Download Full Version ebookmass - Visit ebookmass.

com

Architecting a Modern Data Warehouse for Large

Enterprises: Build Multi-cloud Modern Distributed
Data Warehouses with Azure and AWS 1st Edition
Anjani Kumar
https://ebookmass.com/product/architecting-a-modern-data-
warehouse-for-large-enterprises-build-multi-cloud-modern-
distributed-data-warehouses-with-azure-and-aws-1st-edition-
anjani-kumar-3/

OR CLICK HERE

DOWLOAD NOW

Discover More Ebook - Explore Now at ebookmass.com

 Instant digital products (PDF, ePub, MOBI) ready for you
Download now and discover formats that fit your needs...

Architecting a Modern Data Warehouse for Large

Enterprises: Build Multi-cloud Modern Distributed Data
Warehouses with Azure and AWS 1st Edition Anjani Kumar
https://ebookmass.com/product/architecting-a-modern-data-warehouse-
for-large-enterprises-build-multi-cloud-modern-distributed-data-
warehouses-with-azure-and-aws-1st-edition-anjani-kumar/
ebookmass.com

Architecting a Modern Data Warehouse for Large

Enterprises: Build Multi-cloud Modern Distributed Data
Warehouses with Azure and AWS 1st Edition Anjani Kumar
https://ebookmass.com/product/architecting-a-modern-data-warehouse-
for-large-enterprises-build-multi-cloud-modern-distributed-data-
warehouses-with-azure-and-aws-1st-edition-anjani-kumar-2/
ebookmass.com

Modern Data Architecture on Azure: Design Data-centric

Solutions on Microsoft Azure 1st Edition Sagar Lad

https://ebookmass.com/product/modern-data-architecture-on-azure-
design-data-centric-solutions-on-microsoft-azure-1st-edition-sagar-
lad/
ebookmass.com

Petroleum Refining Design and Applications Handbook,

Volume 4 A. Kayode Coker

https://ebookmass.com/product/petroleum-refining-design-and-
applications-handbook-volume-4-a-kayode-coker/

ebookmass.com
Climate Economics: A Call for More Pluralism And
Responsibility 1st ed. Edition Michael Roos

https://ebookmass.com/product/climate-economics-a-call-for-more-
pluralism-and-responsibility-1st-ed-edition-michael-roos/

ebookmass.com

Pharmacology and Toxicology of Cytochrome P450 - 60th

Anniversary 1st Edition Hiroshi Yamazaki

https://ebookmass.com/product/pharmacology-and-toxicology-of-
cytochrome-p450-60th-anniversary-1st-edition-hiroshi-yamazaki/

ebookmass.com

Basic Immunology E-Book: Functions and Disorders of the

Immune System 6th Edition – Ebook PDF Version

https://ebookmass.com/product/basic-immunology-e-book-functions-and-
disorders-of-the-immune-system-6th-edition-ebook-pdf-version/

ebookmass.com

The Christmas Wish Sharon Sala

https://ebookmass.com/product/the-christmas-wish-sharon-sala/

ebookmass.com

A False Tree of Liberty: Human Rights In Radical Thought

1st Edition Edition Susan Marks

https://ebookmass.com/product/a-false-tree-of-liberty-human-rights-in-
radical-thought-1st-edition-edition-susan-marks/

ebookmass.com
World-Class Warehousing and Material Handling, 2E – Ebook
PDF Version

https://ebookmass.com/product/world-class-warehousing-and-material-
handling-2e-ebook-pdf-version/

ebookmass.com
Anjani Kumar, Abhishek Mishra and Sanjeev Kumar

Architecting a Modern Data Warehouse

for Large Enterprises
Build Multi-cloud Modern Distributed Data
Warehouses with Azure and AWS
Anjani Kumar
Gurgaon, India

Abhishek Mishra
Thane West, Maharashtra, India

Sanjeev Kumar
Gurgaon, Haryana, India

ISBN 979-8-8688-0028-3 e-ISBN 979-8-8688-0029-0

https://doi.org/10.1007/979-8-8688-0029-0

© Anjani Kumar, Abhishek Mishra, and Sanjeev Kumar 2024

This work is subject to copyright. All rights are solely and exclusively
licensed by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in
any other physical way, and transmission or information storage and
retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.

The use of general descriptive names, registered names, trademarks,

service marks, etc. in this publication does not imply, even in the
absence of a specific statement, that such names are exempt from the
relevant protective laws and regulations and therefore free for general
use.

The publisher, the authors, and the editors are safe to assume that the
advice and information in this book are believed to be true and accurate
at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, expressed or implied, with respect to the
material contained herein or for any errors or omissions that may have
been made. The publisher remains neutral with regard to jurisdictional
claims in published maps and institutional affiliations.

This Apress imprint is published by the registered company APress

Media, LLC, part of Springer Nature.
The registered company address is: 1 New York Plaza, New York, NY
10004, U.S.A.
I dedicate this book to my mother, Prabhawati; my aunt, Sunita; and my
wife, Suchi.
— Anjani Kumar
I dedicate this book to my lovely daughter, Aaria.
— Abhishek Mishra
I dedicate this book to my late father, Shri Mahinder Nath.
— Sanjeev Kumar
Any source code or other supplementary material referenced by the
author in this book is available to readers on GitHub
(https://github.com/Apress). For more detailed information, please
visit https://www.apress.com/gp/services/source-code.
Acknowledgments
We would like to thank Apress for giving us the opportunity to work on
this book. Also, thanks to the technical reviewer and the editor and the
entire Apress team for supporting us on this journey.
Table of Contents
Chapter 1:​Introduction
Objective
Origin of Data Processing and Storage in the Computer Era
Evolution of Databases and Codd Rules
Transitioning to the World of Data Warehouses
Data Warehouse Concepts
Data Sources (Data Format and Common Sources)
ETL (Extract, Transform, Load)
Data Mart
Data Modeling
Cubes and Reporting
OLAP
Metadata
Data Storage Techniques and Options
Evolution of Big Data Technologies and Data Lakes
Transition to the Modern Data Warehouse
Traditional Big Data Technologies
The Emergence of Data Lakes
Data Lake House and Data Mesh
Transformation and Optimization between New vs.​Old
(Evolution to Data Lake House)
A Wider Evolving Concept Called Data Mesh
Building an Effective Data Engineering Team
An Enterprise Scenario for Data Warehousing
Summary
Chapter 2:​Modern Data Warehouses
Objectives
Introduction to Characteristics of Modern Data Warehouse
Data Velocity
Data Variety
Volume
Data Value
Fault Tolerance
Scalability
Interoperability​
Reliability
Modern Data Warehouse Features:​Distributed Processing,
Storage, Streaming, and Processing Data in the Cloud
Distributed Processing
Storage
Streaming and Processing
Autonomous Administration Capabilities
Multi-tenancy and Security
Performance
What Are NoSQL Databases?​
Key–Value Pair Stores
Document Databases
Columnar DBs
Graph Databases
Case Study:​Enterprise Scenario for Modern Cloud-based Data
Warehouse
 Advantages of Modern Data Warehouse over Traditional Data
Warehouse
Summary
Chapter 3:​Data Lake, Lake House, and Delta Lake
Structure
Objectives
Data Lake, Lake House, and Delta Lake Concepts
Data Lake, Storage, and Data Processing Engines Synergies and
Dependencies
Implement Lake House in Azure
Create a Data Lake on Azure and Ingest the Health Data CSV
File
Create an Azure Synapse Pipeline to Convert the CSV File to
a Parquet File
Attach the Parquet File to the Lake Database
Implement Lake House in AWS
Create an S3 Bucket to Keep the Raw Data
Create an AWS Glue Job to Convert the Raw Data into a Delta
Table
Query the Delta Table using the AWS Glue Job
Summary
Chapter 4:​Data Mesh
Structure
Objectives
The Modern Data Problem and Data Mesh
Data Mesh Principles
Domain-driven Ownership
Data-as-a-Product
 Self-Serve Data Platform
Federated Computational Governance
Design a Data Mesh on Azure
Create Data Products for the Domains
Create Self-Serve Data Platform
Federated Governance
Summary
Chapter 5:​Data Orchestration Techniques
Structure
Objective
Data Orchestration Concepts
Modern Data Orchestration in Detail
Evolution of Data Orchestration
Data Integration
Middleware and ETL Tools
Enterprise Application Integration (EAI)
Service-Oriented Architecture (SOA)
Data Warehousing
Real-Time and Streaming Data Integration
Cloud-Based Data Integration
Data Integration for Big Data and NoSQL
Self-Service Data Integration
Data Pipelines
Data Processing using Data Pipelines
Benefits and Advantages of Data Pipelines
Common Use Cases for Data Pipelines
 Data Governance Empowered by Data Orchestration:​
Enhancing Control and Compliance
Achieving Data Governance through Data Orchestration
Tools and Examples
Azure Data Factory
Azure Synapse
Summary
Chapter 6:​Data Democratization, Governance, and Security
Objectives
Introduction to Data Democratization
Factors Driving Data Democratization
Layers of Democratization Architecture
Self-Service
Data Catalog and Data Sharing
People
Tools and Technology:​Self-Service Tools
Data Governance Tools
Introduction to Data Governance
Ten Key Factors that Ensure Successful Data Governance
Data Stewardship
Models of Data Stewardship
Data Security Management
Security Layers
Data Security Approach
Types of Controls
Data Security in Outsourcing Mode
 Popular Information Security Frameworks
Major Privacy and Security Regulations
Major Modern Security Management Concepts
Practical Use Case for Data Governance and Data
Democratization
Problem Statement
High-Level Proposed Solution
Summary
Chapter 7:​Business Intelligence
Structure
Objectives
Introduction to Business Intelligence
Descriptive Reports
Predictive Reports
Prescriptive Reports
Business Intelligence Tools
Query and Reporting Tools
Online Analytical Processing (OLAP) Tools
Analytical Applications
Trends in Business Intelligence (BI)
Business Decision Intelligence Analysis
Self-Service
Advanced BI Analytics
BI and Data Science Together
Data Strategy
Data and Analytics Approach and Strategy
 Summary
Index
About the Authors
Anjani Kumar
is the managing director and founder of
MultiCloud4u, a rapidly growing startup
that helps clients and partners
seamlessly implement data-driven
solutions for their digital businesses.
With a background in computer science,
Anjani began his career researching and
developing multi-lingual systems that
were powered by distributed processing
and data synchronization across remote
regions of India. He later collaborated
with companies such as Mahindra
Satyam, Microsoft, RBS, and Sapient to
create data warehouses and other data-based systems that could
handle high-volume data processing and transformation.

