Re: Terms.

2011/2/20 Bruce Momjian <bruce(at)momjian(dot)us>

> Alvaro Herrera wrote:
> > Excerpts from Dmitriy Igrishin's message of lun sep 20 15:52:43 -0400
> 2010:
> > > Hey Bruce,
> > >
> > > Its interesting indeed, because I always wanted to know why libpq
> > > named "libpq". Thanks!
> > > But I wrote not about libpq functions, I just wanted to say about
> > > unnecessary (IMO) redundancy - the combinations of terms "backend
> server"
> > > seems to me inaccuracy.
> >
> > could be, but we're already using that term elsewhere. Maybe we should
> > consistently use some other term.
> >
> >
> > $ git grep 'backend server'
> >
> > doc/src/sgml/libpq.sgml: backend server and to receive the results of
> these queries.
> > doc/src/sgml/libpq.sgml: <productname>PostgreSQL</productname> backend
> server. An
> > doc/src/sgml/libpq.sgml: server process</><tertiary>in libpq</></>
> of the backend server
> > doc/src/sgml/problems.sgml: Here <quote>program</quote> refers to any
> executable, not only the backend server.
> > doc/src/sgml/problems.sgml: are specifically talking about the backend
> server, mention that, do not
> > doc/src/sgml/problems.sgml: backend server process is quite different
> from crash of the parent
> > doc/src/sgml/query.sgml: backend server machine, not the client, since
> the backend server
> > doc/src/sgml/ref/pg_ctl-ref.sgml: backend server (<xref
> linkend="app-postgres">), or displaying the
> > src/backend/libpq/pqsignal.c: * the frontend commands as well as the
> backend server.
> > src/backend/main/main.c: printf(_(" -T send SIGSTOP to
> all backend servers if one dies\n"));
> > src/backend/po/de.po:msgid " -T send SIGSTOP to all backend
> servers if one dies\n"
> > src/backend/po/es.po:msgid " -T send SIGSTOP to all backend
> servers if one dies\n"
> > src/backend/po/fr.po:msgid " -T send SIGSTOP to all backend
> servers if one dies\n"
> > src/backend/po/ja.po:msgid " -T send SIGSTOP to all backend
> servers if one dies\n"
> > src/backend/po/pt_BR.po:msgid " -T send SIGSTOP to all
> backend servers if one dies\n"
> > src/backend/po/tr.po:msgid " -T send SIGSTOP to all backend
> servers if one dies\n"
> > src/bin/psql/settings.h: int sversion; /* backend server
> version */
> > src/test/examples/testlibpq4.c: pghost = NULL; /* host name
> of the backend server */
> > src/test/examples/testlibpq4.c: pgport = NULL; /* port of
> the backend server */
> > src/test/examples/testlibpq4.c: pgtty = NULL; /* debugging
> tty for the backend server */
>
> Any ideas on this? Is "backend" better than "backend server"?
>
Better, undoubtedly.
Another alternative is just "server". Why? Well, for consistency, e.g. with
http://www.postgresql.org/docs/9.0/static/errcodes-appendix.html
Here you may see phrase "All messages emitted by the PostgreSQL server ..."
and furthermore, there are some SQL standard's "SQLSTATE" code
which name includes "server" rather than "backend":
SQLSERVER REJECTED ESTABLISHMENT OF SQLCONNECTION
But anyway, phrase "backend server" is incorrect.

>
> --
> Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
> EnterpriseDB http://enterprisedb.com
>
> + It's impossible for everything to be true. +
>

--
// Dmitriy.