[tor-bugs] #21942 [User Experience/Website]: Sitemapping the current site layout

[Tor Bug Tracker & Wiki]

#21942: Sitemapping the current site layout
-------------------------------------+------------------------
 Reporter:  linda                    |          Owner:  linda
     Type:  task                     |         Status:  closed
 Priority:  Medium                   |      Milestone:
Component:  User Experience/Website  |        Version:
 Severity:  Normal                   |     Resolution:  fixed
 Keywords:                           |  Actual Points:
Parent ID:  #21222                   |         Points:
 Reviewer:                           |        Sponsor:
-------------------------------------+------------------------

Old description:

> = Objective =
> * generate a digraph representation of the current sitemap of
> www.torproject.org.
>
> We are doing this to visualize how the website is currently laid out,
> analyze the pros and cons of the layout, and to see user paths throughout
> the website.
>
> = Definitions =
> * a ''sitemap'' is a list of pages of a web site accessible to crawlers
> or users, typically organized in hierarchical fashion. It can be a
> document in any form used as a planning tool for Web design.
> * a ''directed graph (or digraph)'' is a graph with a set of vertices
> connected by edges, where the edges have a direction associated with
> them.
>
> = Methodology =
>
> To generate the sitemap digraph, one person (linda) manually visited the
> website and manually wrote the code for generating the visualization. The
> manual crawl began by visiting all the pages reachable with one click
> from the front page, then visiting all the pages reachable with two
> clicks from the front page, and so on. This continued until there were no
> additional pages to visit.
>
> External pages (any site that wasn't www.torproject.org/stuff, so
> donations.torproject.org would be considered an external link) and
> duplicate paths (if one page was reachable from the header, and also from
> the footer, for instance) were noted along the way.
>
> There was existing work done to sitemap the website (#10591), and this
> was taken into consideration. The previous work was used to check that
> there were not any sites that were not accounted for, but since the
> digraphs were not generated in the same way (the old method did add nodes
> for external links, whereas this digraph does, for instance), they do not
> look identical.
>
> = Results =
> * a digraph sitemap of www.torproject.org
>
> [[Image(tpo-digraph-before.png, 600px)]]
>
> * observations about site structure along the way:
>  - 38 links on the front page, which leads to 30+ pages--that's a bit too
> much.
>  - 3-4 ways to get to one page (header, footer, inline, from a subpage),
> sometimes with different text ('volunteer' and 'get involved' lead to the
> same page).
>  - there are site headers, subheaders, AND side headers, which compete
> for attention.
>  - the header, subheader, and footer stay the same throughout the site,
> and are visible everywhere.
>  - the side headers sometimes appear, and also are different depending on
> which page you are on
> (https://www.torproject.org/docs/documentation.html.en vs
> https://www.torproject.org/about/overview.html.en).
>
> * ideas about structure for the new www.torproject.org
>  - keep it simple: nothing more than sub-sub-sub pages (3 clicks away)
>  - less is more: reduce the different amount of content on each page, and
> expand on the select few topics that remain on each page.
>  - consistency: use the same phrasing to refer to the same pages and
> topics throughout.
>  - minimize: use only a header and footer.
>  - put things in the footer that are not as important, and link to a leaf
> page. The current footer links to pages linked to by the header, which is
> kind of confusing.
>  - organize by target audience: a lot of the existing content can be
> organized into the developer, support, and outreach portals. (for
> instance, the manuals, project pages, wiki, can all be in the developer
> portal.)

New description:

 = Objective =
 * generate a digraph representation of the current sitemap of
 www.torproject.org.

 We are doing this to visualize how the website is currently laid out,
 analyze the pros and cons of the layout, and to see user paths throughout
 the website.

 = Definitions =
 * a ''sitemap'' is a list of pages of a web site accessible to crawlers or
 users, typically organized in hierarchical fashion. It can be a document
 in any form used as a planning tool for Web design.
 * a ''directed graph (or digraph)'' is a graph with a set of vertices
 connected by edges, where the edges have a direction associated with them.

