An update to my original slideshow of about 2 years ago. I've added some refinements and (I hope) made the design process a little more clear.

1. Adding
Bi-Temporal
Functionality
to RDBMS

2. ©2008-2015 2
References
●Dr. Richard Snodgrass
● Developing Time-Oriented Database
Applications
● Temporal Data Types
● Kinds of Time: Instant, Interval, Period
● Temporal Statements
● Current, Sequenced, Non-sequenced
● Non-temporal
● Temporal: Past, Current, Future
●Tom Johnston
● Time and Again series

3. ©2008-2015 3
Terminology
●Temporal: fancy-schmancy word
for “time”
● “We’re approaching the temporal anomaly,
Captain!”
●Versioning: tracking changes
● Hoarding data, don't throw anything away
● Mono-temporality (valid/effective). When the
fact or state change took effect.
● Example: source control systems (csv, svn,
etc.)

4. ©2008-2015 4
Terminology
●Bi-Temporal
● Effective time: when did/will it happen?
● Transaction time: when did we record it?
●Version/Temporal Normal Form
(vnf/tnf)
● Really... a new normalization pattern?
● Not really. Special case of 2nf.
● Differences and similarities

5. ©2008-2015 5
Terminology
●Temporally Insensitive Data
● Never changes
● Changes are ignored
●Temporally Sensitive Data
● Changes are tracked
● Changes are scheduled
●Temporal Key
● Composite key with at least one date field
● Version key contains effective date.
● Bi-temporal key is version key with addition
of transaction date field.
} static
} versioned

6. ©2008-2015 6
Why Versioning?
●All changes to TS data are tracked
●“Look back” queries
● Valid/Effective time
● Transaction time
●Database “undo” even after commit
●Changes may be “pre-inserted” or
“pre-updated” to automatically
become current when the time
arrives.

7. ©2008-2015 7
Why Versioning?
● No extraordinary means (such as separate
“history” tables).
● Normal referential constraints.
● All queries follow a consistent pattern.
● Extend CVS/SVN to database.
● All this in a transactional database without
maintaining separate snapshot, history
or audit tables, and with only minor
impact on performance.

8. ©2008-2015 8
Characteristics of Versioning
● Logical / physical differences.
● The row is the entity
● The row is one version of the entity.
● Versions cannot overlap. There cannot be two
different “truths” in effect at the same time.
● There cannot be gaps between versions. The
entity does not simply disappear at some
point then reappear later in a different state.
● One and only one relative current version.
Current to the reference or “as of” time.

9. ©2008-2015 9
Versioning State of the Art
● From Wikipedia: Slowly Changing Dimension
(SCD).
● Data Warehouse patterns
● Type 1: Do nothing
● Type 6: One version almost workable
● Two tables: current versions and past versions.
● Two dates: start and end. (Row spanning
dependency)
● Lot of data changing for “undo.”
● Referential integrity: FK to “current” table.

10. ©2008-2015 10
Minimum Requirements
● One and only one date / time value which
signifies when the version took effect
(no row spanning dependencies).
● NO separate field used to indicate the
current version!
● New version involves one record (version
independence).
● Foreign key references same as always.
● Versioning pattern: consistent and
universal.

11. ©2008-2015 11
1. There cannot be more than one version of the
truth at any given time. In its simplest form,
this means that any query that returns more
than one version of the same entity for the
same time is erroneous.
2. We can’t change the past. This means simply
that physical updates to versioned data should
be prohibited.
i. This does not apply to versions set to
become effective in the future. Plans
change all the time!
Versioning Rules

12. ©2008-2015 12
Temporal Normal Form (tnf)
➲ How it is implemented

13. ©2008-2015 13
Building up to VNF
EmpNo (pk) FName LName HireDate PayRate Other Info
1001 Sally Jones 1998-10-10 15.35 …
1002 Sam Spade 1998-11-11 10.56 …
1004 Katherine Great 1998-12-12 9.00 …
Figure1: Original Non-versioned Data

