"Count what is countable. Measure what is measureable. What is not measureable, make measureable." -- Galileo

Friday, February 20, 2009

Plone-site Tags

A tip o' the hat to Sam, who left a suggestion during last weekend's Great Backyard Plone Count. He suggested mining Plone URLs from Del.icio.us where plone-site is the requisite tag in question.

A couple lines of Python later and we've got a Google Docs spreadsheet with 1,427 Plone URLs. Del.icio.us reported 2326 tagged items, but about 900 were duplicates. Not sure how they're reporting things; perhaps they're getting subordinate pages with tags folded in as well.

At any rate, there's a whole lotta Plone out there. That also means the GBPC netted 163 sites (183 less the 20 plone-site tagged items I manually entered). Another tip o' the hat goes out to everyone who took the time to enter their "sitings" this weekend. We'll see you next year and then we'll be able to start spotting some trends.

Meanwhile, I've got some homework to do--comparing Plone.net listings with the Del.icio.us set, running some QA/QC, and checking to see if all the sites are live. I might even have a shot at comparing numbers per country against things like Llakomy's stats for Britain.

Also in my things-to-do list is finishing up the tasks that Nate Aune got me to sign up for during this week's Plone Marketing Committee conference call. With the Marketing Committee reinvigorated, expect to see more advocacy and evangelism for Plone.

And finally, I've been inspired to start tracking Twitter tweets that mention Plone. Should be good for the occasional blog posting.

Tuesday, February 17, 2009

The GBPC is Ovah!

The Great Backyard Plone Count is over and I'd like to thank everyone who participated. Just in a brief once-over of the data shows an incredible diversity of sites reporting in. Plone Community, you guys rock!

The data collection spreadsheet is still up and available to all, but I've turned off the online form until next year. In the graph above blue is the cummulative number of sites reported during the 96-hour long weekend and red is the number of people contributing.

One side-effect included Sam inspiring me to data mine Del.icio.us for "plone-site" tags. I've done the basic sweep of all 2326 links, but I've also collected a bunch of Delicious anchors that have to be purged before I publish the full list. Stay tuna'd for that later this week.

Another side-effect was Finn Arild publishing a link to his own list of sites for both Plone (CMS) and Plone (music).

Saturday, February 14, 2009

The Great Backyard Plone Count Continues


Data continues to come in for this weekend's GBPC. In the graph above the blue line is the cummulative count of sites that have been added. The red line is the cummulative number of contributors. Hop over to the online form and add your two-cent's worth.

I've manually added a handful of items from Del.icio.us and by the end of the weekend I hope to have a simple Python script to automate that data mining task. The main problem is that I only will get a site name and URL. There won't be any additional identifying information: site description, developer, owner, or location.

There's still plenty of time to contribute. Thanks in advance.

Thursday, February 12, 2009

Great Backyard Plone Count


What is the GBPC?

The Great Backyard Plone Count is a shameless copy of the Great Backyard Bird Count. Just as GBBC is an effort by the Audubon Society and the Cornell Lab of Ornithology to track birds, GBPC is a fledgling effort to catalog Plone sites worldwide. Its a voluntary, self-reporting project (with all the weaknesses that go along with it) that attempts to find as many Plone sites as possible, including and especially intranet sites behind firewalls.

The Great Backyard Plone Count is a hoped-to-be annual four-day event that engages Plone watchers of all ages in counting Plone sites to create a real-time snapshot of where the sites are across the world. This year the GBPC is February 13-16, 2009, the same dates as the GBBC. Anyone can participate, from beginning Plonistas to experts. It takes as little as 15 minutes on one day, or you can count for as long as you like each day of the event. It’s free, fun, and easy—and it helps Plone. We'll be adding updated 2009 GBPC materials on this blog and on the Plone marketing site as they become available.
Participants count Plone sites anywhere for as little or as long as they wish during the four-day period. They tally the site name, URL, and other statistics. To report their counts, they fill out an online checklist at the Great Backyard Plone Count web site.

As the count progresses, anyone with Internet access can explore what is being reported from their own towns or anywhere in the World. Next year they will also see how one year's numbers compare with those from previous years.

Statistians and Plone enthusiasts can learn a lot by knowing where the sites are. Websites are dynamic; they are constantly in flux. No single observer or team of observers could hope to document the complex distribution and changes of so many sites in such a short time.

We need your help. Make sure the sites you visit from your online community are well represented in the count. It doesn't matter whether you report the 5 sites you routinely visit or the 75 websites you see during a day's browsing and searching specifically for Plone portals.

Your counts can help us answer many questions:
  • How does the number and types of Plone sites compare with past years? (Well, at least next year we can do this!)
  • How are other applications, especially CMSs, affecting Plone sites in different application domains as well as geographical regions?
  • What kinds of differences in Plone diversity are apparent in various domains and usage areas?
  • Are any types of Plone sites undergoing worrisome declines that point to the need for attention?
Plone developers will use the counts, along with observations from other citizen-web projects, such as Plone.net, CMS Matrix, and Ohloh, to give us an immense picture of our Plone environment. Each year that these data are collected makes them more meaningful and allows Plone and Web researchers to investigate far-reaching questions.

Disclaimer: All information collected in the Google Docs spreadsheet will be publicly available. Identifying information is voluntary (but we do need at least a unique "handle") and all fields are optional.

