[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[lmi-commits] [4881] Consolidate and simplify documentation
From: |
Greg Chicares |
Subject: |
[lmi-commits] [4881] Consolidate and simplify documentation |
Date: |
Sat, 01 May 2010 11:31:43 +0000 |
Revision: 4881
http://svn.sv.gnu.org/viewvc/?view=rev&root=lmi&revision=4881
Author: chicares
Date: 2010-05-01 11:31:43 +0000 (Sat, 01 May 2010)
Log Message:
-----------
Consolidate and simplify documentation
Modified Paths:
--------------
lmi/trunk/dbvalue.cpp
lmi/trunk/dbvalue.hpp
lmi/trunk/ihs_dbvalue.cpp
Modified: lmi/trunk/dbvalue.cpp
===================================================================
--- lmi/trunk/dbvalue.cpp 2010-05-01 10:42:43 UTC (rev 4880)
+++ lmi/trunk/dbvalue.cpp 2010-05-01 11:31:43 UTC (rev 4881)
@@ -154,23 +154,6 @@
return &data_[z];
}
-/*
-Implementation
- TODO ?? duration should be ***last*** axis
- for a C array with dimensions d0, d1, d2
- int a[d0][d1][d2]
- index
- a[i0][i1][i2]
- is
- *( ((int*)a) + ((((i0*d1)+i1)*s2)+i2) )
- so we can do
- int* p =( ((int*)a) + (((i0*d1)+i1)*s2) )
- for(int j = length; 0 < j; j--)
- *output++ = *p++;
-
- TODO ?? must-be-scalar flag?
-*/
-
int TDBValue::GetKey() const
{
return key_;
Modified: lmi/trunk/dbvalue.hpp
===================================================================
--- lmi/trunk/dbvalue.hpp 2010-05-01 10:42:43 UTC (rev 4880)
+++ lmi/trunk/dbvalue.hpp 2010-05-01 11:31:43 UTC (rev 4881)
@@ -132,18 +132,16 @@
};
/*
-Database items should be allowed to vary across numerous axes, such as
- gender
- underwriting class (e.g. preferred, standard, and various substd tables)
- smoker
- issue age (or attained age as an optional alternative?)
- medical/paramedical/nonmedical
+Some data are scalar by nature--for example, maturity age (DB_EndtAge).
+Someday we might add a flag here to express that constraint; until then,
+it's enforced passim by using a scalar-only Query() function.
+
+Database items might include:
+ attained age as an optional alternative to issue age
rate bands (see below)
-and maybe
+or even:
months (e.g. lapse skewness)
mode (e.g. for lapse rate or mode weighting)
-and last of all
- duration
Does it make sense to use one axis each for
issue age--every year
Modified: lmi/trunk/ihs_dbvalue.cpp
===================================================================
--- lmi/trunk/ihs_dbvalue.cpp 2010-05-01 10:42:43 UTC (rev 4880)
+++ lmi/trunk/ihs_dbvalue.cpp 2010-05-01 11:31:43 UTC (rev 4881)
@@ -492,58 +492,33 @@
return rc;
}
-// potential inlines
-
-//============================================================================
int TDBValue::GetKey() const
{
return key_;
}
-//============================================================================
int TDBValue::GetNDims() const
{
return axis_lengths_.size();
}
-//============================================================================
int TDBValue::GetLength() const
{
LMI_ASSERT(0 < axis_lengths_.size());
return axis_lengths_.back();
}
-//============================================================================
int TDBValue::GetLength(int axis) const
{
LMI_ASSERT(0 <= axis && axis < static_cast<int>(axis_lengths_.size()));
return axis_lengths_[axis];
}
-//============================================================================
std::vector<int> const& TDBValue::GetAxisLengths() const
{
return axis_lengths_;
}
-/*
-Implementation
- duration should be ***last*** axis
- for a C array with dimensions d0, d1, d2
- int a[d0][d1][d2]
- index
- a[i0][i1][i2]
- is
- *( ((int*)a) + ((((i0*d1)+i1)*s2)+i2) )
- so we can do
- int* p =( ((int*)a) + (((i0*d1)+i1)*s2) )
- for(int j = length; 0 < j; j--)
- *output++ = *p++;
-
- bands on...each item?
- scalar marker?
-*/
-
void TDBValue::read(xml::element const& e)
{
std::string short_name;
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [lmi-commits] [4881] Consolidate and simplify documentation,
Greg Chicares <=