14. ©2008-2015 14
Building up to VNF
EmpNo (pk) FName LName HireDate PayRate Other Info
1001 Sally Jones 1998-10-10 15.35 …
1002 Sam Spade 1998-11-11 10.56 …
1004 Katherine Great 1998-12-12 9.00 …
Figure1: Original Non-versioned Data
EmpNo (pk) FName LName HireDate PayRate Other Info
1001 Sally Smith 1998-10-10 15.35 …
1002 Sam Spade 1998-11-11 10.56 …
1004 Katherine Great 1998-12-12 9.00 …
Figure 2: “Jones” becomes “Smith” ― option 1

15. ©2008-2015 15
Building up to VNF
EmpNo (pk) FName LName HireDate PayRate Other Info
1001 Sally Jones 1998-10-10 15.35 …
1002 Sam Spade 1998-11-11 10.56 …
1004 Katherine Great 1998-12-12 9.00 …
Figure1: Original Non-versioned Data
EmpNo (pk) FName LName HireDate PayRate Other Info
1001 Sally Jones 1998-10-10 15.35 …
1002 Sam Spade 1998-11-11 10.56 …
1004 Katherine Great 1998-12-12 9.00 …
1005 Sally Smith 1998-10-10 15.35 …
Figure 3: “Jones” becomes “Smith” ― option 2

16. ©2008-2015 16
Building up to VNF
EmpNo (pk) FName LName HireDate PayRate Other Info
1001 Sally Jones 1998-10-10 15.35 …
1002 Sam Spade 1998-11-11 10.56 …
1004 Katherine Great 1998-12-12 9.00 …
Figure1: Original Non-versioned Data
EmpNo (pk) Effective(pk) FName LName HireDate PayRate Other Info
1001 2010-06-19 Sally Jones 1998-10-10 15.35 …
1002 2007-01-01 Sam Spade 1998-11-11 10.56 …
1004 2008-02-26 Katherine Great 1998-12-12 9.00 …
1001 2012-01-01 Sally Smith 1998-10-10 15.35 …
Figure 4: Simple Versioned Data

17. ©2008-2015 17
Temporal Normal Forms
● A table is in Temporal Normal Form when
it is in at least second normal form and
the primary key is a Temporal key.
● Version Normal Form (VNF): primary key
is a version key.
● Temporal Normal Form (TNF): primary
key is a bi-temporal key.
TNF can be used as a generic term, short for “VNF or
TNF.” VNF always is effective time only.

18. ©2008-2015 18
Version Normal Form
EmpNo (pk) Created Deleted FName HireDate Other Info
1001 2008-10-01 9999-12-31 Sally 2008-10-10 …
1002 2008-11-01 9999-12-31 Sam 2008-11-11 …
1004 2008-12-01 9999-12-31 Katherine 2008-12-12 …
Employees Entity Table – after VNF
EmpNo (pk) Effective(pk) Created LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
Temporally Normalized (Versioned) Employee Data
EmpNo (pk) Effective(pk) Created LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
1001 2012-01-01 2012-02-01 Smith 15.35 …

19. ©2008-2015 19
Version DML
declare @EmpNo int;
-- insert static record
insert into Employees( Created, Deleted, HireDate, FName, ... )
values( SysDate(), '9999-12-31', '2008-10-01', 'Sally', ... );
set @EmpNo = @@Identity;
-- insert first version
insert into EmployeeVersions( EmpNo, Effective, Created, LName, ... )
values( @EmpNo, '2008-10-01', SysDate(), 'Jones', ... );
Creating an entity (insert):
Modifying entity data (update):
-- insert new version
insert into EmployeeVersions( EmpNo, Effective, LName, ... )
values( 1001, '2012-01-01', 'Smith', ... );

20. ©2008-2015 20
Version Normal Form
EmpNo (pk) Created Deleted FName HireDate Other Info
1001 2008-10-01 9999-12-31 Sally 2008-10-10 …
1002 2008-11-01 9999-12-31 Sam 2008-11-11 …
1004 2008-12-01 9999-12-31 Katherine 2008-12-12 …
Employees Entity Table – after VNF
Temporally Normalized (Versioned) Employee Data
EmpNo (pk) Effective(pk) Created LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
1001 2012-01-01 2012-02-01 Smith 15.35 …

