Jump to content

User:Amire80/2020-12-05-lightning

From Meta, a Wikimedia project coordination wiki

Hello!

Wikimedia sites run on software, which consists of several layers.

Now, you don't need to know anything about software to understand this lightning talk, so please don't run away.

The bottom layer is the core MediaWiki platform and it's shared by all wikis and languages. It lets you do the basics: log in, read, edit, and perform admin tasks.

Above it is the layer of extensions, which is also shared by all the languages. Some familiar extensions are the Mobile frontend; Visual editor; Math formulas; Cite, which lets you add basic references; and so forth.

But there's also another layer: the templates. People don't often think about them as a software layer, even though some of our sites' most prominent features are implemented there:

infoboxes; formatted references; "citation needed"; Wikisource navigation; and many others.

These useful and innovative features are familiar not just to editors, but to the hundreds of millions of our readers. They are used in nearly all the pages.

Despite their prominence, they have an acute problem: they are developed on each wiki separately, even though much of their functionality is needed everywhere.

Because of this, the bigger wikis, like English, German, or Spanish Wikipedias, waste the editors' time as they develop the same features repeatedly, while the small ones, in which there are no people who are skilled at template development, often don't have these features at all.

So, while many people think that all the wikis have the same software, only the lower software platform layer is the same for everyone, and the templates layer is not.

And if the software is not the same for everyone, we don't even have equality between projects, not to mention equity. Knowledge equity.

Without addressing this, it will be impossible to have an effective Methodology to improve user experience research and design, which is initiative number nine. The templates are such a central part of the experience that they cannot be ignored in design. But if you do consider them, you'd have to do this for each wiki separately, which is extremely inefficient.

Without addressing this, it will be impossible to build Resources for newcomers effectively, which is initiative number eleven. Newcomers have to deal with templates from the very start, and if they are different in every wiki, you'd have to build these Resources for each wiki separately, too.

Without addressing this, it will be impossible to develop an equitable Enterprise-level API and the Partnerships around it, initiatives number five and fifteen. The templates are intertwined with the content so much that an API that makes it machine-readable for partners cannot be built effectively if they are different in every language.

All these initiatives are important, and must be addressed, but this problem of the lack of cross-wiki sharing of templates is a blocker for addressing them well. Luckily, there's an initiative that corresponds to addressing it, and it's number fourteen: Cross-project tool development and reuse.

I implore anyone who hears this: Please make sure that you don't ignore the templates layer of our software when talking about other improvements to technology and user experience. It has been ignored for sixteen years, and we shouldn't wait any longer to fix it. The rest of the strategic plan will be much easier to implement if we prioritize it now.

Thank you.