Open sternk opened 10 years ago
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:7
Igor doesn't know how to "send a development graph or at least some structured theory to the provers, according to their abilities". (And I think, that needs more discussion.) Currently, we can create one Isabelle theory file out of a single development graph node. Assuming we are going to create one Isabelle theory file for every node, we need to create "imports"-parts for inclusions. That is the translation of a single node needs to know the names of incomining nodes and the corresponding morphisms (if they are more than just an inclusions). Another problem is that the signature of an extension should translate only new signature elements and not those from incoming nodes. Maybe therefore a signature_difference method should be added to the class Logic (to be implemented by every Logic instance).
Comment by till Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:8
It is easier to display the resulting development graph in a new window and forget about the proof history. Users want to keep the old DG anyway.
Comment by till Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:9
Try to keep the structure of the libraries, such that reference nodes refering to nodes of other development graphs remain as they are (unless development graph links are completely flattened out).
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:10
This needs to be finished (at least for VSE) until the end of January
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:11
My last comment really applies to #311. The transformations are basically implemented (but may be improved).
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:12
When flattening heterogeneity i.e. in spechub/Hets:VSE/test/List_Rev.casl the nodes are changed in the history, but do no need to change.
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:13
"Proofs -> Flattening -> Disjoint unions" of spechub/Hets:VSE/test/List_Rev.casl creates one additional node. But the same operation after "Proofs -> Automatic" creates 11 additional nodes!
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:14
"Proofs -> Flattening -> Heterogeneity" on spechub/Hets:VSE/test/Imports.casl moves the sentences from Nat into the local theory of List. This should not happen for structured proofs.
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:15
Replying to maeder:
"Proofs -> Flattening -> Heterogeneity" on source:trunk/VSE/test/Imports.casl moves the sentences from Nat into the local theory of List. This should not happen for structured proofs.
This problem was fixed by spechub/Hets@d6ca60f1259d30c191869dce7ec2205ec60aeb37 (but leaves #735)
Comment by maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508#comment:16
I've created #769 and #770 from some comments above. A generic prover template (also see #311) for structured proofs is still missing. (Only VSE can be used for structured proofs, currently.)
Reported by till and assigned to maeder Migrated from http://trac.informatik.uni-bremen.de:8080/hets/ticket/508
There are the following levels of support of structuring mechanisms:
ProverTemplate
in spechub/Hets:Logic/Prover.hs with data indicating provers' abilibties. For each level, realise translation functions translating away the features specific to this level:similar to those in spechub/Hets:Static/DGTranslation.hs Start with translating away featues of level 2. Then, implement translating away featues of level 4.