21. ©2008-2015 21
Version Query
EmpNo Effective FName LName PayRate HireDate Other Info
1001 2010-06-19 Sally Jones 15.35 2008-10-10 …
1002 2011-01-01 Sam Spade 10.56 2008-11-11 …
1004 2011-02-26 Katherine Great 9.00 2008-12-12 …
1001 2012-01-01 Sally Smith 15.35 2008-10-10 …
Figure 5: Simple Join Query – All Rows
Select e.EmpNo, ev.Effective, ev.FName,
ev.Lname, e.HireDate, ev.PayRate, ...
from Employees e
join EmployeeVersions ev
on e.EmpNo = ev.EmpNo
where e.Deleted > GetDate()
and ev.Effective = (
select Max( Effective )
from EmployeeVersions ev1
where ev1.EmpNo = e.EmpNo
and ev1.Effective < GetDate() );

22. ©2008-2015 22
EmpNo(pk) Effective(pk) Created LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
1001 2012-01-01 2012-02-01 Smith 15.35 …
EmpNo(pk) Effective(pk) Created(pk) LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
1001 2012-01-01 2012-02-01 Smith 15.35 …
(Bi-)Temporal Normal Form
EmpNo(pk) Created Deleted FName HireDate Other Info
1001 2008-10-01 9999-12-31 Sally 2008-10-10 …
1002 2008-11-01 9999-12-31 Sam 2008-11-11 …
1004 2008-12-01 9999-12-31 Katherine 2008-12-12 …
Employees Entity Table – after VNF
Bi-Temporally Normalized Employee Data
EmpNo(pk) Effective(pk) Created(pk) LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
1001 2012-01-01 2012-02-01 Smith 15.35 …
1001 2012-01-01 2012-11-17 Smythe 15.35 …

23. ©2008-2015 23
Bi-Temporal Query
EmpNo Effective Created FName LName PayRate HireDate Other Info
1002 2011-01-01 2008-11-01 Sam Spade 10.56 2008-11-11 …
1004 2011-02-26 2008-12-01 Katherine Great 9.00 2008-12-12 …
1001 2012-01-01 2012-11-17 Sally Smythe 15.35 2008-10-10 …
Figure 6: Bi-Temporal Join Query
Select e.EmpNo, ev.Effective, ev.Created, ev.FName,
ev.Lname, e.HireDate, ev.PayRate, ...
from Employees e
join EmployeeVersions ev
on e.EmpNo = ev.EmpNo
where e.Deleted > GetDate()
and ev.Effective = (
select Max( Effective )
from EmployeeVersions ev1
where ev1.EmpNo = e.EmpNo
and ev1.Effective < GetDate() )
and ev.Created = (
select Max( Created )
from EmployeeVersions ev2
where ev2.EmpNo = e.EmpNo
and ev2.Effective = ev.Effective
and ev2.Created < GetDate() );

24. ©2008-2015 24
Versioned One-to-many

25. ©2008-2015 25
Versioned One-to-many
OrderID CustID Shipped Created Deleted
12 123 2012-01-02 2011-11-10 9999-12-31
Orders:
OrdersV:
OrderItems:
OrderItemsV:
OrderID Effective Shipped Total Terms
12 2011-11-15 2012-01-08 $1030.00 5net30
OrderID ItemNo Created Deleted
12 1 2011-11-15 9999-12-31
12 2 2011-11-15 9999-12-31
OrderID ItemNo Effective Product Qty Price
12 1 2011-11-15 1705 10 $30
12 2 2011-11-15 309 2 $500
12 2 2011-11-20 309 1 $515

26. ©2008-2015 26
Versioned One-to-many
select *
from Orders o
join OrdersV ov on o.OrderID = ov.OrderID
where ov.OrderID = @OrderID
and o.Deleted > @AsOf
and ov.Effective = (
select Max( Effective )
from OrdersV ov1
where ov1.OrderID = o.OrderID
and ov1.Effective <= @AsOf );
select *
from OrderItems oi
join OrderItemsV oiv on oi.OrderID = oiv.OrderID and
oi.ItemNo = oiv.ItemNo
where oi.OrderID = @OrderID
and oi.Deleted > @AsOf
and oiv.Effective = (
select Max( Effective )
from OrderItemsV oiv1
where oiv1.OrderID = oi.OrderID
and oiv1.ItemNo = oi.ItemNo
and oiv1.Effective <= @AsOf );