Saturday, February 7, 2009

Usability of CMS Home Sites

Color in the News

An interesting article on color and cognition turned up in Science today and its making its way into the mainstream media. The long and the short of it is that the color blue unconsciously enhances the performance of creative tasks. The color red unconsciously enhances the performance of detail-oriented tasks.

I think the take home lesson in terms of home site usability is that, all other things being equal, one should emphasize blue on those pages whose purpose is to elicit a positive decision about a CMS product. I believe that evaluating software and making a go/no-go decision is a creative task. Detail-oriented tasks (which would involve red pages) are in-depth analysis and feature comparison, more like reading and digesting CMS Matrix results and such.

Usability, Part 3

I'm continuing with my review of CMS and usability, triggered by my elderly parent's difficulties with accessing information on the web. I certainly can't survey hundreds of various CMS implementations and I don't have a handy panel of users to conduct timed usability tests, so I'm looking at something a little less ambitious (and a good bit less rigorous). I have scored each of the parent sites of a few top-flight CMSs against Jakob Neilsen's "Top Ten Mistakes in Web Design."

For the moment, I've limited my study to plone.org, drupal.org, joomla.org, and dotnetnuke.com. The results are displayed in two tables, again due to my CSS not displaying wide tables. Italics are simply to visually separate rows. The snippets of text are to help readers link a particular score with a discussion item in the paragraphs below.

The results have been enlightening and they mirror my earlier heuristic look-and-feel analysis. In the tables below, I've scored a point for the presence of each of Neilsen's top ten mistakes. Half points were given where the flaw was not deemed egregious or of minor impact. I urge readers to go to AlertBox and get the full description of Jakob's top ten mistakes.

Mistake

Plone


Drupal


Bad search



PDF files for online reading



Not changing the color of visited links

1 #0062A0 and #027AC6
Non-scannable text 0.5 No "F pattern" 1 Walls of text everywhere
Fixed font size



Page titles with low search engine visibility 1 "How to" 1 "Join"
Anything that looks like an advertisement

1 True advertisements
Violating design conventions



Opening new browser windows



Not answering users questions 1 "Market-ese" 1 Reverse chronological listing on homepage
Total
2.5
5

Plone. The homepage is mercifully uncluttered, but doesn't display information in the highly scannable "F-shaped" pattern that eye-tracking studies have shown is common when users are looking for something. Page titles in the documentation could be better and its often difficult to find a specific article unless you remember the exact wording. Try to "lock down your site" so people can't "join as users?" You have to search for "Close your site." Plone also has far too many items that start with the phrase "How to...." And finally items in the marketing box at the bottom of the page use catchy phrases that don't necessarily fit with the questions or concepts that a prospective new user might have. None-the-less, Plone.org had only 2.5 of Neilsen's baddies.

Drupal. Meanwhile, the heaviest hitter among CMS has a much weaker home site. Visited links barely change color, large walls of unscannable text are everywhere, and true advertisements are prominent. While it has an obvious simple search box, trying to "lock down your site" so people can't "join as users" results instead in hundreds of items about using "join" in queries. The lengthy reverse chronological listing of news items only serves to obscure important, basic information that first-time site visitors might be looking for. Drupal comes in with a disappointing 5 out of 10.

Mistake
Joomla
DotNetNuke

Bad search 1 No search box

PDF files for online reading



Not changing the color of visited links

1 No change at all
Non-scannable text 0.5 Mini-walls of text 1 Icons scattered; menu bar; "toolbar"
Fixed font size



Page titles with low search engine visibility 1 No "How to" 1 No "How to"
Anything that looks like an advertisement 1 "Contribute" icon 1 True advertisements
Violating design conventions 1 No "Home" 1 Pulldowns interfere with toolbar icons; no "Home"
Opening new browser windows



Not answering users questions 1 "What is?" 1 Questionable relevance; Where is the documentation? "Downloadable files"
Total
5.5

6


Joomla! This was the only home site that lacked an obvious simple search feature. When I finally found a search box one level down, its advanced search was useless to an outsider. How would I know to search a "chunk talk," a "transwiki," or a "JDOC?" Its first page text areas were reasonably deployed, although not in an "F-pattern," but each was a monolithic mini wall of text. The look-and-feel of the documentation area was completely different from other pages. The "Contribute" link looked like a pitch for money, and their was no explicit "Home" link on many pages nor easy to recognize bread crumbs. It took some getting used to the non-standard nav menu with its confusing "Main page" and "Main" links that went to different places. Finding "How to" documents was frustrating. Perhaps worst of all, there are things like the "Getting Started" link, which takes you to a "What is" section where the download link is buried below many column-inches of scrolling text. Joomla's site was beautiful and balanced but inconsistent not very usable with a score of 5.5.

DotNetNuke. This site had a number of Neilsen's sins. Visited links didn't change even slightly, icons were scattered across the page in both a panel of images and in a toolbar-like menu. The text menu bar expanded down and interfered with seeing the icons below and was awkward to clear. Finding "How to" information or any documentation at all was difficult. Why would I look under "Downloadable files" to find documentation? When I did search for items, the returned set was of questionable relevance. My question about "locking down a site" returned results that were absolutely off mark. The presence of true advertisements detracted from the main purpose of the site, unless the main purpose is to generate revenue for sponsors. DotNetNuke came in at the bottom of our four sites with 6 critical mistakes.

Summary. Clearly and not unexpectedly, all four CMSs have some weaknesses. Surprisingly, Plone comes out ahead of the pack while the other three had 5 to 6 of Neilsen's top ten usability mistakes.