Debian Codegraph @

doc/src/sgml/history.sgml

raw
1 <!-- doc/src/sgml/history.sgml -->
2
3 <sect1 id="history">
4 <title>A Brief History of <productname>PostgreSQL</productname></title>
5
6 <indexterm zone="history">
7 <primary>history</primary>
8 <secondary>of PostgreSQL</secondary>
9 </indexterm>
10
11 <para>
12 The object-relational database management system now known as
13 <productname>PostgreSQL</productname> is derived from the
14 <productname>POSTGRES</productname> package written at the
15 University of California at Berkeley. With decades of
16 development behind it, <productname>PostgreSQL</productname> is now
17 the most advanced open-source database available anywhere.
18 </para>
19
20 <para>
21 Another take on the history presented here can be found in Dr. Joe
22 Hellerstein's paper <quote>Looking Back at Postgres</quote>
23 <xref linkend="hell18"/>.
24 </para>
25
26 <sect2 id="history-berkeley">
27 <title>The Berkeley <productname>POSTGRES</productname> Project</title>
28
29 <indexterm zone="history-berkeley">
30 <primary>POSTGRES</primary>
31 </indexterm>
32
33 <para>
34 The <productname>POSTGRES</productname> project, led by Professor
35 Michael Stonebraker, was sponsored by the Defense Advanced Research
36 Projects Agency (<acronym>DARPA</acronym>), the Army Research
37 Office (<acronym>ARO</acronym>), the National Science Foundation
38 (<acronym>NSF</acronym>), and ESL, Inc. The implementation of
39 <productname>POSTGRES</productname> began in 1986. The initial
40 concepts for the system were presented in <xref linkend="ston86"/>,
41 and the definition of the initial data model appeared in <xref
42 linkend="rowe87"/>. The design of the rule system at that time was
43 described in <xref linkend="ston87a"/>. The rationale and
44 architecture of the storage manager were detailed in <xref
45 linkend="ston87b"/>.
46 </para>
47
48 <para>
49 <productname>POSTGRES</productname> has undergone several major
50 releases since then. The first <quote>demoware</quote> system
51 became operational in 1987 and was shown at the 1988
52 <acronym>ACM-SIGMOD</acronym> Conference. Version 1, described in
53 <xref linkend="ston90a"/>, was released to a few external users in
54 June 1989. In response to a critique of the first rule system
55 (<xref linkend="ston89"/>), the rule system was redesigned (<xref
56 linkend="ston90b"/>), and Version 2 was released in June 1990 with
57 the new rule system. Version 3 appeared in 1991 and added support
58 for multiple storage managers, an improved query executor, and a
59 rewritten rule system. For the most part, subsequent releases
60 until <productname>Postgres95</productname> (see below) focused on
61 portability and reliability.
62 </para>
63
64 <para>
65 <productname>POSTGRES</productname> has been used to implement many
66 different research and production applications. These include: a
67 financial data analysis system, a jet engine performance monitoring
68 package, an asteroid tracking database, a medical information
69 database, and several geographic information systems.
70 <productname>POSTGRES</productname> has also been used as an
71 educational tool at several universities. Finally, Illustra
72 Information Technologies (later merged into
73 <ulink url="https://www.ibm.com/analytics/informix"><productname>Informix</productname></ulink>,
74 which is now owned by <ulink
75 url="https://www.ibm.com/">IBM</ulink>) picked up the code and
76 commercialized it. In late 1992,
77 <productname>POSTGRES</productname> became the primary data manager
78 for the Sequoia 2000 scientific computing project described in
79 <xref linkend="ston92"/>.
