I've mail to the guy wrote gimd project and he told me that it is now
an abandoned project so it is not a solution.

I would like to ask about the different models between regular
database and git database.
Usually in relational database (or object database) each entity has a
primary key which identify it and does not every change.
On the other hand git database has objects which identify by the
result of hash function on their content so if the content change (say
the name of the entity) the identity change so there is no constant
identification of entity.
The constant identification in git achieved by embedding file names
into the directory objects.

Another major difference is that in relational database in a one-to-
many relation the many points on the one (order has customer_id), but
in git the one points on the many (directory has file hash).

The last but not least is that relational database has index to
quickly find entities, git does not since it was design to reconstruct
file hierarchy and not execute arbitrary queries (I know that git has
index or cache but it is not index in the sense of databases).

So my question is how can git still be used as database backend.
I'm open for ideas, I just don't see it yet.

Thank you,


You received this message because you are subscribed to the Google Groups "Git 
for human beings" group.
To post to this group, send email to git-us...@googlegroups.com.
To unsubscribe from this group, send email to 
For more options, visit this group at 

Reply via email to