Dr. Abhishek Mishra

is a cloud architect at a leading
organization and has more than a decade
and a half of experience building and
architecting software solutions for large
and complex enterprises across the
globe. He has deep expertise in enabling
digital transformations for his customers
using the cloud and artificial intelligence.
Sanjeev Kumar
heads up a global data and analytics
practice at the leading and oldest
multinational shoe company with
headquarters in Switzerland. He has 19+
years of experience working for
organizations in multiple industries
modeling modern data solutions. He has
consulted with some of the top
multinational firms and enabled digital
transformations for large enterprises
using modern data warehouses in the
cloud. He is an expert in multiple fields
of modern data management and
execution, including data strategy,
automation, data governance,
architecture, metadata, modeling,
business intelligence, data management,
and analytics.
About the Technical Reviewer
Viachaslau Matsukevich
is an industry expert with over a decade
of experience in various roles, including
DevOps, cloud, solutions architecture,
tech leadership, and infrastructure
engineering.
As a cloud solutions architect,
Viachaslau has delivered 20+ DevOps
projects for a number of Fortune 500
and Global 2000 enterprises. He holds
certifications from Microsoft, Google,
and the Linux Foundation, including
Solutions Architect Expert, Professional
Cloud Architect, and Kubernetes
Administrator.
Viachaslau authors technology articles about cloud-native
technologies and Kubernetes, for platforms such as Red Hat Enable
Architect, SD Times, Hackernoon, and Dzone.
In addition to his technical expertise, Viachaslau serves as a
technical reviewer for technology books, ensuring the quality and
accuracy of the latest publications.
He has also made significant contributions as an industry expert
and judge for esteemed awards programs, including SIIA CODiE Awards
and Globee Awards (including IT World Awards, Golden Bridge Awards,
Disruptor Company Awards, and American Best in Business Awards).
Viachaslau has also lent his expertise as a judge in over 20 hackathons.
Viachaslau is also the author of online courses covering a wide array
of topics related to cloud, DevOps and Kubernetes tools.
Follow Viachaslau on LinkedIn:
https://www.linkedin.com/in/viachaslau-
matsukevich/
© Anjani Kumar, Abhishek Mishra, and Sanjeev Kumar 2024
A. Kumar et al., Architecting a Modern Data Warehouse for Large Enterprises
https://doi.org/10.1007/979-8-8688-0029-0_1

1. Introduction
Anjani Kumar1 , Abhishek Mishra2 and Sanjeev Kumar3
(1) Gurgaon, India
(2) Thane West, Maharashtra, India
(3) Gurgaon, Haryana, India

In the early days of computing, businesses struggled to keep up with

the flood of data. They had few options for storing and analyzing data,
hindering their ability to make informed decisions. As technology
improved, businesses recognized the value of data and needed a way to
make sense of it. This led to the birth of data warehousing, coined by
Bill Inmon in the 1980s. Inmon’s approach was focused on structured,
relational data for reporting and analysis. Early data warehouses were
basic but set the stage for more advanced solutions as businesses
gained access to more data. Today, new technologies like Big Data and
data lakes have emerged to help deal with the increasing volume and
complexity of data. The data lakehouse combines the best of data lakes
and warehouses for real-time processing of both structured and
unstructured data, allowing for advanced analytics and machine
learning. While the different chapters of this book cover all aspects of
modern data warehousing, this chapter specifically focuses on the
transformation of data warehousing techniques from past to present to
future, and how it impacts building a modern data warehouse.
In this chapter we will explore the following:
History and Evolution of Data Warehouse
Basic Concepts and Features of Data Warehouse
Advantages and Examples of Cloud-based Data Warehouse
Enterprise Scenario for Data Warehouse
Objective
This chapter provides an overview of data warehouses and familiarizes
the readers with the terminologies and concepts of data warehouses.
The chapter further focuses on the transformation of data warehousing
techniques from past to present to future, and how it impacts building a
modern data warehouse.
After studying this chapter, you should be able to do the following:
Understand the basics of data warehousing, from the tools,
processes, and techniques used in modern-day data warehousing to
the different roles and responsibilities of a data warehouse team.
Set up a synergy between engineering and operational communities,
even when they’re at different stages of learning and implementation
maturity.
Determine what to adopt and what to ignore, ensuring your team
stays up to date with the latest trends in data warehousing.
Whether you’re starting a data warehouse team or just looking to
expand your knowledge, this guide is the perfect place to start. It will
provide you with a background on the topics covered in detail in
further chapters, allowing you to better understand the nuances of data
warehousing and become an expert in the field.

Origin of Data Processing and Storage in the

Computer Era
The history of data processing and storage dates back to the early 20th
century when mechanical calculators were used for basic arithmetic
operations. However, it wasn’t until the mid-20th century that
electronic computers were developed, which revolutionized data
processing and storage.
The first electronic computer, the ENIAC (Electronic Numerical
Integrator and Computer), was built in 1946 by J. Presper Eckert and
John Mauchly. It was a massive machine that filled an entire room and
used vacuum tubes to perform calculations. ENIAC was primarily used
for military purposes, such as computing artillery firing tables.
 In the 1950s and ’60s, the development of smaller and faster
transistors led to the creation of smaller and more efficient computers.
The introduction of magnetic tape and magnetic disks in the late 1950s
allowed for the storage of large amounts of data, which could be
accessed much more quickly than with punched cards or paper tape.
In the 1970s, the development of integrated circuits (ICs) made it
possible to create even smaller and more powerful computers. This led
to the development of personal computers in the 1980s, which were
affordable and accessible to a wide range of users.
Today, data processing and storage are essential to nearly every
aspect of modern life, from scientific research to business and
commerce to entertainment. The rapid growth of the modern storage
solution powered by SSD and flash memory and by internet and cloud
computing has made it possible to store and access vast amounts of
data from almost anywhere in the world.
In conclusion, the origin of data processing and storage can be
traced back to the early 20th century, but it wasn’t until the
development of electronic computers in the mid-20th century that
these processes became truly revolutionary. From massive room-sized
machines to powerful personal computers, data processing and storage
have come a long way and are now essential to almost every aspect of
modern life.

Evolution of Databases and Codd Rules

The evolution of databases began with IBM’s development of the first
commercially successful database management system (DBMS) in the
1960s. The relational model of databases, introduced by E.F. Codd in
the 1970s, organized data into tables consisting of rows and columns,
leading to the development of Structured Query Language (SQL). The
rise of the internet and e-commerce in the 1990s led to the
development of NoSQL databases for handling vast amounts of
unstructured data. The Chord protocol, proposed in 2001, is a
distributed hash table (DHT) algorithm used for maintaining the
consistency and reliability of data across multiple nodes in distributed
databases.
 Codd’s 12 principles for relational databases established a
framework for designing and implementing a robust, flexible, and
scalable data management system. These principles are relevant in data
warehousing today because they provide a standard for evaluating data
warehousing systems and ensuring that they can handle large volumes
of data, support complex queries, maintain data integrity, and evolve
over time to meet changing business needs.
The 12 principles of Codd’s rules for relational databases are as
follows:
1. Information Rule: All data in the database should be represented
as values in tables. This means that the database should be
structured as a collection of tables, with each table representing a
single entity or relationship.

2. Guaranteed Access Rule: Each value in the database should be

accessible by specifying its table name, primary key value, and
column name. This ensures that every piece of data in the
database is uniquely identifiable and can be accessed efficiently.

3. Systematic Treatment of Null Values: The database should

support the use of null values to represent missing or unknown
data. These null values should be treated consistently throughout
the system, with appropriate support for operations such as null
comparisons and null concatenations.

4. Dynamic Online Catalog Based on the Relational Model: The

database should provide a dynamic online catalog that describes
the structure of the database in terms of tables, columns, indexes,
and other relevant information. This catalog should be accessible
to users and applications and should be based on the relational
model.

5. Comprehensive Data Sublanguage Rule: The database should

support a comprehensive data sublanguage that allows users to
define, manipulate, and retrieve data in a variety of ways. This
sublanguage should be able to express complex queries, data
definitions, and data modifications.
6. View Updating Rule: The database should support the updating
of views, which are virtual tables that are defined in terms of
other tables. This allows users to modify data in a flexible and
intuitive way, without having to worry about the underlying
structure of the database.

7. High-Level Insert, Update, and Delete Rule: The database

should support high-level insert, update, and delete operations
that allow users to modify multiple rows or tables at once. This
simplifies data management and improves performance by
reducing the number of database interactions required.

8. Physical Data Independence: The database should be able to

store and retrieve data without being affected by changes to the
physical storage or indexing structure of the database. This allows
the database to evolve over time without requiring significant
changes to the application layer.

