DokuWiki

It's better when it's simple

User Tools

Site Tools


Sidebar

Translations of this page?:

Learn about DokuWiki

Advanced Use

Corporate Use

Our Community


Follow us on Facebook, Twitter and other social networks.

Our Privacy Policy

devel:plugin_survey:friendliness

Plugin Survey 2011 - Friendliness

(This is a part of the plugin survey 2011)

The plugin system was created to make it easy to share and reuse extensions and modification made by the DokuWiki community. Counted from the first DokuWiki release in august 2004 there has been an average of 2.6 new plugins every week! Today there is a wealth of plugins of different kinds.

Plugin repository

How do I find my favourite plugin? It is not an easy task to navigate and choose among 825 plugins. Powered by repository plugin, a plugin list is maintained. Unfortunately not every plugin is a plugin.

Tags

One of the main search mechanisms for plugins are a tag cloud in the plugin list. Each plugin may have one or more tags attached to the plugin homepage. Two years ago there were 745 different tags. Thanks to a tag cleanup campain, 360 tags were found last year. During this survey 503 different tags where found, 59% of these were only used by a single plugin each. media continues to be the most popular tag with 105 plugins, that is more than 12%. Here are the current top ten:

Tag Plugins
media 105
images 82
embed 78
listing 56
links 53
users 40
search 38
editing 34
navigation 33
diagram 33

The !experimental and formatting tag are no longer on top 10 list.

Excluded plugins

There are also some plugins that are not shown in the plugin list. 12 plugins (2%) are excluded from showing in plugins list for security reasons. All these have problems with XSS vulnerability allowing arbitrary JavaScript insertion. Authors are informed.

Gladly we can see that 2 plugins have had their security problems fixed since last year (dir & select). There are also a new securitywarning field available in the repository to inform users of inherent security risks with some plugins. For example it is not advisable to have a plugin exposing server folders in a public wiki, but might be handy in a more private installation. So far only 4 plugins use this feature.

Homepage appearance

New users may become intimidated by large pages with a lot of confusing comments, plugins marked “experimental” or not stating explicit compatibility.

Homepage size

There are at least 121 plugin homepages that could be considered thin, most contain only a “Details & Download” link. The following 23 plugins represent the other extreme having the largest pages with lots of <code> blocks and comments. They are in a desperate need of cleanup.

Readability

Readability is the ease in which text can be read and understood. Dave Child's text statistics PHP class was used to analyze the plugin homepages. The code blocks were removed before analysis. But even then, this is not an exact calculation. Syllable counting is not always correct, so scores are a guide line - treat them as such.

First Gunning fog index - The index estimates the years of formal education needed to understand the text on a first reading. The median score for all plugins is 7.6.

Gunning-Fog Plugins
< 8 480 Suitable for a universal audiance
8 - 12 337 Suitable for a wide audiance
>= 12 8 Hard to understand
doctree2filelist, ltr, meta, guestbook, gnuplot, top10, issuetracker, nbsp

An alternative measure is the Flesch Reading Ease score, higher scores indicate material that is easier to read. The median score is 65.2.

Flesh score Plugins
> 60 596 Easily understandable by 13- to 15-year-old students
> 30 220 Hard to understand by 13- to 15-year-old students
<= 30 9 Best understood by university graduates
shibbolethauth, screenplay, arguments, top10, operawidget, picasawebalbum, starred, info, reliwa

Below a graph of the distribution of SMOG score that estimates the years of education needed to understand a piece of writing. The median score is 6.4.

Images

“A picture is worth a thousand words” –- Only 161 (25%) of the syntax and admin plugin homepages have one or more images showing the plugin in action on their homepage. This is a great way to explain complicated stuff.

Devel & experimental

Last year 8 plugins referred to develonly, this has grown to 21 in this survey. These plugins rely on features which are currently only available in the development version of DokuWiki.

32 plugins (4%) are marked experimental.

Plugin content

What does the downloaded plugin contain? Is it localized or did it contain a darcs repro?

Unrelated content

All plugin authors have not removed unnecessary content like SVN directories before zipping their plugins. These 15 plugins (one less than previous year) contained unrelated content at time of survey.

Correct name

Although most users never notice, there are 47 plugins that have a homepage name that differs from the plugin class name. Beside the risk of compatibility problems, the developer will not see any popularity rating in the plugin list. It will also confuse users that actually read the code, question arises “Have I downloaded the right file?”.

See also: Shared class names

Localization

Plugins can be localized by using the plugin framework, this feature was found in approximately 46% of the plugins. Compared to the 60 languages shipped with the DokuWiki 2011-05-25 release, the plugin authors use slightly more languages - 68 in total. Top 10 popular languages supported are same as 2010 but the active Russian community changed the order a little.

Language Plugins
en 374 English
de 160 German
fr 97 French
ru 58 Russian
it 49 Italian
es 48 Spanish
cs 38 Czech
nl 32 Dutch
ja 28 Japanese
pt-br 27 Brazilian Portuguese

Finally a list of plugins with 10 or more languages. The five plugins, beside the bundled, with most translations are same as previous year. The discussion plugin has added 4 languages though.

Continue to ⇒ developers

devel/plugin_survey/friendliness.txt · Last modified: 2016-01-27 22:38 by 81.32.237.8