27. ©2008-2015 27
Versioned One-to-many
select o.OrderID, o.CustID, o.Shipped, ov.Terms,
oi.ItemNo, oi.Product, oiv.Qty, oi.Price,
oiv.Qty * oi.Price as ItemTotal
from Orders o
join OrdersV ov on o.OrderID = ov.OrderID
left join OrderItems oi on o.OrderID = oi.OrderID
join OrderItemsV oiv on oi.OrderID = oiv.OrderID and
oi.ItemNo = oiv.ItemNo
where o.OrderID = @OrderID
and ov.Effective = (
select Max( Effective )
from OrdersV ov1
where ov1.OrderID = o.OrderID and
ov1.Effective <= @AsOf )
and oiv.Effective = (
select Max( Effective )
from OrderItemsV oiv1
where oiv1.OrderID = oi.OrderID
and oiv1.ItemNo = oi.ItemNo
and oiv1.Effective <= @AsOf );

28. ©2008-2015 28
Version Analysis
Version Slicing
Time
A
VA5VA5
VA4VA4
VA3VA3
VA2VA2
VA1VA1
B
VB5VB5
VB4VB4
VB2VB2
VB1VB1
VB3VB3
C
VC3VC3
VC1VC1
VC2VC2
Current
As of

29. ©2008-2015 29
Version Analysis
Version History
V11
V1
V2
V3
V4
V5
V6
V7
V8
V9
V10
A
VA5VA5
VA4VA4
VA3VA3
VA2VA2
VA1VA1
B
VB5VB5
VB4VB4
VB2VB2
VB1VB1
VB3VB3
C
VC3VC3
VC1VC1
VC2VC2
Time

30. ©2008-2015 30
Version Analysis
Effective Data
02/01/2008 VA1
02/02/2008 VA2
02/05/2008 VA3
02/07/2008 VA4
02/11/2008 VA5
Effective Data
02/01/2008 VB1
02/03/2008 VB2
02/06/2008 VB3
02/08/2008 VB4
02/10/2008 VB5
Effective Data
02/01/2008 VC1
02/04/2008 VC2
02/09/2008 VC3
Table A Table B Table C

31. ©2008-2015 31
Version Analysis
select a.Effective as AEff, a.Data as AData,
b.Effective as BEff, b.Data as BData,
c.Effective as CEff, c.Data as CData
from A_Data a
join B_Data b on a.A_ID = b.B_ID
join C_Data c on a.A_ID = c.C_ID
where b.Effective = (
select Max( Effective )
from B_Data
where Effective <= a.Effective
)
and c.Effective = (
select Max( Effective )
from C_Data
where Effective <= a.Effective
)
union all
...

32. ©2008-2015 32
Version Analysis
select a.Effective as AEff, a.Data as AData,
b.Effective as BEff, b.Data as BData,
c.Effective as CEff, c.Data as CData
from A_Data a
join B_Data b on a.A_ID = b.B_ID
join C_Data c on a.A_ID = c.C_ID
where a.Effective = (
select Max( Effective )
from A_Data
where Effective <= b.Effective
)
and c.Effective = (
select Max( Effective )
from C_Data
where Effective <= b.Effective
)
union all

33. ©2008-2015 33
Version Analysis
...
select a.Effective as AEff, a.Data as AData,
b.Effective as BEff, b.Data as BData,
c.Effective as CEff, c.Data as CData
from A_Data a
join B_Data b on a.A_ID = b.B_ID
join C_Data c on a.A_ID = c.C_ID
where a.Effective = (
select Max( Effective )
from A_Data
where Effective <= c.Effective
)
and b.Effective = (
select Max( Effective )
from B_Data
where Effective <= c.Effective
);