9. Logical Data Independence: The database should be able to

store and retrieve data without being affected by changes to the
logical structure of the database. This means that the database
schema can be modified without requiring changes to the
application layer.

10. Integrity Independence: The database should be able to enforce

integrity constraints such as primary keys, foreign keys, and other
business rules without being affected by changes to the
application layer. This ensures that data is consistent and accurate
at all times.

11. Distribution Independence: The database should be able to

distribute data across multiple locations without being affected by
changes to the application layer. This allows the database to scale
horizontally and geographically without requiring changes to the
application layer.

12. Non-Subversion Rule: The database should not be susceptible to

subversion by unauthorized users or applications. This means
 that the database should enforce access controls, encryption, and
other security measures to protect against unauthorized access or
modification of data.

Traditional tabular systems based on Codd rules were relevant, but

with the rise of the internet and e-commerce, there was a huge
increase in the volume and variety of data being generated. To
handle this data, new NoSQL databases were developed, which are
more flexible and scalable, especially for unstructured data. In
building a universally accepted data warehouse, it’s important to
consider the strengths and weaknesses of both traditional and
NoSQL databases and follow best practices, such as data quality, data
modeling, data governance, and security measures. In the upcoming
section of this chapter, we will explore this transition in a step-by-
step manner while giving special attention to the areas that remain
relevant for creating a strong and widely accepted modern data
warehouse.

Transitioning to the World of Data Warehouses

In the 1970s, the dominant form of database used in business was the
hierarchical database, which organized data in a tree-like structure,
with parent and child nodes. However, as businesses began to collect
more data and as the need for complex querying and reporting
increased, it became clear that the hierarchical database was not
sufficient.
This led to the development of the network database, which allowed
for more complex relationships between data, but it was still limited in
its ability to handle large volumes of data and complex querying. As a
result, the relational database model was developed, which organized
data into tables consisting of rows and columns, allowing for more
efficient storage and easier retrieval of information.
However, the relational model was not without its limitations. As
businesses continued to collect more data, the need for a centralized
repository to store and manage data became increasingly important.
This led to the development of the data warehouse, which is a large,
centralized repository of data that is optimized for reporting and
analysis.
The data warehouse is designed to handle large volumes of data
from multiple sources and to provide a single source of truth for
reporting and analytics. Data warehouses use specialized technologies,
such as extract, transform, load (ETL) processes, to extract data from
multiple sources, transform it into a common format, and load it into
the data warehouse.
Data warehouses also use specialized tools for querying and
reporting, such as online analytical processing (OLAP), which allows
users to analyze data across multiple dimensions, and data mining,
which uses statistical and machine learning techniques to identify
patterns and relationships in the data.

The world transitioned to data warehousing from databases in the

1970s as businesses realized the limitations of the hierarchical and
network database models when handling large volumes of data and
complex querying. The development of the data warehouse provided
a centralized repository for storing and managing data, as well as
specialized tools for reporting and analysis. Today, data warehouses
are a critical component of modern businesses, enabling them to
make data-driven decisions and stay competitive in a rapidly
changing market.

During this pivotal transition in the world of data management,

numerous scientists and experts made significant contributions to the
field. Notable among them are Bill Inmon, revered as the originator of
the data warehouse concept, which focuses on a single source of truth
for reporting and analysis; Ralph Kimball, a renowned data
warehousing expert who introduced dimensional modeling, which
emphasizes optimized data modeling for reporting, star schemas, and
fact tables; and Dan Linstedt, who invented the data vault modeling
approach, which combines elements of Inmon and Kimball’s
methodologies and is tailored for handling substantial data volumes
and historical reporting. In addition, Claudia Imhoff, a business
intelligence and data warehousing expert, founded the Boulder BI Brain
Trust, offering thought leadership; Barry Devlin pioneered the business
data warehouse concept, which highlights business metadata’s
importance and aligns data warehousing with business objectives; and,
lastly, Jim Gray, a computer scientist and database researcher, who
contributed significantly by introducing the data cube, a
multidimensional database structure for enhanced analysis and
reporting. In conclusion, these luminaries represent just a fraction of
the visionary minds that shaped modern data warehousing,
empowering businesses to harness data for informed decision-making
in a dynamic market landscape.

Data Warehouse Concepts

In today’s world, businesses collect more data than ever before. This
data can come from a variety of sources, such as customer transactions,
social media, and Internet of Things (IoT) devices. However, collecting
data is only the first step; to truly unlock the value of this data,
businesses must be able to analyze and report on it. This is where the
data warehouse comes in. The following are aspects of the data
warehouse:
A data warehouse is a large, centralized repository of data that is
optimized for reporting and analysis. The data warehouse is
designed to handle large volumes of data from multiple sources, and
to provide a single source of truth for reporting and analytics. It is a
critical component of modern business intelligence, enabling
businesses to make data-driven decisions and stay competitive in a
rapidly changing market.
Data warehouses use specialized technologies, such as extract,
transform, load (ETL) processes, to extract data from multiple
sources, transform it into a common format, and load it into the data
warehouse. This allows businesses to bring together data from
disparate sources and create a single, unified view of the data.
Data warehouses also use specialized tools for querying and
reporting, such as online analytical processing (OLAP), which allows
users to analyze data across multiple dimensions, and data mining,
which uses statistical and machine learning techniques to identify
patterns and relationships in the data.
 One of the key features of the data warehouse is its ability to handle
historical data. Traditional transactional databases are optimized for
handling current data, but they are not well suited to handling large
volumes of historical data. Data warehouses, however, are optimized
for handling large volumes of historical data, which is critical for
trend analysis and forecasting.
In addition, data warehouses are designed to be easy to use for
business users. They use specialized reporting tools that allow users
to create custom reports and dashboards, and to drill down into the
data to gain deeper insights. This makes it easy for business users to
access and analyze the data they need to make informed decisions.
There are several common concepts in data warehouses that are
essential to understanding their architecture. Here are some of the
most important concepts:
Data Sources: A data warehouse collects data from a variety of
sources, such as transactional databases, external data sources, and
flat files. Data is extracted from these sources and transformed into a
standardized format before being loaded into the data warehouse.
ETL (Extract, Transform, Load): This is the process used to collect
data from various sources and prepare it for analysis in the data
warehouse. During this process, data is extracted from the source
systems, transformed into a common format, and loaded into the
data warehouse.
Data Marts: A data mart is a subset of a data warehouse that is
designed to meet the needs of a particular department or group
within an organization. Data marts are typically organized around
specific business processes or functions, such as sales or marketing.
Data Modeling: In the field of data warehousing, there are two main
approaches to modeling data: tabular modeling and dimensional
modeling. Tabular modeling is a relational approach to data
modeling, which means it organizes data into tables with rows and
columns. Dimensional modeling involves organizing data around
dimensions (such as time, product, or location) and measures (such
as sales revenue or customer count) and using a star or snowflake
schema to represent the data.
OLAP (Online Analytical Processing): OLAP is a set of tools and
techniques used to analyze data in a data warehouse. OLAP tools
 allow users to slice and dice data along different dimensions and to
drill down into the data to gain deeper insights.
Data Mining: Data mining is the process of analyzing large datasets
to identify patterns, trends, and relationships in the data. This
technique uses statistical and machine learning algorithms to
discover insights and make predictions based on the data.
Metadata: Metadata is data about the data in a data warehouse. It
provides information about the source, structure, and meaning of the
data in the warehouse, and is essential for ensuring that the data is
accurate and meaningful.

Data Sources (Data Format and Common Sources)

In a data warehouse, data source refers to any system or application
that provides data to the data warehouse. A data source can be any type
of system or application that generates data, such as a transactional
system, a customer relationship management (CRM) application, or an
enterprise resource planning (ERP) system.
The data from these sources is extracted and transformed before it
is loaded into the data warehouse. This process involves cleaning,
standardizing, and consolidating the data to ensure that it is accurate,
consistent, and reliable. Once the data has been transformed, it is then
loaded into the data warehouse for storage and analysis.
In some cases, data sources may be connected to the data
warehouse using extract, transform, and load (ETL) processes, while in
other cases, they may be connected using other data integration
methods, such as data replication, data federation, or data
virtualization.

