A more elaborate statement , with examples, is available.
The best we can expect from various areas of research like theorem proving, lambda-calculus, type theory, etc., is that slowly some ideas from these domains will inspire people who are actually working in software development and that they provide a language to talk about various phenomena which have to be dealt with when designing or handling software. None of these formal areas will ever provide a solution to a practical problem. Topological models are neither better nor worse than order theoretic models and it is not clear that they are useful at all in software technology.
The only attitude is to believe that the mathematical models we are building will be useful mainly in providing a language to describe real phenomena and understand them.
My view of the AMAST identity is that of a group of people who believe that algebraic and syntactic methods will be useful, and should be useful, and that this usefulness may be an aim, but is also our main motivation; let mathematicians develop theories per se and for the only beauty of the art (though I believe in mathematics too the possible use in various application fields like physics or mechanics is essential), whereas let us privilege those papers which are motivated by an eventual usefulness, this being completely independent of the content of the paper and the formal techniques it uses. I can very well imagine that papers in traditionally applied fields (e.g. automata theory or algebraic specification) may prove unsuitable under this light, for their real motivation may be a generalization or extension which is not justified by any real need, whilst a paper using methods that are more sophisticated and a priori farther from applications (such as topological methods or geometrical algebra) may prove perfect for AMAST purposes, for it may bring some new ideas at least on how some real phenomenon has already been described and dealt with in another setting.