34. ©2008-2015 34
Version Analysis
AEff AData BEff BData CEff CData
02/01/2008 VA1 02/01/2008 VB1 02/01/2008 VC1
02/02/2008 VA2 02/01/2008 VB1 02/01/2008 VC1
02/02/2008 VA2 02/03/2008 VB2 02/01/2008 VC1
02/02/2008 VA2 02/03/2008 VB2 02/04/2008 VC2
02/05/2008 VA3 02/03/2008 VB2 02/04/2008 VC2
02/05/2008 VA3 02/06/2008 VB3 02/04/2008 VC2
02/07/2008 VA4 02/06/2008 VB3 02/04/2008 VC2
02/07/2008 VA4 02/08/2008 VB4 02/04/2008 VC2
02/07/2008 VA4 02/08/2008 VB4 02/09/2008 VC3
02/07/2008 VA4 02/10/2008 VB5 02/09/2008 VC3
02/11/2008 VA5 02/10/2008 VB5 02/09/2008 VC3

35. ©2008-2015 35
Version Analysis
Select
case
when AEff > BEff and AEff > CEff
then AEff
when BEff > AEff and BEff > CEff
then BEff
else CEff
end as Effective,
AData, BData, CData
from (
-- previous query
) as History;
Effective AData BData CData
02/01/2008 VA1 VB1 VC1
02/02/2008 VA2 VB1 VC1
02/03/2008 VA2 VB2 VC1
02/04/2008 VA2 VB2 VC2
02/05/2008 VA3 VB2 VC2
02/06/2008 VA3 VB3 VC2
02/07/2008 VA4 VB3 VC2
02/08/2008 VA4 VB4 VC2
02/09/2008 VA4 VB4 VC3
02/10/2008 VA4 VB5 VC3
02/11/2008 VA5 VB5 VC3
History

36. ©2008-2015 36
Version Views
create view Orders as
select os.OrderID, os.CustID, os.Shipped, ov.Total,
ov.Terms, ov.Created as DateModified
from OrdersS os
join OrdersV ov on ov.OrderID = os.OrderID
where os.Deleted > GetDate() and ov.Effective = (
select Max( ov1.Effective )
from OrdersV ov1
where ov1.OrderID = ov.OrderID and
ov1.Effective <= GetDate() )
Orders View (current)

37. ©2008-2015 37
Version Views
create view OrderItems as
select oiv.OrderID, oiv.ItemID, oiv.Product, oiv.Qty,
oiv.Price, oiv.Effective as DateModified
from OrderItemsS ois
join OrderItemsV oiv on oiv.OrderID = ois.OrderID and
oiv.ItemNo = ois.ItemNo
where oiv.Deleted > GetDate()
and oiv.Effective = (
select Max( Effective )
from OrderItemsV oiv1
where oiv1.OrderID = oiv.OrderID
and oiv1.ItemID = oiv.ItemID
and oiv1.Effective <= GetDate() )
OrderItems View (current)

38. ©2008-2015 38
Version Views
select o.OrderID, o.CustID, o.Shipped, o.Total, o.Terms,
oi.ItemID, oi.Product, oi.Qty, oi.Price
from Orders o
left join OrderItems oi
on o.OrderID = oi.OrderID;
Querying Current Views

39. ©2008-2015 39
Version Views
● Insert
● Separate static from versioned
● Expose 'Effective' so it can be entered or use current
timestamp?
● Update
● Separate static from versioned
● Update static data, insert new version data as needed
● If 'Effective' exposed, ignore or use updated value?
● Delete
● Hard, firm, soft
● Hard: physical delete. Not recommended with temporal data.
● Firm: enforce “once deleted, stays deleted.”
● Soft: Delete/undelete operations are available.
View triggers

40. ©2008-2015 40
● How it is implemented
● When it is implemented
● Data must be tracked through time
● Past data must be reconstructed
● Undo ― even after commit
● Planned changes pre-inserted
● Time-related analysis
Temporal Normal Form (tnf)

41. ©2008-2015 41
● How it is implemented
● When it is implemented
● Check against requirements
● One and only one date/time value (no rsd)
● No separate field to flag “current”
● Version independence
● No FK tricks
● Consistent pattern
Temporal Normal Form (tnf)