Note Data sources are a critical component of a data warehouse, as

they provide the data that is needed to support business intelligence
and analytics. By consolidating data from multiple sources into a
single location, a data warehouse enables organizations to gain
insights into their business operations and make more-informed
decisions.
There are various types and formats of data sources that can be used in
a data warehouse. Here are some examples:
Relational databases: A common data source for a data warehouse
is a relational database, such as Oracle, Microsoft SQL Server, or
MySQL. These databases store data in tables with defined schemas
and can be queried using SQL.
Flat files: Data can also be sourced from flat files, such as CSV files,
Parquet, Excel, or any other formatted text files. These files typically
have a delimited format with columns and rows.
Cloud storage services: Cloud storage services, such as Amazon S3
or Azure Data Lake Storage, can also be used as a data source for a
data warehouse. These services can store data in a structured or
unstructured format and can be accessed through APIs.
NoSQL databases: NoSQL databases, such as MongoDB or
Cassandra, can be used as data sources for data warehouses. These
databases are designed to handle large volumes of unstructured data
and can be queried using NoSQL query languages.
Real-time data sources: Real-time data sources, such as message
queues or event streams, can be used to stream data into a data
warehouse in real-time. This type of data source is often used for
applications that require up-to-date data.
APIs: APIs can also be used as a data source, providing access to data
from third-party applications or web services.
Format of the data coming from multiple sources can also vary
depending on the type of data. For example, data can be structured or
unstructured, semi-structured, such as JSON or XML. The data format
needs to be considered when designing the data warehouse schema
and the ETL processes. It is important to ensure that the data is
properly transformed and loaded into the data warehouse in a format
that is usable for analysis.
Data can flow to the data warehouse through different systems,
some of the most used of which include the following:
Transactional databases: Transactional databases are typically the
primary source of data for a data warehouse. These databases
capture and store business data generated by various systems, such
as sales, finance, and operations.
 ERP systems: Enterprise resource planning (ERP) systems are used
by many organizations to manage their business processes. ERP
systems can provide a wealth of data that can be used in a data
warehouse, including information on customer orders, inventory, and
financial transactions.
CRM systems: Customer relationship management (CRM) systems
provide data on customer interactions that can be used to support
business analytics and decision-making.
Legacy systems: Legacy systems are often used to store important
historical data that needs to be incorporated into the data
warehouse. This data may be stored in a variety of formats, including
flat files or proprietary databases.
Cloud-based systems: Cloud-based systems, such as software-as-a-
service (SaaS) applications, are becoming increasingly popular as
data sources for data warehouses. These systems can provide access
to a variety of data, including customer behavior, website traffic, and
sales data.
Social media: Social media platforms are another source of data that
can be used in a data warehouse. This data can be used to gain
insights into customer behavior, sentiment analysis, and brand
reputation.
One effective approach for documenting data-related artifacts, such
as data sources and data flows, is using data dictionaries and data
catalogs. These tools can capture relevant information about data
elements, including their structure and meaning, as well as provide more
comprehensive details about data sources, flows, lineage, and ownership.
By leveraging these tools, implementation teams and data operations
teams can gain a better understanding of this information, leading to
improved data quality, consistency, and collaboration across various
teams and departments within an organization.

Note When categorizing data into structured or unstructured

sources, you’ll find that older systems like transactional, ERP, CRM,
and legacy tend to have well-organized and -classified data
compared to that sourced from cloud-based systems or social media.
It’s not entirely accurate to say that all data from cloud platforms
and website analytics activities are unstructured, but analyzing such
 data requires additional computing power to derive significant
insights. With the adoption of cloud computing, organizations are
increasingly storing unstructured data.

ETL (Extract, Transform, Load)

ETL stands for extract, transform, load. It is a process used to move data
from one or more source systems, transform the data to fit business
needs, and load the data into a target system, such as a data warehouse.
The ETL process is an essential component of a data warehouse, as
it enables organizations to consolidate and integrate data from multiple
sources into a single, unified view of their business operations. Here is
a brief overview of the ETL process:
Extract: The first step in the ETL process is to extract the data from
the source systems. This can be done using various methods, such as
APIs, file transfers, or direct database connections.
Transform: Once the data has been extracted, it needs to be
transformed to fit the needs of the data warehouse. This may involve
cleaning the data, consolidating duplicate records, converting data
types, or applying business rules and calculations.
Load: After the data has been transformed, it is loaded into the target
system, such as a data warehouse. This can be done using various
methods, such as bulk inserts, incremental updates, or real-time
streaming.
The ETL process can be complex and time-consuming, particularly
for large datasets or complex data models. However, modern ETL tools
and technologies, such as cloud-based data integration platforms, have
made the process more efficient and scalable.

A well-designed ETL process is critical to the success of a data

warehouse, as it ensures that the data is accurate, consistent, and
reliable. By providing a unified view of business data, a data
warehouse enables organizations to gain insights into their
operations, identify trends and patterns, and make more informed
decisions.
 There are many ETL (extract, transform, load) software tools
available, both commercial and open source. Here are some examples:
Informatica PowerCenter: Informatica PowerCenter is a popular
ETL tool that offers a wide range of data integration and
transformation features, including data profiling, data quality, and
metadata management.
Microsoft SQL Server Integration Services (SSIS): SSIS is a
powerful ETL tool that is part of the Microsoft SQL Server suite. It
provides a wide range of data integration and transformation
features, including data cleansing, data aggregation, and data
enrichment.
Talend Open Studio: Talend Open Studio is an open source ETL tool
that offers a broad range of data integration and transformation
features, including support for Big Data platforms like Hadoop and
Spark.
IBM InfoSphere DataStage: IBM InfoSphere DataStage is a
comprehensive ETL tool that offers advanced data integration and
transformation features, including support for real-time data
processing and complex data structures.
Oracle Data Integrator (ODI): ODI is a powerful ETL tool that offers
a broad range of data integration and transformation features,
including support for Big Data and cloud platforms.
Apache NiFi: Apache NiFi is an open-source data integration and
transformation tool that provides a flexible, web-based interface for
designing and executing data workflows. It supports a wide range of
data sources and destinations and can be used for real-time data
processing and streaming.
Azure Data Factory: Azure Data Factory is a cloud-based data
integration service offered by Microsoft Azure. It allows you to
create, schedule, and manage data integration pipelines. It provides
90+ built-in connectors for seamless data integration from various
sources, including on-premises data stores. Azure Data Factory
enables easy design, deployment, and monitoring of data integration
pipelines through an intuitive graphical interface or code. This helps
you manage your data more efficiently, reduce operational costs, and
accelerate business insights.
 AWS Glue: AWS Glue is a serverless ETL service by Amazon Web
Services that automates time-consuming ETL tasks for preparing
data for analytics, machine learning, and application development. It
enables you to create data transformation workflows that can
extract, transform, and load data from various sources into data
lakes, warehouses, and other stores. You can use pre-built
transformations or custom code with Python or Scala for ETL. AWS
Glue is based on Apache Spark, allowing for fast and scalable data
processing, and integrates with other AWS services like Amazon S3,
Amazon Redshift, and Amazon RDS. This service simplifies the ETL
process and frees up time for analyzing data to make informed
business decisions.
These are just a few examples of the many ETL tools available for data
integration and transformation. The choice of ETL tool depends on the
specific needs and requirements of the organization, as well as the
available resources and budget.

ETL and ELT

ETL (extract, transform, load) and ELT (extract, load, transform) are
both data integration techniques that are used to transfer data from
source systems to target systems. The main difference between ETL
and ELT lies in the order of the data processing steps:
In ETL, data is first extracted from source systems, then transformed
into the desired format, and finally loaded into the target system.
This means that the transformation step takes place outside of the
target system and can involve complex data manipulation and
cleansing.
In ELT, data is first extracted from source systems and loaded into the
target system, and then transformed within the target system. This
means that the transformation step takes place within the target
system, using its processing power and capabilities to transform the
data.
The choice between ETL and ELT depends on several factors,
including the complexity and size of the data being processed, the
capabilities of the target system, and the processing and storage
costs. ETL is more suitable for complex data transformation, while
 ELT is more suitable for large data volumes and systems with
advanced processing capabilities.

Data Mart
A data mart is a subset of a larger data warehouse and is designed to
serve the needs of a particular business unit or department. Data marts
are used to provide targeted, specific information to end users, allowing
them to make better, data-driven decisions.
A data mart is typically designed to store data that is relevant to a
specific business area or function, such as sales, marketing, or finance.
Data marts can be created using data from the larger data warehouse,
or they can be created as standalone systems that are populated with
data from various sources.

Data Mart Architecture

The architecture of a data mart can vary depending on the specific
needs of the business unit or department it serves. However, some
common elements are typically found in a data mart architecture:
Data sources: The data sources for a data mart can come from
various systems and applications, such as transactional systems,
operational databases, or other data warehouses.
ETL process: The ETL process is used to extract data from the source
systems, transform it to meet the needs of the data mart, and load it
into the data mart.
Data mart database: The data mart database is the repository that
stores the data for the specific business unit or department. It is
typically designed to be optimized for the types of queries and
analyses that are performed by the end users. (In modern day, in
some cases this may be a temporarily transformed datastore
refreshed periodically with no history.)
Business intelligence tools: Business intelligence (BI) tools are
used to analyze the data in the data mart and provide reports,
dashboards, and other visualizations to end users.

Advantages of Data Marts

Data marts are a crucial component of modern data management and
analytics strategies, offering several advantages that organizations can
leverage to drive informed decision-making and enhance their
competitive edge. These streamlined subsets of data warehouses are
designed to cater to specific business units or departments, providing a
focused and efficient approach to data access and analysis. Some of the
key advantages of a data mart, with examples, are as follows:
Targeted data: Data marts provide a subset of the larger data
warehouse that is specifically designed to meet the needs of a
particular business unit or department. This makes it easier for end
users to find the data they need and make well-informed decisions.
Improved performance: Data marts are designed to be optimized
for the types of queries and analyses that are performed by the end
users. This can improve query performance and reduce the time it
takes to access and analyze data.
Reduced complexity: By focusing on a specific business area or
function, data marts can simplify the data architecture and make it
easier to manage and maintain.

Examples of Data Marts

An organization comprises various departments, including Sales,
Marketing, Finance, etc., each with distinct analytics needs and specific
information consumption requirements. Therefore, to effectively
address these diverse needs, different datamarts are essential.
Sales data mart: A sales data mart might be used to provide
information on customer orders, product sales, and revenue by
region or by salesperson.
Marketing data mart: A marketing data mart might be used to
provide information on customer demographics, campaign
performance, and customer acquisition costs.
Finance data mart: A finance data mart might be used to provide
information on budgeting, forecasting, and financial reporting.
In conclusion, data marts are an essential component of a data
warehouse, providing targeted, specific information to end users and
enabling them to make better, data-driven decisions. By designing data
marts to meet the specific needs of individual business units or
departments, organizations can improve performance, reduce complexity,
and achieve their business objectives more effectively.

Data Modeling
In the field of data warehousing, there are two main approaches to
modeling data: tabular modeling and dimensional modeling. Both
approaches have their strengths and weaknesses, and choosing the
right one for your specific needs is crucial to building an effective data
warehouse.

