers to Exercises, Chapter 11
Buy Chapter 11, Web Site Design, in PDF format: now available from the MacAvon Media Downloads Store.
- The fundamental reason why Web sites rarely exhibit a purely hierarchical structure is that the structure of the site models the structure of the information and services that it presents, which in turn model some aspect of the real world, and the real world is not built neatly out of pure hierarchies. The untidiness of the world manifests itself in site structures in at least four ways. First, some parts of the site will impose a linear ordering (for example, a sign-up process), which introduces sequences within an overall hierarchy. Second, some pages will belong in more than one place, so a site structure that reflects this will be composed of superposed hierarchies, with more than one route to some pages. Third, in an absolutely pure hierarchy, all the information is in the leaves (pages with no outgoing links); pages higher in the structure only contain links to their descendants. It is usually more efficient (in terms of the number of clicks a visitor must make to reach some information, or the number of pages that must be fetched from the servers) to move information up in the hierarchy, so that most pages consist of a mixture of links and information. Finally, a pure hierarchy only has links up and down the tree structure, but most sites provide cross-links (often in the navbar) for immediate access to important pages.
- Any collection of pages where each page (except the first and last) has both next and previous links forms a sequence that can be traversed in either direction. Familiar examples include the pages of results from a search engine, pages of responses to a post in an online forum, product listings on a shopping site where there are too many products in one category to fit on a single page or an image gallery arranged as a slide show. Sequences that can only be traversed forwards are typically processes where each step depends on the preceding steps. Examples of such sequences include registration procedures, checking out from an online shop and filling in multi-page online surveys.
- The problem with calling the home page
homepage.htmlis that Web servers and clients make certain default assumptions about the naming of home pages, which enable them to support certain convenient shorthands for users. In particular, if the name of the home page is
index.php, then, as we mentioned in Chapter 2, it can be omitted from HTTP
GETrequests, and a site can be accessed simply by typing its domain name. If the home page of the Desperate Software site was called
homepage.html, then a user who simply typed
desperatesw.co.ukinto their browser would get a 404 error, and probably assume that the site did not exist. However, most (if not all) Web servers can be configured to look for a different set of file names when none is given, so you could call your home page
homepage.htmland avoid any problems if you had access to the server's configuration files and could tell it to look for
homepage.htmlas well as the default names when a request did not specify a particular file.
- A link checker that works by scanning the source of the site's pages and following the
aelements would not be able to determine whether every site could be reached from the home page, because some pages might only be reachable via server-side scripts that required data. For instance, the home page of the site might be a login page, where the user had to enter their name and password before being transferred to the inner pages of the site. All those inner pages would appear to be unreachable. It is possible to imagine a link checker that had its own name and password for this case, but it appears to be the case that checkers that follow dynamic links would have to be built individually for each site.
- A hierarchy is a simple and elegant structure: each node has exactly one parent, there is exactly one path from the root to every node. It follows that to get from one node to another it is necessary to go up the hierarchy and back down again. Users don't like having to click repeatedly to move from one page to another, and fetching the pages that correspond to the intermediate steps en route is inefficient and a waste of bandwidth. Hence, it is customary to provide direct links from any page to other popular or important destinations in the hierarchy – often these links comprise the navbar. Doing so has the considerable additional benefit of providing navigational information that helps users orient themselves. Imagine if, on every page of a site with as many pages as this one, the only links you could see were to its parent and children. How could you know how to get from this page to the feedback form? You would have to work your way up the home page first. Would you be likely to do so?
- No, a site map should not necessarily contain a link to every page on the site. If the site includes sequences that must be traversed in order, it would be wrong to include a link to the intermediate pages of such a sequence in the site map, since it would provide a way of jumping into the middle of a sequence, such as a payment procedure, without having gone through the prerequisite steps. Where there are such sequences, the site map should only include links to their entry points.
Discussion Topics: Hints and Tips
- This question is by no means a purely hypothetical one. Many Web applications and Web application frameworks work in precisely the way described. It is instructive to look at the form of URLs shown in the address bar for complex dynamic sites and see which parts change and which don't. If you have sufficient technical knowledge, you can get an idea of what might be going on and the advantages to implementors by looking at how Ruby on Rails or some other Web application framework rewrites URLs. To see the disdvantages of this approach think about how people use URLs. If you needed to make a note of a site's URL, would you like to have to write down the sort of raw URL that relies on elaborate query strings to choose a page?
- What benefits do Web sites offer small businesses? Are there any aspects of a plumber's business that distinguishes it from other types of business, such as Web design agencies or hotels? In particular, how do plumbers get work? When someone in your part of the world needs a plumber, are they likely or able to look on the Internet?
- This is not a technical question, but is concerned with organization and management. Adding specialists to a team brings skills that other members might not have, but introduces new problems of communication. Information architects have even been called a curse. What sort of team structures can cope with these problems? For an unconventional but popular view of this area, take a look at Getting Real, but don't take it all seriously.
- Look at a lot of home pages and see whether there is anything about their content and purpose that would require or benefit from a separate design. Does the fact that many visitors to a site actually arrive at an inside page by following a link from a search engine affect whether the home page needs its own design?
Practical Tasks: Hints and Tips
- This is, admittedly, a tedious job, but it can be illuminating. To get the most out of it, visit as many sites as you can and tabulate the number of occurrences of each navbar label. (You could perhaps use a spreadsheet or write a little script to collate the data.) You should end up with an idea of which sections appear commonly and how they are labelled. If you come across unusual labels for which there is a more common alternative (for example, if you find Q&A instead of FAQs) ask yourself whether the less commonly used label is more accurate and/or confusing.
- It's probably not a good idea to try this with a huge corporate site. You could try analyzing this site and compare the way it finally came out with the initial design ideas in the book.
- This is quite ambitious and could be a group project, bringing together different specialists. Think carefully about the structure of the site before you start.
- Sites like this are still a large part of the work of small agencies. Although they may seem dull compared to applications like the previous one, they are valued by the clients and keep Web designers fed. You have to take the job seriously and pay attention to details, otherwise the client won't see why they couldn't get their teenage son to do the site using Microsoft Frontpage. This project will be much more valuable if it is possible to find a real client, preferably one who says things like "That's really great, there's just one other thing...".