update TODO and README

README.md

@@ -5,3 +5,12 @@ This is done in a reproducible way using `datalad`.
 In principle, a dataset is created, that is automatically administered by the backlogger, in which data from differnt projects are held together.
 Everything is catalogued by a searchable SQL database, which holds the paths to the respective measurements.
 The original projects can be linked to the dataset and the data may be imported using wrapper functions around the read methonds of pyerrors.
+
+We work with the following nomenclature in this project:
+- Measurement
+    A setis of Observables, including the appropriate metadata.
+- Project
+    A series of measurements that was done by one person as part of their research.
+- Record
+    An entry of a single Correlator in the database of the backlogger.
+- 

TODO.md

@@ -15,3 +15,7 @@
 ## Bugfixes
 - [ ] revisit the reimport function for single files
 - [ ] drop record needs to look if no records are left in a json file.
+
+## Rough Ideas
+- [ ] multitable could provide a high speed implementation of an HDF5 based format
+- [ ] implement also a way to include compiled binaries in the archives.