Aster: Automatic abstract syntax
| dc.contributor.advisor | William R. Cook | en |
| dc.creator | Barman, Shaon | en |
| dc.date.accessioned | 2011-09-02T15:09:36Z | en |
| dc.date.available | 2011-09-02T15:09:36Z | en |
| dc.date.issued | 2009-05 | en |
| dc.description.abstract | Modern parser generators typically require two components: an abstract syntax data structure and a concrete syntax grammar specification. Usually the abstract syntax and the concrete syntax are closely related but are specified independently because it is diffcult to distinguish what parts of the concrete syntax are associated with the abstract syntax. There are a few add-ons to parser generators, such as ANTLR, which allow the programmer to specify this relationship through annotations in the concrete syntax, but their scope is limited. In this honors thesis, I propose a new way of generating the abstract syntax data structure. The concrete syntax grammar is used as a base for the abstract syntax data structure. The grammar undergoes a series of transformations, where each transformation analyzes the grammar and removes excess information, such as precedence, associativity, or literals used only to disambiguate potential parsings. The final result is an abstract syntax data structure. In addition, since each of these transformations can be linked together, parser actions can be inserted into the original grammar so that an abstract syntax tree is created when the generated parser is run. | en |
| dc.description.department | Computer Sciences | en |
| dc.identifier.uri | http://hdl.handle.net/2152/13381 | en |
| dc.language.iso | eng | en |
| dc.subject | College of Natural Sciences | en |
| dc.subject | abstract syntax | en |
| dc.subject | concrete syntax | en |
| dc.subject | abstract syntax data structure | en |
| dc.subject | parser | en |
| dc.title | Aster: Automatic abstract syntax | en |
| dc.type | Thesis | en |