80 </para>
81
82 <para>
83 The size of the external user community nearly doubled during 1993.
84 It became increasingly obvious that maintenance of the prototype
85 code and support was taking up large amounts of time that should
86 have been devoted to database research. In an effort to reduce
87 this support burden, the Berkeley
88 <productname>POSTGRES</productname> project officially ended with
89 Version 4.2.
90 </para>
91 </sect2>
92
93 <sect2 id="history-postgres95">
94 <title><productname>Postgres95</productname></title>
95
96 <indexterm zone="history-postgres95">
97 <primary>Postgres95</primary>
98 </indexterm>
99
100 <para>
101 In 1994, Andrew Yu and Jolly Chen added an SQL language interpreter
102 to <productname>POSTGRES</productname>. Under a new name,
103 <productname>Postgres95</productname> was subsequently released to
104 the web to find its own way in the world as an open-source
105 descendant of the original <productname>POSTGRES</productname>
106 Berkeley code.
107 </para>
108
109 <para>
110 <productname>Postgres95</productname> code was completely ANSI C
111 and trimmed in size by 25%. Many internal changes improved
112 performance and
113 maintainability. <productname>Postgres95</productname> release
114 1.0.x ran about 30&ndash;50% faster on the Wisconsin Benchmark compared
115 to <productname>POSTGRES</productname>, Version 4.2. Apart from
116 bug fixes, the following were the major enhancements:
117
118 <itemizedlist>
119 <listitem>
120 <para>
121 The query language PostQUEL was replaced with
122 <acronym>SQL</acronym> (implemented in the server). (Interface
123 library <link linkend="libpq">libpq</link> was named after PostQUEL.)
124 Subqueries
125 were not supported until <productname>PostgreSQL</productname>
126 (see below), but they could be imitated in
127 <productname>Postgres95</productname> with user-defined
128 <acronym>SQL</acronym> functions. Aggregate functions were
129 re-implemented. Support for the <literal>GROUP BY</literal>
130 query clause was also added.
131 </para>
132 </listitem>
133
134 <listitem>
135 <para>
136 A new program
137 (<application>psql</application>) was provided for interactive
138 SQL queries, which used <acronym>GNU</acronym>
139 <application>Readline</application>. This largely superseded
140 the old <application>monitor</application> program.
141 </para>
142 </listitem>
143
144 <listitem>
145 <para>
146 A new front-end library, <filename>libpgtcl</filename>,
147 supported <acronym>Tcl</acronym>-based clients. A sample shell,
148 <command>pgtclsh</command>, provided new Tcl commands to
149 interface <application>Tcl</application> programs with the
150 <productname>Postgres95</productname> server.
151 </para>
152 </listitem>
153
154 <listitem>
155 <para>
156 The large-object interface was overhauled. The inversion large
157 objects were the only mechanism for storing large objects. (The
158 inversion file system was removed.)
159 </para>
160 </listitem>
161
162 <listitem>
163 <para>
164 The instance-level rule system was removed. Rules were still
165 available as rewrite rules.
166 </para>
167 </listitem>
168
169 <listitem>
170 <para>
171 A short tutorial introducing regular <acronym>SQL</acronym>
172 features as well as those of
173 <productname>Postgres95</productname> was distributed with the
174 source code.
175 </para>
176 </listitem>
177
178 <listitem>
179 <para>
180 <acronym>GNU</acronym> make (instead of <acronym>BSD</acronym>
181 make) was used for the build. Also,
182 <productname>Postgres95</productname> could be compiled with an
183 unpatched <productname>GCC</productname> (data alignment of
184 doubles was fixed).
185 </para>
186 </listitem>
187 </itemizedlist>
188 </para>
189 </sect2>
190
191 <sect2 id="history-postgresql">
192 <title><productname>PostgreSQL</productname></title>
193
194 <para>
195 By 1996, it became clear that the name <quote>Postgres95</quote>
196 would not stand the test of time. We chose a new name,
197 <productname>PostgreSQL</productname>, to reflect the relationship
198 between the original <productname>POSTGRES</productname> and the
199 more recent versions with <acronym>SQL</acronym> capability. At
200 the same time, we set the version numbering to start at 6.0,
201 putting the numbers back into the sequence originally begun by the
202 Berkeley <productname>POSTGRES</productname> project.
203 </para>
204
205 <para>
206 <productname>Postgres</productname> is still considered an official
207 project name, both because of tradition and because people find it
208 easier to pronounce <productname>Postgres</productname> than
209 <productname>PostgreSQL</productname>.
210 </para>
211
212 <para>
213 The emphasis during development of
214 <productname>Postgres95</productname> was on identifying and
215 understanding existing problems in the server code. With
216 <productname>PostgreSQL</productname>, the emphasis has shifted to
217 augmenting features and capabilities, although work continues in
218 all areas.
219 </para>
220
221 <para>
222 Details about what has happened in
223 each <productname>PostgreSQL</productname> release since then
224 can be found at
225 <ulink url="https://www.postgresql.org/docs/release/">https://www.postgresql.org/docs/release/</ulink>.
226 </para>
227 </sect2>
228 </sect1>