The word "virgin" used incorrectly and probably better off replaced

On 2019-Nov-07, PG Doc comments form wrote:

Page: https://www.postgresql.org/docs/12/manage-ag-templatedbs.html
Description:

The use of the word virgin as an adjective is incorrect here and also an
anachronism. It is better off replaced with the word pristine - quotes
unnecessary. Note the word virgin appears in another page of the
documentation (a search will find it).

Just because a word has sexual connotations does not imply that it
doesn't have non-sexual meanings.

Merriam-Webster lists https://www.merriam-webster.com/dictionary/virgin

2: FRESH, UNSPOILED
specifically : not altered by human activity

6: free of impurity or stain : UNSULLIED

which seems to apply well to all cases at hand. Also:
First Known Use of virgin
Noun: 13th century, in the meaning defined at sense 2a
Adjective: 14th century, in the meaning defined at sense 6

History and Etymology for virgin
Noun: Middle English, from Anglo-French virgine, from Latin virgin-, virgo young woman, virgin

That said, in two of the three phrases where the word appears, the quoted
adjective adds no value. We could remove the quoted word entirely in
all three places and nothing would be lost. But if we do that, then the
third occurrence of the word would become inintelligible:

"This is particularly handy when restoring a pg_dump dump: the dump
script should be restored in a virgin database to ensure that one
recreates the correct contents of the dumped database, without
conflicting with objects that might have been added to template1 later
on."

because we have not explained what a "virgin database" is. We could say
"empty", which seems better suited than both "virgin" and "pristine"
anyway.

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On 7 Nov 2019, at 16:03, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

because we have not explained what a "virgin database" is.

I think this is the key observation.

We could say "empty", which seems better suited than both "virgin" and
"pristine" anyway.

empty is a lot better, but still isn't conveying the state of the database
without there being room for interpretation. (My grasp of the english language
isn't enough to suggest a better alternative however).

cheers ./daniel

On Thu, Nov 7, 2019 at 07:55:22PM +0100, Daniel Gustafsson wrote:

On 7 Nov 2019, at 16:03, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

because we have not explained what a "virgin database" is.

I think this is the key observation.

We could say "empty", which seems better suited than both "virgin" and
"pristine" anyway.

empty is a lot better, but still isn't conveying the state of the database
without there being room for interpretation. (My grasp of the english language
isn't enough to suggest a better alternative however).

I am thinking "pristine" would be a good word here.

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ As you are, so once was I.  As I am, so you will be. +
+                      Ancient Roman grave inscription +

On 2019-Nov-07, Bruce Momjian wrote:

On Thu, Nov 7, 2019 at 07:55:22PM +0100, Daniel Gustafsson wrote:

On 7 Nov 2019, at 16:03, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

We could say "empty", which seems better suited than both "virgin" and
"pristine" anyway.

empty is a lot better, but still isn't conveying the state of the database
without there being room for interpretation. (My grasp of the english language
isn't enough to suggest a better alternative however).

I am thinking "pristine" would be a good word here.

But you would have to explain that a database created as a copy of
template1 may somehow not be pristine. Maybe we should just use a
phrase that describes what we mean, something like "a database that
doesn't contain objects other than default system ones."

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On Thu, Nov 7, 2019 at 06:50:10PM -0300, Alvaro Herrera wrote:

On 2019-Nov-07, Bruce Momjian wrote:

On Thu, Nov 7, 2019 at 07:55:22PM +0100, Daniel Gustafsson wrote:

On 7 Nov 2019, at 16:03, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

We could say "empty", which seems better suited than both "virgin" and
"pristine" anyway.

empty is a lot better, but still isn't conveying the state of the database
without there being room for interpretation. (My grasp of the english language
isn't enough to suggest a better alternative however).

I am thinking "pristine" would be a good word here.

But you would have to explain that a database created as a copy of
template1 may somehow not be pristine. Maybe we should just use a
phrase that describes what we mean, something like "a database that
doesn't contain objects other than default system ones."

True.

--
Bruce Momjian <bruce@momjian.us> http://momjian.us
EnterpriseDB http://enterprisedb.com

+ As you are, so once was I.  As I am, so you will be. +
+                      Ancient Roman grave inscription +

On 7 Nov 2019, at 22:50, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

On 2019-Nov-07, Bruce Momjian wrote:

On Thu, Nov 7, 2019 at 07:55:22PM +0100, Daniel Gustafsson wrote:

On 7 Nov 2019, at 16:03, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

We could say "empty", which seems better suited than both "virgin" and
"pristine" anyway.

empty is a lot better, but still isn't conveying the state of the database
without there being room for interpretation. (My grasp of the english language
isn't enough to suggest a better alternative however).

I am thinking "pristine" would be a good word here.

But you would have to explain that a database created as a copy of
template1 may somehow not be pristine. Maybe we should just use a
phrase that describes what we mean, something like "a database that
doesn't contain objects other than default system ones."

Agreed. I like your suggestion, or the inverse of it: "a database without any
user defined objects".

cheers ./daniel

On 2019-Nov-08, Daniel Gustafsson wrote:

Agreed. I like your suggestion, or the inverse of it: "a database without any
user defined objects".

Here's a proposed patch.