42. ©2008-2015 42
● How it is implemented
● When it is implemented
● Check against requirements
● The future
Temporal Normal Form (tnf)

43. ©2008-2015 43
Temporal (TSQL/SQL2/SQL3)
NONSEQUENCED TRANSACTION AND VALID PERIOD
'2005-01-01 - 2006-01-01'
SELECT e1.ename, e1.street AS old_street,
e2.street AS new_street,
BEGIN(TRANSACTION(e2)) AS trans_time
FROM employee AS e1, employee AS e2
WHERE e1.eno = e2.eno AND
TRANSACTION(e1) MEETS TRANSACTION(e2)
AND e1.street <> e2.street;
Find employees who changed address during 2005.
Display both the previous and the new addresses.

44. ©2008-2015 44
Temporal (TNF)
Select e.ename, ev1.street as old_street,
ev2.street as new_street,
ev2.Effective as move_time
From Employees e
Join EmployeeVersions ev1
on ev1.EmpNo = e.EmpNo
and ev1.Effective >= '2005-01-01'
and ev1.Effective < '2006-01-01'
Join EmployeeVersions ev2
on ev2.EmpNo = e.EmpNo
and ev2.Effective =(
Select Max( ev.Effective )
From EmployeeVersions ev
Where ev.Effective > ev1.Effective
And ev.Effecitve < '2006-01-01')
Where e.Deleted > GetDate();

45. ©2008-2015 45
Temporal (Suggested)
Select e1.ename, e1.street as old_street,
e2.street as new_street,
e2.Effective as move_time,
e2.Transaction as trans_time
From Employees e1
Join Employees e2
On e2.EmpNo = e1.EmpNo
And e2.AsOf > e1.AsOf
Where Asof Effective between
'2005-01-01' and '2006-01-01';
Compare to the SQL2/3 query!
“AsOf” is both a keyword signifying a temporal query (in the WHERE clause)
and a pseudo-column for either the effective date field (in an effective look-
back) or the transaction date field (in a transaction look-back). “Effective” and
“Transaction” are also pseudo-columns to allow more precision and are not
dependent on the type of look-back.

46. ©2008-2015 46
Version Normal Form
EmpNo(pk) Created Deleted FName HireDate Other Info
1001 2008-10-01 9999-12-31 Sally 2008-10-10 …
1002 2008-11-01 9999-12-31 Sam 2008-11-11 …
1004 2008-12-01 9999-12-31 Katherine 2008-12-12 …
Static Employees Table
Versioned Employees (Sub)Table
EmpNo (pk) Effective(pk) Created LName PayRate Other Info
1001 2008-10-10 2008-10-01 Jones 15.35 …
1002 2008-11-11 2008-11-01 Spade 10.56 …
1004 2008-12-12 2008-12-01 Great 9.00 …
1001 2012-01-01 2012-02-01 Smith 15.35 …

47. ©2008-2015 47
create temporal index NAME on Employees
( LName, PayRate, ... ) -- PK field(s) not specified
Effective eff_date_field
Transaction trans_date_field;
select EmpNo, Effective, FName,
Lname, HireDate, PayRate, ...
from Employees
where ...
[asof [effective|transaction] [date|all|between d1 and d2]];
-- Defaults to "asof eff current_timestamp"
update Employees
set LName = 'Smith'
where ...
[asof [effective] [date|all|between d1 and d2]];
-- "asof trans" not allowed
insert Employees (...)
values (...)
[asof [effective] date];
-- "asof trans" not allowed
Temporal Index

48. ©2008-2015 48
Proof of Concept
● Open Source DBMS
● Download source
● Take apart
● Reassemble with t-index
● Two/thirds complete
● Remaining to be done...

49. ©2008-2015 49
Temporal Termination
Questions
Google Group
VRDBMS:
https://groups.google.com/forum/?hl=en&fromgroups#!forum/vrdbms
Sources
Wikipedia:
● http://en.wikipedia.org/wiki/Temporal_database
Time and Again series, Randy Weis & Tom Johnson:
● http://www.inbaseinc.com/dm-series.htm
Richard Snodgrass:
● http://www.cs.arizona.edu/people/rts/tdbbook.pdf