 = Methodology =

 To generate the sitemap digraph, one person (linda) manually visited the
 website and manually wrote the code for generating the visualization. The
 manual crawl began by visiting all the pages reachable with one click from
 the front page, then visiting all the pages reachable with two clicks from
 the front page, and so on. This continued until there were no additional
 pages to visit.

 External pages (any site that wasn't www.torproject.org/stuff, so
 donations.torproject.org would be considered an external link) and
 duplicate paths (if one page was reachable from the header, and also from
 the footer, for instance) were noted along the way.

 There was existing work done to sitemap the website (#10591), and this was
 taken into consideration. The previous work was used to check that there
 were not any sites that were not accounted for, but since the digraphs
 were not generated in the same way (the old method did add nodes for
 external links, whereas this digraph does, for instance), they do not look
 identical.

 = Results and observations=
 * a digraph sitemap of www.torproject.org (key: black = webpage, grey =
 external webpage, pink = duplicate link to a webpage).

 [[Image(tpo-digraph-before.png, 600px)]]

 The three main observations about the structure were that it was
 abnormally structured, too flat, and messily interlinked. More details
 about this below:

 1. The current structure of the website does not follow any of the
 standard design patterns:

 [[Image(an-example-hierarchy.png, 400px)]]
 An example of a hierarchy pattern, additional ones are
 [http://adellefrank.com/blog/review-information-architecture-patternsh
 here].

 Currently, the website structure is asymmetrical, and of various depths.
 This can be irritating to users where some pages just "end" whereas other
 pages require 2-3 clicks to find the information that they need.

 2. the website structure is very flat.

 [[Image(flat-vs-deep.png, 600px)]]
 This conveys the same amount of pages, but in a flat vs deep hierarchy.

 Content is more discoverable when it's not buried under multiple
 intervening layers. Users can become overwhelmed with cluttered menus.
 Hierarchies can be helpful if categories are specific and do not overlap,
 which I do think is the case for many of the content in torproject.org.

 3. there is a lot of inter-linking and duplicate links to various pages.

 You can get to to a pages' subpage from another page's subpage. There are
 links with different text ("learn more" and "about tor" both lead to the
 same place) that lead to the same place. On one page, there are multiple
 ways to get to the same page (you can get to the donate page from the
 header, subheader, and footer, and occasionally a side bar tip). All of
 these things are confusing, and we should find out where the best
 placement for something is, and keep it there.

--

Comment (by linda):

 I'm done with sitemapping. Here are some things I didn't want to forget
 about:

 * observations about site structure along the way:
  - 38 links on the front page, which leads to 30+ pages--that's a bit too
 much.
  - 3-4 ways to get to one page (header, footer, inline, from a subpage),
 sometimes with different text ('volunteer' and 'get involved' lead to the
 same page).
  - there are site headers, subheaders, AND side headers, which compete for
 attention.
  - the header, subheader, and footer stay the same throughout the site,
 and are visible everywhere.
  - the side headers sometimes appear, and also are different depending on
 which page you are on
 (https://www.torproject.org/docs/documentation.html.en vs
 https://www.torproject.org/about/overview.html.en).

 * ideas about structure for the new www.torproject.org
  - keep it simple: nothing more than sub-sub-sub pages (3 clicks away)
  - less is more: reduce the different amount of content on each page, and
 expand on the select few topics that remain on each page.
  - consistency: use the same phrasing to refer to the same pages and
 topics throughout.
  - minimize: use only a header and footer.
  - put things in the footer that are not as important, and link to a leaf
 page. The current footer links to pages linked to by the header, which is
 kind of confusing.
  - organize by target audience: a lot of the existing content can be
 organized into the developer, support, and outreach portals. (for
 instance, the manuals, project pages, wiki, can all be in the developer
 portal.)

 The above observations and ideas are just my ideas, and have not been
 decided on as final.

--
Ticket URL: <https://trac.torproject.org/projects/tor/ticket/21942#comment:5>
Tor Bug Tracker & Wiki <https://trac.torproject.org/>
The Tor Project: anonymity online