Avoid using float datatypes to store data

[jaragunde]

There are certain kinds of data that we are storing using floating-point datatypes. We should generally avoid that, as they are imprecise and we could lose data. We should analyze it case by case.

One example:

create table project (
  id                        serial not null,
  activation                boolean default true not null,
  init                      date,
  _end                      date,
  invoice                   double precision,
  est_hours                 double precision,
  areaId                    integer not null,
  description               varchar(256),
  type                      varchar(256),
  moved_hours               double precision,
  sched_type                varchar(256),
  constraint pk_Project primary key (id)
) ;

We could store the "invoice" as an integer of cents, and hours as integers of minutes like we do for tasks.