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

On 8 Nov 2019, at 14:10, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:

On 2019-Nov-08, Daniel Gustafsson wrote:

Agreed. I like your suggestion, or the inverse of it: "a database without any
user defined objects".

Here's a proposed patch.

+1, LGTM

cheers ./daniel

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

Here's a proposed patch.

I don't like this wording much, because "no user-defined objects"
is not a sufficient specification of what we are talking about.
You need to also capture the property that none of the system-
defined objects have been altered. Now that we explicitly support
things like altering the ACLs of system-defined objects, I do not
think it's okay to take that part for granted.

regards, tom lane

On 2019-Nov-08, Tom Lane wrote:

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

Here's a proposed patch.

I don't like this wording much, because "no user-defined objects"
is not a sufficient specification of what we are talking about.
You need to also capture the property that none of the system-
defined objects have been altered. Now that we explicitly support
things like altering the ACLs of system-defined objects, I do not
think it's okay to take that part for granted.

Hmm. Maybe we can say "pristine database" and then add this explanation
in a parenthical comment:

This is particularly handy when restoring a
<literal>pg_dump</literal> dump: the dump script should be restored in a
pristine database (one where no user-defined objects exist and where
system objects have not been altered), to ensure that one recreates
the correct contents of the dumped database, without conflicting
with objects that might have been added to
<literal>template1</literal> later on.

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

Hmm. Maybe we can say "pristine database" and then add this explanation
in a parenthical comment:

This is particularly handy when restoring a
<literal>pg_dump</literal> dump: the dump script should be restored in a
pristine database (one where no user-defined objects exist and where
system objects have not been altered), to ensure that one recreates
the correct contents of the dumped database, without conflicting
with objects that might have been added to
<literal>template1</literal> later on.

So the patch becomes s/virgin/pristine/g plus add a parenthetical
definition for the first use? Works for me.

regards, tom lane

On 8 Nov 2019, at 16:19, Tom Lane <tgl@sss.pgh.pa.us> wrote:

So the patch becomes s/virgin/pristine/g plus add a parenthetical
definition for the first use? Works for me.

Agreed.

cheers ./daniel

On 2019-Nov-08, Tom Lane wrote:

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

Hmm. Maybe we can say "pristine database" and then add this explanation
in a parenthical comment:

This is particularly handy when restoring a
<literal>pg_dump</literal> dump: the dump script should be restored in a
pristine database (one where no user-defined objects exist and where
system objects have not been altered), to ensure that one recreates
the correct contents of the dumped database, without conflicting
with objects that might have been added to
<literal>template1</literal> later on.

So the patch becomes s/virgin/pristine/g plus add a parenthetical
definition for the first use? Works for me.

Well, there are three uses of the word "virgin". The first is for
"virgin user", and the patch turns that into just "user". The second
one is for "virgin database" and the patch has the effect you describe.
The third one is also s/virgin//.

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

On 2019-Nov-08, Tom Lane wrote:

So the patch becomes s/virgin/pristine/g plus add a parenthetical
definition for the first use? Works for me.

Well, there are three uses of the word "virgin". The first is for
"virgin user", and the patch turns that into just "user".

Uh, no, read the next lines. In both cases those are referring
to "virgin user database" or "virgin database", and this patch
is removing an important qualifier. It needs to be s/virgin/pristine/
in all these places.

Since the third case is well separated from the other two, maybe
we need to repeat the parenthetical definition there too.

regards, tom lane

Everyone,

Thank you for the attention paid to this.

Brian

On 2019-Nov-08, Tom Lane wrote:

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

On 2019-Nov-08, Tom Lane wrote:

So the patch becomes s/virgin/pristine/g plus add a parenthetical
definition for the first use? Works for me.

Well, there are three uses of the word "virgin". The first is for
"virgin user", and the patch turns that into just "user".

Uh, no, read the next lines. In both cases those are referring
to "virgin user database" or "virgin database", and this patch
is removing an important qualifier. It needs to be s/virgin/pristine/
in all these places.

Doh, right.

One problem with doing it that way is that the proposed parenthical
comment partly duplicates the text immediately following it, so I'm no
longer so sure that adding it is good; I think that changing "local
additions" to "local additions and changes" might be sufficient, or
maybe that is too obscure for novices? For create_database.sgml it does
seem to make a little more sense, but I'm not 100% there either.

Maybe "changes" can become "database-local system changes"? i.e.,

By instructing CREATE DATABASE to copy template0 instead of
template1, you can create a pristine user database that contains
none of the site-local additions and database-local system changes
in template1.

... though, argh, "-local" appearing twice makes that look bad too :-(
(I'm not sure that it is clear what a "database-local system change" is.)

--
�lvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Alvaro Herrera <alvherre@2ndquadrant.com> writes:

One problem with doing it that way is that the proposed parenthical
comment partly duplicates the text immediately following it, so I'm no
longer so sure that adding it is good; I think that changing "local
additions" to "local additions and changes" might be sufficient, or
maybe that is too obscure for novices? For create_database.sgml it does
seem to make a little more sense, but I'm not 100% there either.

I think we should stick to the wording we've agreed to be clearer.
Maybe do manage-ag.sgml like so:

By instructing CREATE DATABASE to copy template0 instead of
template1, you can create a pristine user database, that is one where
no user-defined objects exist and where system objects have not been
altered. This is particularly handy ...

regards, tom lane