Tabular Modeling
Tabular modeling is a relational approach to data modeling, which
means it organizes data into tables with rows and columns. This
approach is well suited to handling large volumes of transactional data
and is often used in OLTP (online transaction processing) systems. In a
tabular model, data is organized into a normalized schema, where each
fact is stored in a separate table, and the relationships between the
tables are established through primary and foreign keys.
The advantages of tabular modeling include its simplicity, ease of use,
and flexibility. Because data is organized into a normalized schema, it is
easier to add or modify data fields, and it supports complex queries and
reporting. However, tabular models can become more complex to query
and maintain as the number of tables and relationships increases, and it
can be slower to process queries on large datasets.

Dimensional Modeling
Dimensional modeling is a more specialized approach that is optimized
for OLAP (online analytical processing) systems. Dimensional models
organize data into a star or snowflake schema, with a fact table at the
center and several dimension tables surrounding it. The fact table
contains the measures (i.e., numerical data) that are being analyzed,
while the dimension tables provide the context (i.e., descriptive data)
for the measures.
Also, dimensional modeling is optimized for query performance,
making it well suited for OLAP and especially reporting systems.
Because data is organized into a star or snowflake schema, it is easier to
perform aggregations and analyses, and it is faster to query large
datasets. Dimensional models are also easier to understand and
maintain, making them more accessible to business users. However,
dimensional models can be less flexible and more complex to set up,
and they may not perform as well with transactional data.
In conclusion, both tabular and dimensional modeling have their
places in data warehousing, and the choice between them depends on
the specific needs of your organization. Tabular modeling is more
suited to handling large volumes of transactional data, while
dimensional modeling is optimized for OLAP systems and faster query
performance.

In modern warehousing with data and delta lakes, tabular models

structured in facts and dimensions are still effective. There are
multiple tools available to balance between extreme normalization
and extreme classification. While tabular models provide a simpler
structure and facilitate querying of data, dimensional models make it
more ready for analytics and reporting needs.

Understanding Dimensional Modeling in Brief

In data warehousing, dimensions, facts, and measures are essential
concepts that are used to organize and analyze data. A dimension is a
category of data that provides context for a fact, while a fact is a value
that describes a specific event or activity. Measures are numerical
values that quantify facts.
Dimensions
A dimension is a grouping or category of data that provides context for
a fact. Dimensions can be thought of as the “who, what, when, where,
and why” of a dataset.
For example, a sales dataset might include dimensions such as date,
product, customer, and location. Each of these dimensions provides
additional information about the sales data and helps to contextualize
it.
Dimensions can be further classified into the following types:
Degenerate Dimension: A degenerate dimension is a dimension
that is not stored in a separate dimension table but is included in the
fact table.
Conformed Dimension: A conformed dimension is a dimension that
is used in multiple fact tables in the same data warehouse. It is
designed to ensure the consistency and accuracy of data across the
different fact tables. For example, let’s consider a retail company that
sells products through multiple channels, such as brick and mortar
stores, online stores, and mobile apps. The company has a data
warehouse that stores data about sales, inventory, and customer
behavior.
In this scenario, the “customer” dimension is a good example of a
conformed dimension. The customer dimension contains attributes
such as customer name, address, age, gender, and purchase history.
This dimension is used in multiple fact tables, such as sales fact table,
customer behavior fact table, and inventory fact table.

By using a conformed dimension for customer data, the data

warehouse ensures that all the information related to customers is
consistent and accurate across all the fact tables. It also simplifies the
data model and reduces the risk of data inconsistencies and errors.
Another advantage of using conformed dimensions is that they can
be reused across multiple data marts or data domains. This means
that the same customer dimension can be used for sales analysis,
customer behavior analysis, and inventory management analysis
without duplicating the data or creating a separate dimension for
each fact table.
Junk Dimension: A junk dimension is a collection of flags and
indicators that are not related to any specific dimension. These flags
and indicators are grouped together into a single dimension table to
simplify the data model and improve query performance.
Junk dimensions are used when you have many low-cardinality
flags that are not related to any specific dimension, and it’s not
worthwhile to create a separate dimension for each flag.
The name junk comes from the fact that the dimension contains
seemingly unrelated attributes that don’t fit neatly into any other
dimension. Examples of attributes that can be included in a junk
dimension are as follows:
Boolean indicators: “yes” or “no” flags that describe the presence
or absence of a particular condition
Flags: “on” or “off” indicators that specify the status of a particular
process or workflow
Codes: short codes that describe the result of a particular event or
transaction
Dates: dates or timestamps that indicate when a particular event
occurred
By consolidating these attributes into a single dimension table,
you can simplify the data model and improve query performance.
The junk dimension table acts as a bridge table between the fact
table and the other dimensions in the data model.
For example, let’s consider an e-commerce website that sells
products online. The website has a fact table that records the sales
transactions and several dimensions, such as product, customer, and
time. The fact table contains several low-cardinality flags, such as
“shipped,” “cancelled,” “returned,” and “gift-wrapped,” which are not
related to any specific dimension. Instead of creating a separate
dimension table for each flag, these flags can be consolidated into a
junk dimension table. The junk dimension table will have a record for
each unique combination of these flags, and the fact table will
reference the junk dimension table using a foreign key.
Junk dimensions can be an effective way to simplify a data model
and reduce the number of dimension tables required in a data
warehouse. However, care should be taken to ensure that the attributes
in the junk dimension are truly unrelated and do not belong in any
other dimension. Otherwise, the use of a junk dimension can lead to
data-quality issues and analysis errors.
Role-Playing Dimension: A role-playing dimension is a dimension
that is used in different ways in the same fact table. For example, a
date dimension can be used to analyze sales by order date, ship
date, or delivery date. Role-playing dimensions are useful when the
same dimension table is used in different contexts with different
meanings. By reusing the same dimension table, the data model can
be simplified, and the data can be more easily analyzed and
understood. However, it’s important to ensure that the meaning of
each use of the dimension table is clearly defined to avoid confusion
and errors in data analysis.
Slowly Changing Dimension (SCD): A slowly changing dimension is
a dimension that changes slowly over time. SCDs are classified into
six types:
Type 1 SCD: In a Type 1 SCD, the old values are simply overwritten
with new values when a change occurs. This approach is suitable
for dimensions where historical information is not required.
Suppose you have a price master table that contains
information about products such as name, price, and details. If the
price of a product changes, you might simply update the price in
the price master table without keeping track of the historical price.
Type 2 SCD: In a Type 2 SCD, a new row is added to the dimension
table when a change occurs, and the old row is marked as inactive.
This approach is suitable for dimensions where historical
information is required.
Continuing with the price master table example, if you want to
keep track of the historical price of each product, you might create
a new row for each price change. For example, you might add a
new row with a new product version number and a new price
whenever the price changes. This way, you can keep track of the
historical prices of each product.
Type 3 SCD: In a Type 3 SCD, a limited amount of historical
information is maintained by adding columns to the dimension
table to store previous values. This approach is suitable for
dimensions where only a limited amount of historical information
is required.
Suppose you have an employee table that contains information
about employees such as name, address, and salary. If an employee
gets a promotion, you might add a new column to the table to store
the new job title. You would only store the most recent job title,
and not keep track of historical job titles.
Type 4 SCD: Create a separate table to store historical data. This
type of SCD is useful when historical data needs to be stored
separately for performance reasons.
 Suppose you have a customer table that contains information
about customers such as name, address, and phone number. If you
want to keep track of historical addresses, you might create a new
table to store the historical addresses. The new table would
contain the customer ID, the old address, and the date the address
was changed.
Type 5 SCD: Combine SCD Types 1 and 2 by adding an additional
column to track the current and historical values. This type of SCD
can be useful when there are a large number of historical changes,
but only the current value is needed for most queries.
Continuing with the customer table example, if you want to
keep track of the current and historical phone numbers for each
customer, you might create a new column in the customer table to
store the current phone number, and then create a new row in a
separate phone number table for each phone number change. The
phone number table would contain the customer ID, the phone
number, the start date, and the end date.
Type 6 SCD: Combine SCD Types 2 and 3 by adding an additional
column to track the current and previous values. This type of SCD
is useful when historical data is important, but only the current
and previous values are needed for most queries.
Suppose you have a product table that contains information
about products such as name, price, and description. If the price of
a product changes, you might create a new row for the new
product version and store the new price in that row. You might also
add a new column to the product table to store the previous price.
This way, you can easily access the current and previous prices of
each product.
Time Dimension: A time dimension is a special type of dimension
that is used to track time-related data. It provides a way to group and
filter data based on time periods such as hours, days, weeks, months,
and years.
Hierarchical Dimension: A hierarchical dimension is a dimension
that has a parent–child relationship between its attributes. For
example, a product dimension can have a hierarchy that includes
product category, sub-category, and product.
Random documents with unrelated
content Scribd suggests to you:
depend on the proportion of the power to the resistance; in short,
they had no clear idea of accelerating force. This defect runs through
all their mechanical speculations, and renders them entirely
valueless.

