DennisHeimbigner: Created page with "DevelopmentDAP4 << Back to OPULS Development ==Background== Currently, the syntax for "projection" constraints ..."

2013-02-08T20:44:50Z

Created page with "Development DAP4 << Back to OPULS Development ==Background== Currently, the syntax for "projection" constraints ..."

New page

[[Category:Development|Development]][[Category:DAP4|DAP4]]
[[OPULS_Development| << Back to OPULS Development]]

==Background==
Currently, the syntax for "projection" constraints in the spec
looks roughly like this:
: /group1/.../groupn/struct1[a1:b1].field2[a2:b2].struct3[a3:b3].field4[a4:b4]

In looking how John handled this in CDM, I noticed that he used a different but equivalent notation that strikes me a superior.

==Proposal==
Basically, the proposal is to push all of the indexing to the end of the expression. So we now would have (using the above example):
: /group1/.../groupn/struct1.field2.struct3.field4[a1:b1][a2:b2][a3:b3][a4:b4]

==Rationale==
Assuming one has the rank info for the intermediate structs and fields, the two
forms are equivalent. The advantage of this format is that it has a simpler syntax and is easier to parse. It is basically an FQN followed by a sequence of slice specs.

[[User:dmh|Dennis]]

DAP4: DAP4 Projection Syntax - Revision history

DennisHeimbigner: Created page with "DevelopmentDAP4 << Back to OPULS Development ==Background== Currently, the syntax for "projection" constraints ..."