Categorise distributions into "broad groups of utilisation" for easier searching

I'd like to see more emphasis on the page given to searching for distributions (aka "modules") of interest, and a step that would be needed for that is the categorisation of modules into a relatively few "broad groups of utilisation". (eg. Data Capture, Database, Web framework, Web Presentation, Networking, Scientific, Multimedia, Publishing, etc.). The current tags assigned each module do not provide for this generic search functionality. Perhaps the current tags could provide a second level categorisation.

In Issue #55 @jjatria provided the following guidance to the above suggestion re. 'broad groups of utilisation': "That's an interesting idea, but one that I'm not sure raku.land should implement it. We are trying to index Raku distributions, not necessarily curate them. Maybe what we need is some sort of Task::Kensho-like approach. If such a community resource existed, we could link to it."

I do have some sympathy with the above reasoning, however I also think that it's imperative to put some generic grouping scheme for distributions in place, and that such a scheme to aid searching would provide the most benefit to Raku.Land given that it is the one that is compiling and presenting a long list of items for Raku users to explore.

Therefore, I would say that Raku.Land should be the application / initiative to drive such a classification, and tailor it so it fits the plans for improving the distribution index search capabilities of the website.

It should be a goal that the community can provide help to classify / reclassify distributions according to the scheme so devised. In order for that to happen, the full list of classification groupings should be presented on the webpage. This will also serve to help users decide the categories to search. A drop-down list wouldn't cut it, I'm thinking. I'd say a maximum of 20 groups, perhaps starting with about 15 to begin with.

Then how to populate the groups, well it's a little subjective. Therefore it doesn't need to be exact. We can apply the "80-20" rule, close enough for start up with an initial grouping of categories. Perhaps allow a distribution to belong to a maximum of 2 categories. Also there can be a sizeable 'Uncategorized' group, and so focus the community to provide guidance on how to place these into the current scheme, or perhaps suggest new category groups.

The classification scheme should be able to evolve as feedback on the effectiveness of the search capability begins to trickle back.

Ideally the category groups should split the index fairly evenly. Each category group should have a very short definition abstract.

The overall aim should be that distributions are grouped by their utilisation in a way that resonates with the users visiting the website, and searching for functionality to address their specific need.

Many websites have similar need to group items to help users to search through long lists. One example is on the BBC IPlayer, on their Explore page, where content is grouped under headings such as films / documentaries / comedies / etc. Clearly there is overlap, but the Beeb still deem it worthwhile to group the items like this, because of the benefits for the users having to wade through so much content.