We may exemplify the same confusion of thought on mechanical

subjects in writers of a less technical character. Thus, if men had any
distinct idea of mechanical action, they could not have accepted for a
moment the fable of the Echineis or Remora, a little fish which was
said to be able to stop a large ship merely by sticking to it. 1 Lucan
refers to this legend in a poetical manner, and notices this creature
only in bringing together a collection of monstrosities; but Pliny
relates the tale gravely, and moralizes upon it after his manner.
“What,” he cries, 2 “is more violent than the sea and the winds? what
a greater work of art than a ship? Yet one little fish (the Echineis) can
hold back all these when they all strain the same way. The winds
may 191 blow, the waves may rage; but this small creature controls
their fury, and stops a vessel, when chains and anchors would not
hold it: and this it does, not by hard labor, but merely by adhering to
it. Alas, for human vanity! when the turreted ships which man has
built, that he may fight from castle-walls, at sea as well as at land,
are held captive and motionless by a fish a foot and a half long! Such
a fish is said to have stopped the admiral’s ship at the battle of
Actium, and compelled Antony to go into another. And in our own
memory, one of these animals held fast the ship of Caius, the
emperor, when he was sailing from Astura to Antium. The stopping
of this ship, when all the rest of the fleet went on, caused surprise;
but this did not last long, for some of the men jumped into the water
to look for the fish, and found it sticking to the rudder; they showed it
to Caius, who was indignant that this animal should interpose its
prohibition to his progress, when impelled by four hundred rowers. It
was like a slug; and had no power, after it was taken into the ship.”
1Lucan is describing one of the poetical compounds produced in
incantations.
Huc quicquid fœtu genuit Natura sinistro
Miscetur: non spuma canum quibus unda timori est,
Viscera non lyncis, non duræ nodus hyænæ
Defuit, et cervi pasti serpente medullæ;
Non puppes retinens, Euro tendente rudentes
In mediis Echineis aquis, oculique draconum.
Etc. Pharsalia, vi. 670.

2 Plin. Hist. N. xxxii. 5.

A very little advance in the power of thinking clearly on the force

which it exerted in pulling, would have enabled the Romans to see
that the ship and its rowers must pull the adhering fish by the hold
the oars had upon the water; and that, except the fish had a hold
equally strong on some external body, it could not resist this force.

3. Indistinctness of Ideas shown in Architecture.—Perhaps it may

serve to illustrate still further the extent to which, under the Roman
empire, men’s notions of mechanical relations became faint,
wavered, and disappeared, if we observe the change which took
place in architecture. All architecture, to possess genuine beauty,
must be mechanically consistent. The decorative members must
represent a structure which has in it a principle of support and
stability. Thus the Grecian colonnade was a straight horizontal beam,
resting on vertical props; and the pediment imitated a frame like a
roof, where oppositely inclined beams support each other. These
forms of building were, therefore, proper models of art, because they
implied supporting forces. But to be content with colonnades and
pediments, which, though they imitated the forms of the Grecian
ones, were destitute of their mechanical truth, belonged to the
decline of art; and showed that men had lost the idea of force, and
retained only that of shape. Yet this was what the architects of the
Roman empire did. Under their hands, the pediment was severed at
its vertex, and divided into separate halves, so that it was no longer
a mechanical possibility. The entablature no longer lay straight from
pillar to pillar, but, projecting over each 192 column, turned back to
the wall, and adhered to it in the intervening space. The splendid
remains of Palmyra, Balbec, Petra, exhibit endless examples of this
kind of perverse inventiveness; and show us, very instructively, how
the decay of art and of science alike accompany this indistinctness
of ideas which we are now endeavoring to illustrate.

4. Indistinctness of Ideas in Astronomy.—Returning to the

sciences, it may be supposed, at first sight, that, with regard to
astronomy, we have not the same ground for charging the stationary
period with indistinctness of ideas on that subject, since they were
able to acquire and verify, and, in some measure, to apply, the
doctrines previously established. And, undoubtedly, it must be
confessed that men’s notions of the relations of space and number
are never very indistinct. It appears to be impossible for these chains
of elementary perception ever to be much entangled. The later
Greeks, the Arabians, and the earliest modern astronomers, must
have conceived the hypotheses of the Ptolemaic system with
tolerable completeness. And yet, we may assert, that during the
stationary period, men did not possess the notions, even of space
and number, in that vivid and vigorous manner which enables them
to discover new truths. If they had perceived distinctly that the
astronomical theorist had merely to do with relative motions, they
must have been led to see the possibility, at least, of the Copernican
system; as the Greeks, at an earlier period, had already perceived it.
We find no trace of this. Indeed, the mode in which the Arabian
mathematicians present the solutions of their problems, does not
indicate that clear apprehension of the relations of space, and that
delight in the contemplation of them, which the Greek geometrical
speculations imply. The Arabs are in the habit of giving conclusions
without demonstrations, precepts without the investigations by which
they are obtained; as if their main object were practical rather than
speculative,—the calculation of results rather than the exposition of
theory. Delambre 3 has been obliged to exercise great ingenuity, in
order to discover the method by which Ibn Iounis proved his solution
of certain difficult problems.
3 Delamb. M. A. p. 125–8.

5. Indistinctness of Ideas shown by Skeptics.—The same

unsteadiness of ideas which prevents men from obtaining clear
views, and steady and just convictions, on special subjects, may
lead them to despair of or deny the possibility of acquiring certainty
at all, and may thus make them skeptics with regard to all
knowledge. Such skeptics 193 are themselves men of indistinct
views, for they could not otherwise avoid assenting to the
demonstrated truths of science; and, so far as they may be taken as
specimens of their contemporaries, they prove that indistinct ideas
prevail in the age in which they appear. In the stationary period,
moreover, the indefinite speculations and unprofitable subtleties of
the schools might further impel a man of bold and acute mind to this
universal skepticism, because they offered nothing which could fix or
satisfy him. And thus the skeptical spirit may deserve our notice as
indicative of the defects of a system of doctrine too feeble in
demonstration to control such resistance.
 The most remarkable of these philosophical skeptics is Sextus
Empiricus; so called, from his belonging to that medical sect which
was termed the empirical, in contradistinction to the rational and
methodical sects. His works contain a series of treatises, directed
against all the divisions of the science of his time. He has chapters
against the Geometers, against the Arithmeticians, against the
Astrologers, against the Musicians, as well as against Grammarians,
Rhetoricians, and Logicians; and, in short, as a modern writer has
said, his skepticism is employed as a sort of frame-work which
embraces an encyclopedical view of human knowledge. It must be
stated, however, that his objections are rather to the metaphysical
grounds, than to the details of the sciences; he rather denies the
possibility of speculative truth in general, than the experimental
truths which had been then obtained. Thus his objections to
geometry and arithmetic are founded on abstract cavils concerning
the nature of points, letters, unities, &c. And when he comes to
speak against astrology, he says, “I am not going to consider that
perfect science which rests upon geometry and arithmetic; for I have
already shown the weakness of those sciences: nor that faculty of
prediction (of the motions of the heavens) which belongs to the
pupils of Eudoxus, and Hipparchus, and the rest, which some call
Astronomy; for that is an observation of phenomena, like agriculture
or navigation: but against the Art of Prediction from the time of birth,
which the Chaldeans exercise.” Sextus, therefore, though a skeptic
by profession, was not insensible to the difference between
experimental knowledge and mystical dogmas, though even the
former had nothing which excited his admiration.

The skepticism which denies the evidence of the truths of which

the best established physical sciences consist, must necessarily
involve a very indistinct apprehension of those truths; for such truths,
properly exhibited, contain their own evidence, and are the best
antidote 194 to this skepticism. But an incredulity or contempt
towards the asserted truths of physical science may arise also from
the attention being mainly directed to the certainty and importance of
religious truths. A veneration for revealed religion may thus assume
the aspect of a skepticism with regard to natural knowledge. Such
appears to be the case with Algazel or Algezeli, who is adduced by
Degerando 4 as an example of an Arabian skeptic. He was a
celebrated teacher at Bagdad in the eleventh century, and he
declared himself the enemy, not only of the mixed Peripatetic and
Platonic philosophy of the time, but of Aristotle himself. His work
entitled The Destructions of the Philosophers, is known to us by the
refutation of it which Averrhoes published, under the title of
Destruction of Algazel’s Destructions of the Philosophers. It appears
that he contested the fundamental principles both of the Platonic and
of the Aristotelian schools, and denied the possibility of a known
connection between cause and effect; thus making a prelude, says
Degerando, to the celebrated argumentation of Hume.
4 Degerando, Hist. Comp. de Systèmes, iv. 224.

[2d Ed.] Since the publication of my first edition, an account of

Algazel or Algazzali and his works has been published under the title
of Essai sur les Ecoles Philosophiques chez les Arabes, et
notamment sur la Doctrine d’Algazzali, par August Schmölders.
Paris. 1842. From this book it appears that Degerando’s account of
Algazzali is correct, when he says 5 that “his skepticism seems to
have essentially for its object to destroy all systems of merely
rational theology, in order to open an indefinite career, not only to
faith guided by revelation, but also to the free exaltation of a mystical
enthusiasm.” It is remarked by Dr. Schmölders, following M. de
Hammer-Purgstall, that the title of the work referred to in the text
ought rather to be Mutual Refutation of the Philosophers: and that its
object is to show that Philosophy consists of a mass of systems,
each of which overturns the others. The work of Algazzali which Dr.
Schmölders has published, On the Errors of Sects, &c., contains a
kind of autobiographical account of the way in which the author was
led to his views. He does not reject the truths of science, but he
condemns the mental habits which are caused by laying too much
stress upon science. Religious men, he says, are, by such a course,
led to reject all science, even what relates to eclipses of the moon
and sun; and men of science are led to hate religion. 6
5 Hist. Comp. iv. p. 227.

