New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Improve Doc Reference section #383
Conversation
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
reference_operator.xml \ | ||
reference_output.xml \ | ||
reference_processing.xml \ | ||
reference_raster.xml \ | ||
reference_temporal.xml \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is a historic thing. ST_ initially is SpatioTemporal, but by time to design the spec for temporal writers got tired and just said "ok let it be SpaTial". @estebanzimanyi you know the terminology about time, can you advise here?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The only functions under Temporal were ones dealing with Trajectories, so it seems better to give them a section by that name. That highlights the ability to work with trajectories.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
As for the ST_ naming, maybe that reflected grand ambitions at one time. But AFAIK the specs haven't really fulfilled these dreams, so now it is just a (nice?) way to namespace spatial functions.
<refsynopsisdiv> | ||
<funcsynopsis> | ||
<funcprototype> | ||
<funcdef>geometry <function>ST_Expand</function></funcdef> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
but but but ST_Expand returns rectangular geometry and not a box, if fed with geometry. This minor distinction drove me nuts when I tried to normalize "what the function gets you from which input" graph, after which I decided that boxes are kind of rectangular areas. Also, ST_OrientedEnvelope - it's not a box?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, wasn't sure what was best to do with this. Ideally it would appear under both Geometry and Bounding Box. Too bad the name is overloaded, but there it is.
In the end I decided this was more of a bounding box kind of thing. But can change it back if that's better.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
And as for ST_OrientedEnvelope, I also find that name confusing. Perhaps it should be called ST_OrientedRectangle
? But it's definitely not a box.
My take on naming practice is that the support of box*
types is indicated by using the word "Box" in names. Other names like Rectangle
and Envelope
indicate that a geometry type is used. The use of the term Envelope at all is perhaps unfortunate, since in PG that function is provided by the box
types. But that's what the Simple Features spec uses, so we should use it in a limited way.
<para>Returns the integer SRID of the | ||
specified geometry column by searching through the GEOMETRY_COLUMNS table. | ||
If the geometry column has not been properly added (e.g. with the | ||
<xref linkend="AddGeometryColumn" /> function), this function will not work.</para> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
AddGeometryColumn is not a good way to add columns since PostGIS 2?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That's my understanding too. So this description should indicate this more strongly. But will need to research how to phrase this.
<refnamediv> | ||
<refname>ST_Accum</refname> | ||
|
||
<refpurpose>An aggregate function to construct an array of geometries.</refpurpose> |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ST_Accum is effectively array_agg. I understand it was added when array_agg wasn't available, but how about removing it now that there is no supported Postgres versions without array_agg?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sounds good to me. Can we get a third vote of +1?
@@ -122,19 +122,19 @@ XML_SOURCES = \ | |||
performance_tips.xml \ | |||
postgis.xml \ | |||
reference_accessor.xml \ | |||
reference_bbox.xml \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ST_Envelope, ST_OrientedEnvelope I'd expect nearby. Naming needs tweaking then - how about calling section with something like "rectangle functions" so box / geometry-with-rect-angles distinction is not there?
This may also consume operator section since these are defined on boxes?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
See previous comment. The Bounding Box section does not contain either "Envelope" function, so I think this is ok? When I reorg the Accessors and Processing sections we can think about where they should go.
My opinion is that they belong in different sections, even though they are unfortunately named similarly. One is just a simple accessor, but the OrientedEnvelope requires a full-blown spatial algorithm to compute. I think it should be called OrientedRectangle or something like that. It should not mention Envelope, because that implies an axis-oriented rectangle.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think the operator section should remain distinct, since it's long already. And operators are clearly different things to functions.
A quick answer, I can dig more if needed.
I think we cannot deviate from ST_* which, as Darafei rightfully pointed
out, was meant SpatioTemporal at the beginning. Since it took years (more
precisely decades) to begin addressing the spatiotemporal aspects, the
meaning of the ST_ prefix became SpaTial.
FYI, the current (draft) standards for spatiotemporal aspects can be found
here
https://www.opengeospatial.org/standards/movingfeatures
It may be the case that the prefix for those functions at the SQL level
would be MF_ but according to my understanding this has still to be decided
Esteban
…On Wed, Mar 13, 2019 at 12:00 PM Darafei Praliaskouski < ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In doc/Makefile.in
<#383 (comment)>:
> @@ -122,19 +122,19 @@ XML_SOURCES = \
performance_tips.xml \
postgis.xml \
reference_accessor.xml \
+ reference_bbox.xml \
ST_Envelope, ST_OrientedEnvelope I'd expect nearby. Naming needs tweaking
then - how about calling section with something like "rectangle functions"
so box / geometry-with-rect-angles distinction is not there?
This may also consume operator section since these are defined on boxes?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#383 (review)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AZp2BoDI-buBFamIehtgqR8PLM8U-5wJks5vWNpKgaJpZM4br2nE>
.
--
------------------------------------------------------------
Prof. Esteban Zimanyi
Department of Computer & Decision Engineering (CoDE) CP 165/15
Universite Libre de Bruxelles
Avenue F. D. Roosevelt 50
B-1050 Brussels, Belgium
fax: + 32.2.650.47.13
tel: + 32.2.650.31.85
e-mail: ezimanyi@ulb.ac.be
Internet: http://code.ulb.ac.be/
------------------------------------------------------------
|
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
Signed-off-by: Martin Davis <mtnclimb@gmail.com>
@estebanzimanyi @Komzpa Agreed, there's definitely no mandate to use a prefix other than ST_ right now. I see now where the word Temporal comes from. However, "Temporal" has a much more general meaning in the database world. So it needs to be qualified in the PG context. Perhaps this section should be called Temporal Trajectory Functions? That would tie in with the OGC spec (which I wasn't aware of, and is nice to see). Although are there really any other kind of trajectories? I note however that the PG functions do not seem to follow the OGC naming? It might be nice to "namespace" the Trajectory functions. Maybe by using the prefix "ST_TrajectoryXXX"? |
@Komzpa once again thanks for the detailed and thought-provoking review. Can we make this PR just about reorganizing the reference section, and defer the issues of removing/improving docs to another issue? That will let this get committed and I can work on further improvements. |
To summarize: apart from the reorganization which is the focus of this PR, for reference the other changes suggested are:
|
@dr-jts merged. a ticket linkr mentioned somewhere on the PR would be really helpful for composing a Closes: / References: section of merge svn commit message :) |
De-emphasize it. Go through docs and make sure we only mention it on its own page.
I've raised it on mailing list.
My dream is to take @estebanzimanyi's work and make it part of PostGIS. |
It is my dream also. Working hard toward this. This weekend I transformed
documentation from LaTeX to docbook format to be compatible with PostGIS.
Still have to finalize selectivity functions for the operators. But we are
little by little reaching a first release.
Le mar. 19 mars 2019 à 06:36, Darafei Praliaskouski <
notifications@github.com> a écrit :
… Possibly drop AddGeometryColumn doc, or make clear it is no longer needed?
De-emphasize it. Go through docs and make sure we only mention it on its
own page.
Remove or deprecate ST_Accum
I've raised it on mailing list.
Possibly call section Temporal Trajectory Support ?
My dream is to take @estebanzimanyi <https://github.com/estebanzimanyi>'s
work and make it part of PostGIS.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#383 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AZp2BtvZ0LSx6NQ_B8YW2bltiPEul1FXks5vYHddgaJpZM4br2nE>
.
|
Some more doc improvements. See commits for details.