This paper introduces a new generation of the AsTMa= notation which has evolved from earlier versions based on usage patterns from novices and casual users alike. These experiences have shaped a low-barrier language which is optimized for human-centric encoding of semantically rich Topic Maps content.

1.1 Pidgin English
One objective in the development of AsTMa= was to provide a textual language, which can mimick natural language to a certain extent and so enable non-technical users to express assertions about their knowledge domain. Accordingly, knowledge fragments can be organized in a themed, block-oriented way. That supports effective long-term management of highly irregular information as it encourages to move away from the necessity to manage a high number of microscopic statements. As the language builds on a cleaned-up variant of AsTMa 1.x, we present this first.

At this level, the new language is quite comparable to the future ISO standard 13250-6 “Compact Topic Maps Syntax (CTM)” 1. More innovative (at least in a Topic Maps context) is the use of “natural-language” features. The following block

Paul-McCartney isa person and has born-date “18 June 1942”.
Paul-McCartney plays piano and is-member-of The-Beatles,
which isa pop-group and which originated in liverpool.

leads into a topic map with the topic “Paul-McCartney” being registered as an instance of “person” with an occurrence of type “born-date” which carries a date value. Also “piano”, “The-Beatles”, “pop-group” and “liverpool” with the appropriate associations are added.Further features are the detection of data values, a templating infrastructure to reduce syntactic noise, a telex style to keep AsTMa= code on a single line and a consistent subject identification syntax for topics.

1.2 AsTMa?
In a further step the language is generalized by allowing variables as interrogative pronouns in certain places. This results in AsTMa?, a query language which assists non-technical users to retrieve information from topic maps.The query

$who isa person
is internally translated into the TMQL 2 expression
select $who where $who isa person

The AsTMa? processor answers such an query with a more verbose answer than a TMQL engine would. A result set consists of several topic map fragments encoded in AsTMa=. Example answer:
Paul-McCartney isa person
John-Lennon isa person
[…]
The query results can then be merged into the query statement the user provided. Such responses can then easily be used in a text-to-speech environment or as input for other topic maps.