6 Essai, p. 33.

195 6. Neglect of Physical Reasoning in Christendom.—If the

Arabians, who, during the ages of which we are speaking, were the
most eminent cultivators of science, entertained only such
comparatively feeble and servile notions of its doctrines, it will easily
be supposed, that in the Christendom of that period, where physical
knowledge was comparatively neglected, there was still less
distinctness and vividness in the prevalent ideas on such subjects.
Indeed, during a considerable period of the history of the Christian
Church, and by many of its principal authorities, the study of natural
philosophy was not only disregarded but discommended. The great
practical doctrines which were presented to men’s minds, and the
serious tasks, of the regulation of the will and affections, which
religion impressed upon them, made inquiries of mere curiosity seem
to be a reprehensible misapplication of human powers; and many of
the fathers of the Church revived, in a still more peremptory form, the
opinion of Socrates, that the only valuable philosophy is that which
teaches us our moral duties and religious hopes. 7 Thus Eusebius
says, 8 “It is not through ignorance of the things admired by them, but
through contempt of their useless labor, that we think little of these
matters, turning our souls to the exercise of better things.” When the
thoughts were thus intentionally averted from those ideas which
natural philosophy involves, the ideas inevitably became very
indistinct in their minds; and they could not conceive that any other
persons could find, on such subjects, grounds of clear conviction and
certainty. They held the whole of their philosophy to be, as
Lactantius 9 asserts it to be, “empty and false.” “To search,” says he,
“for the causes of natural things; to inquire whether the sun be as
large as he seems, whether the moon is convex or concave, whether
the stars are fixed in the sky or float freely in the air; of what size and
of what material are the heavens; whether they be at rest or in
motion; what is the magnitude of the earth; on what foundations it is
suspended and balanced;—to dispute and conjecture on such
matters, is just as if we chose to discuss what we think of a city in a
remote country, of which we never heard but the name.” It is
impossible to express more forcibly that absence of any definite
notions on physical subjects which led to this tone of thought.
7 Brucker, iii. 317.

8 Præp. Ev. xv. 61.

9 Inst. 1. iii. init.

7. Question of Antipodes.—With such habits of thought, we are not

to be surprised if the relations resulting from the best established
theories were apprehended in an imperfect and incongruous
manner. 196 We have some remarkable examples of this; and a very
notable one is the celebrated question of the existence of Antipodes,
or persons inhabiting the opposite side of the globe of the earth, and
consequently having the soles of their feet directly opposed to ours.
The doctrine of the globular form of the earth results, as we have
seen, by a geometrical necessity, from a clear conception of the
various points of knowledge which we obtain, bearing upon that
subject. This doctrine was held distinctly by the Greeks; it was
adopted by all astronomers, Arabian and European, who followed
them; and was, in fact, an inevitable part of every system of
astronomy which gave a consistent and intelligible representation of
phenomena. But those who did not call before their minds any
distinct representation at all, and who referred the whole question to
other relations than those of space, might still deny this doctrine; and
they did so. The existence of inhabitants on the opposite side of the
terraqueous globe, was a fact of which experience alone could teach
the truth or falsehood; but the religious relations, which extend alike
to all mankind, were supposed to give the Christian philosopher
grounds for deciding against the possibility of such a race of men.
Lactantius, 10 in the fourth century, argues this matter in a way very
illustrative of that impatience of such speculations, and consequent
confusion of thought, which we have mentioned. “Is it possible,” he
says, “that men can be so absurd as to believe that the crops and
trees on the other side of the earth hang downwards, and that men
there have their feet higher than their heads? If you ask of them how
they defend these monstrosities—how things do not fall away from
the earth on that side—they reply, that the nature of things is such
that heavy bodies tend towards the centre, like the spokes of a
wheel, while light bodies, as clouds, smoke, fire, tend from the centre
towards the heavens on all sides. Now I am really at a loss what to
say of those who, when they have once gone wrong, steadily
persevere in their folly, and defend one absurd opinion by another.” It
is obvious that so long as the writer refused to admit into his
thoughts the fundamental conception of their theory, he must needs
be at a loss what to say to their arguments without being on that
account in any degree convinced of their doctrines.
10 Inst. 1. iii. 23.

In the sixth century, indeed, in the reign of Justinian, we find a

writer (Cosmas Indicopleustes 11 ) who does not rest in this obscurity
of 197 representation; but in this case, the distinctness of the pictures
only serves to show his want of any clear conception as to what
suppositions would explain the phenomena. He describes the earth
as an oblong floor, surrounded by upright walls, and covered by a
vault, below which the heavenly bodies perform their revolutions,
going round a certain high mountain, which occupies the northern
parts of the earth, and makes night by intercepting the light of the
sun. In Augustin 12 (who flourished a. d. 400) the opinion is treated
on other grounds; and without denying the globular form of the earth,
it is asserted that there are no inhabitants on the opposite side,
because no such race is recorded by Scripture among the
descendants of Adam. 13 Considerations of the same kind operated
in the well-known instance of Virgil, Bishop of Salzburg, in the eighth
century. When he was reported to Boniface, Archbishop of Mentz, as
holding the existence of Antipodes, the prelate was shocked at the
assumption, as it seemed to him, of a world of human beings, out of
the reach of the conditions of salvation; and application was made to
Pope Zachary for a censure of the holder of this dangerous doctrine.
It does not, however, appear that this led to any severity; and the
story of the deposition of Virgil from his bishopric, which is circulated
by Kepler and by more modern writers, is undoubtedly altogether
false. The same scruples continued to prevail among Christian
writers to a later period; and Tostatus 14 notes the opinion of the
rotundity of the earth as an “unsafe” doctrine, only a few years
before Columbus visited the other hemisphere.
11 Montfaucon, Collectio Nova Patrum, t. ii. p. 113. Cosmas
Indicopleustes. Christianorum Opiniones de Mundo, sive
Topographia Christiana.

12 Civ. D. xvi. 9.

13 It appears, however, that scriptural arguments were found on

the other side. St. Jerome says (Comm. in Ezech. i. 6), speaking
of the two cherubims with four faces, seen by the prophet, and the
interpretation of the vision: “Alii vero qui philosophorum stultam
sequuntur sapientiam, duo hemispheria in duobus templi
cherubim, nos et antipodes, quasi supinos et cadentes homines
suspicantur.”

14 Montfauc. Patr. t. ii.

8. Intellectual Condition of the Religious Orders.—It must be

recollected, however, that though these were the views and tenets of
many religious writers, and though they may be taken as indications
of the prevalent and characteristic temper of the times of which we
speak, they never were universal. Such a confusion of thought
affects the minds of many persons, even in the most enlightened
times; and in what we call the Dark Ages, though clear views on
such subjects might be more rare, those who gave their minds to
science, entertained the true opinion of the figure of the earth. Thus
Boëthius 15 (in the sixth century) urges the smallness of the globe of
the earth, 198 compared with the heavens, as a reason to repress
our love of glory. This work, it will be recollected, was translated into
the Anglo-Saxon by our own Alfred. It was also commented on by
Bede, who, in what he says on this passage, assents to the doctrine,
and shows an acquaintance with Ptolemy and his commentators,
both Arabian and Greek. Gerbert, in the tenth century, went from
France to Spain to study astronomy with the Arabians, and soon
surpassed his masters. He is reported to have fabricated clocks, and
an astrolabe of peculiar construction. Gerbert afterwards (in the last
year of the first thousand from the birth of Christ) became pope, by
the name of Sylvester II. Among other cultivators of the sciences,
some of whom, from their proficiency, must have possessed with
considerable clearness and steadiness the elementary ideas on
which it depends, we may here mention, after Montucla, 16 Adelbold,
whose work On the Sphere was addressed to Pope Sylvester, and
whose geometrical reasonings are, according to Montucla, 17 vague
and chimerical; Hermann Contractus, a monk of St Gall, who, in
1050, published astronomical works; William of Hirsaugen, who
followed his example in 1080; Robert of Lorraine, who was made
Bishop of Hereford by William the Conqueror, in consequence of his
astronomical knowledge. In the next century, Adelhard Goth, an
Englishman, travelled among the Arabs for purposes of study, as
Gerbert had done in the preceding age; and on his return, translated
the Elements of Euclid, which he had brought from Spain or Egypt.
Robert Grostête, Bishop of Lincoln, was the author of an Epitome on
the Sphere; Roger Bacon, in his youth the contemporary of Robert,
and of his brother Adam Marsh, praises very highly their knowledge
in mathematics.
15 Boëthius, Cons. ii. pr. 7.

16 Mont. i. 502.

17 Ib. i. 503.
 “And here,” says the French historian of mathematics, whom I
have followed in the preceding relation, “it is impossible not to reflect
that all those men who, if they did not augment the treasure of the
sciences, at least served to transmit it, were monks, or had been
such originally. Convents were, during these stormy ages, the
asylum of sciences and letters. Without these religious men, who, in
the silence of their monasteries, occupied themselves in transcribing,
in studying, and in imitating the works of the ancients, well or ill,
those works would have perished; perhaps not one of them would
have come down to us. The thread which connects us with the
Greeks and Romans would have been snapt asunder; the precious
productions of 199 ancient literature would no more exist for us, than
the works, if any there were, published before the catastrophe that
annihilated that highly scientific nation, which, according to Bailly,
existed in remote ages in the centre of Tartary, or at the roots of
Caucasus. In the sciences we should have had all to create; and at
the moment when the human mind should have emerged from its
stupor and shaken off its slumbers, we should have been no more
advanced than the Greeks were after the taking of Troy.” He adds,
that this consideration inspires feelings towards the religious orders
very different from those which, when he wrote, were prevalent
among his countrymen.

Except so far as their religious opinions interfered, it was natural

that men who lived a life of quiet and study, and were necessarily in
a great measure removed from the absorbing and blinding interests
with which practical life occupies the thoughts, should cultivate
science more successfully than others, precisely because their ideas
on speculative subjects had time and opportunity to become clear
and steady. The studies which were cultivated under the name of the
Seven Liberal Arts, necessarily tended to favor this effect. The
Trivium, 18 indeed, which consisted of Grammar, Logic, and Rhetoric,
had no direct bearing upon those ideas with which physical science
is concerned; but the Quadrivium, Music, Arithmetic, Geometry,
Astronomy, could not be pursued with any attention, without a
corresponding improvement of the mind for the purposes of sound
knowledge. 19
18 Bruck. iii. 597.

