When the deposit contains software that was written many years ago but is encapsulated in a container that was created today, dating the content is tricky.
Here an example of the artifacts we receive in a deposit:
```
- metadata in xml
- zipped archive
- Scilab 1.1
- README
- AUTHORS
- LICENCE
- src
- README
[legacy content]
```
`src` directory contains content created in 1980 while the root directory contains data created in 2019.
The idea is the following,
for v1:
- create 1 revision A only with `src`content dated to 1980 with the extracted metadata in `dateCreated`.
- create 1 revision B with the complete deposit `root` with `deposit_date` extracted from SWORD protocol when deposit is received.
- B points to A
for v2:
- create 1 revision C for `src`v2
- create 1 revision D for complete deposit `root`
- C points to A
- D points to B
- C points to D
```
A (1980) -----------C (1981)
/ /
B (2019)-----------D(2019)
```
We need also to consider creating a branch for the `synthetic_history` (A+ C)
And see how we specify exactly which content is legacy inside the deposit itself.