Searching for accessible material for children should be easy. It should be fun as well, if children and illiterate learners are to make use of it when getting started.
Two elements of a catalog for international childrens materials:
Metadata and cataloging standards
Metadata tracked for children's materials -- both those they use and those they create -- should include standard core elements as well as elements indicating how it can be used, how adaptable it is, and its intended audience. Material for international audiences should include detailed information on the cultural background involved.
Cataloging material from diverse sources
Many modern collections are of materials, and from authors, that would not traditionally have been archived or stored in a library. Examples include collections of Scratch or Squeak programs on a website; indexes of articles authored by dozens of people in a collaborative dictionary or encyclopedia; snippets of code and XML uploaded to community sites; repositories of photos and videos uploaded by people often not the original author, sometimes with confusing history; or physical objects that serve as cultural artefacts and can now be stored digitally much more flexibly than in a local museum. [NB -- museums have long dealt with this issue, many in their own ways; only recently has there been convergence with library methods.]
The communities of creators that are currently online are adding things to the global index of URLs at a tremendous rate. Most of this material has little structured metadata, and where it exists this almost always takes the form of tags or categories. If we want something more organized than this for materials that come out of OLPC schools and towns, we need to provide a place to store the resulting records.
Most formally curated catalogs don't accept input from a general audience; the Internet Archive's wayback machine may be an extreme exception.
A library organization interested in storing such materials, focusing on works and elements that would be useful to children in an educational setting, would be a potential host for a catalog of this new type of record. In the first instance, one should be able to find a way to explicitly catalog Scratch and Wikiversity projects that captures the different peculiarities of each.