19 Roger Bacon, in his Specula Mathematica, cap. i., says

“Harum scientiarum porta et clavis est mathematica, quam sancti
a principio mundi invenerunt, etc. Cujus negligentia jam per
triginta vel quadraginta annos destruxit totum studium Latinorum.”
I do not know on what occasion this neglect took place.

9. Popular Opinions.—That, even in the best intellects, something

was wanting to fit them for scientific progress and discovery, is
obvious from the fact that science was so long absolutely stationary.
And I have endeavored to show that one part of this deficiency was
the want of the requisite clearness and vigor of the fundamental
scientific ideas. If these were wanting, even in the most powerful and
most cultivated minds, we may easily conceive that still greater
confusion and obscurity prevailed in the common class of mankind.
They actually adopted the belief, however crude and inconsistent,
that the form of the earth and heavens really is what at any place it
appears to be; that the earth is flat, and the waters of the sky
sustained above a material floor, through which in showers they
descend. Yet the true doctrines of 200 astronomy appear to have had
some popular circulation. For instance, a French poem of the time of
Edward the Second, called Ymage du Monde, contains a metrical
account of the earth and heavens, according to the Ptolemaic views;
and in a manuscript of this poem, preserved in the library of the
University of Cambridge, there are representations, in accordance
with the text, of a spherical earth, with men standing upright upon it
on every side; and by way of illustrating the tendency of all things to
the centre, perforations of the earth, entirely through its mass, are
described and depicted; and figures are exhibited dropping balls
down each of these holes, so as to meet in the interior. And, as
bearing upon the perplexity which attends the motions of up and
down, when applied to the globular earth, and the change of the
direction of gravity which would occur in passing the centre, the
readers of Dante will recollect the extraordinary manner in which the
poet and his guide emerge from the bottom of the abyss; and the
explanation which Virgil imparts to him of what he there sees. After
they have crept through the aperture in which Lucifer is placed, the
poet says,

“Io levai gli occhi e credetti vedere

Lucifero com’ io l’ avea lasciato,
E vidile le gambe in su tenere.”
. . . . “Questi come è fitto
Si sottasopra!” . . . . .
“Quando mi volsi, tu passast’ il punto
Al qual si traggon d’ ogni parte i pesi.”
Inferno, xxxiv.

. . . . “I raised mine eyes,

Believing that I Lucifer should see
Where he was lately left, but saw him now
With legs held upward.” . . . .
“How standeth he in posture thus reversed?”
. . . . . . . . . . . . . .
“Thou wast on the other side so long as I
Descended; when I turned, thou didst o’erpass
That point to which from every part is dragged
All heavy substance.” Cary.
 This is more philosophical than Milton’s representation, in a more
scientific age, of Uriel sliding to the earth on a sunbeam, and sliding
back again, when the sun had sunk below the horizon.

. . . . . “Uriel to his charge

Returned on that bright beam whose point now raised,
Bore him slope downward to the sun, now fallen
Beneath the Azores.” Par. Lost, B. iv. 201

The philosophical notions of up and down are too much at

variance with the obvious suggestions of our senses, to be held
steadily and justly by minds undisciplined in science. Perhaps it was
some misunderstood statement of the curved surface of the ocean,
which gave rise to the tradition of there being a part of the sea
directly over the earth, from which at times an object has been
known to fall or an anchor to be let down. Even such whimsical
fancies are not without instruction, and may serve to show the reader
what that vagueness and obscurity of ideas is, of which I have been
endeavoring to trace the prevalence in the dark ages.

We now proceed to another of the features which appears to me

to mark, in a very prominent manner, the character of the stationary
period.
 CHAPTER II.

The Commentatorial Spirit of the Middle Ages.

W E have already noticed, that, after the first great achievements

of the founders of sound speculation, in the different
departments of human knowledge, had attracted the interest and
admiration which those who became acquainted with them could not
but give to them, there appeared a disposition among men to lean on
the authority of some of these teachers;—to study the opinions of
others as the only mode of forming their own;—to read nature
through books;—to attend to what had been already thought and
said, rather than to what really is and happens. This tendency of
men’s minds requires our particular consideration. Its manifestations
were very important, and highly characteristic of the stationary
period; it gave, in a great degree, a peculiar bias and direction to the
intellectual activity of many centuries; and the kind of labor with
which speculative men were occupied in consequence of this bias,
took the place of that examination of realities which must be their
employment, in order that real knowledge may make any decided
progress.

In some subjects, indeed, as, for instance, in the domains of

morals, poetry, and the arts, whose aim is the production of beauty,
this opposition between the study of former opinion and present
reality, may not be so distinct; inasmuch as it may be said by some,
that, in these subjects, opinions are realities; that the thoughts and
feelings which 202 prevail in men’s minds are the material upon
which we must work, the particulars from which we are to generalize,
the instruments which we are to use; and that, therefore, to reject the
study of antiquity, or even its authority, would be to show ourselves
ignorant of the extent and mutual bearing of the elements with which
we have to deal;—would be to cut asunder that which we ought to
unite into a vital whole. Yet even in the provinces of history and
poetry, the poverty and servility of men’s minds during the middle
ages, are shown by indications so strong as to be truly remarkable;
for instance, in the efforts of the antiquarians of almost every
European country to assimilate the early history of their own state to
the poet’s account of the foundation of Rome, by bringing from the
sack of Troy, Brutus to England, Bavo to Flanders, and so on. But
however this may be, our business at present is, to trace the varying
spirit of the physical philosophy of different ages; trusting that,
hereafter, this prefatory study will enable us to throw some light upon
the other parts of philosophy. And in physics the case undoubtedly
was, that the labor of observation, which is one of the two great
elements of the progress of knowledge, was in a great measure
superseded by the collection, the analysis, the explanation, of
previous authors and opinions; experimenters were replaced by
commentators; criticism took the place of induction; and instead of
great discoverers we had learned men.

1. Natural Bias to Authority.—It is very evident that, in such a bias

of men’s studies, there is something very natural; however strained
and technical this erudition may have been, the propensities on
which it depends are very general, and are easily seen. Deference to
the authority of thoughtful and sagacious men, a disposition which
men in general neither reject nor think they ought to reject in
practical matters, naturally clings to them, even in speculation. It is a
satisfaction to us to suppose that there are, or have been, minds of
transcendent powers, of wide and wise views, superior to the
common errors and blindness of our nature. The pleasure of
admiration, and the repose of confidence, are inducements to such a
belief. There are also other reasons why we willingly believe that
there are in philosophy great teachers, so profound and sagacious,
that, in order to arrive at truth, we have only to learn their thoughts,
to understand their writings. There is a peculiar interest which men
feel in dealing with the thoughts of their fellow-men, rather than with
brute matter. Matter feels and excites no sympathies: in seeking for
mere laws of nature, there is nothing of mental intercourse with the
great spirits of the past, as there is in 203 studying Aristotle or Plato.
Moreover, a large portion of this employment is of a kind the most
agreeable to most speculative minds; it consists in tracing the
consequences of assumed principles: it is deductive like geometry:
and the principles of the teachers being known, and being
undisputed, the deduction and application of their results is an
obvious, self-satisfying, and inexhaustible exercise of ingenuity.

These causes, and probably others, make criticism and

commentation flourish, when invention begins to fail, oppressed and
bewildered by the acquisitions it has already made; and when the
vigor and hope of men’s minds are enfeebled by civil and political
changes. Accordingly, 20 the Alexandrian school was eminently
characterized by a spirit of erudition, of literary criticism, of
interpretation, of imitation. These practices, which reigned first in
their full vigor in “the Museum,” are likely to be, at all times, the
leading propensities of similar academical institutions.
20 Degerando, Hist. des Syst. de Philos. iii. p. 134.

How natural it is to select a great writer as a paramount authority,

and to ascribe to him extraordinary profundity and sagacity, we may
see, in the manner in which the Greeks looked upon Homer; and the
fancy which detected in his poems traces of the origin of all arts and
sciences, has, as we know, found favor even in modern times. To
pass over earlier instances of this feeling, we may observe, that
Strabo begins his Geography by saying that he agrees with
Hipparchus, who had declared Homer to be the first author of our
geographical knowledge; and he does not confine the application of
this assertion to the various and curious topographical information
which the Iliad and Odyssey contain, concerning the countries
surrounding the Mediterranean; but in phrases which, to most
persons, might appear the mere play of a poetical fancy, or a casual
selection of circumstances, he finds unquestionable evidence of a
correct knowledge of general geographical truths. Thus, 21 when
Homer speaks of the sun “rising from the soft and deep-flowing
ocean,” of his “splendid blaze plunging in the ocean;” of the northern
constellation

“Alone unwashen by the ocean wave;”

and of Jupiter, “who goes to the ocean to feast with the blameless
Ethiopians;” Strabo is satisfied from these passages that Homer
knew the dry land to be surrounded with water: and he reasons in
like manner with respect to other points of geography.
21 Strabo, i. p. 5.

204 2. Character of Commentators.—The spirit of commentation,

as has already been suggested, turns to questions of taste, of
metaphysics, of morals, with far more avidity than to physics.
Accordingly, critics and grammarians were peculiarly the growth of
this school; and, though the commentators sometimes chose works
of mathematical or physical science for their subject (as Proclus,
who commented on Euclid’s Geometry, and Simplicius, on Aristotle’s
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookmass.com