[GRASS5] adding metadata in GRASS 5.7

GRASS GIS GRASS GIS developer
1

Dear list,

I would like to add metadata to maps in GRASS 5.7 but I can't find any modules to edit related history files in the directory /hist.

For raster data in GRASS 5.0 and 5.3 r.support allows to add additional comments, but this module doesn't exist in GRASS 5.7 anymore.

For vector data, the old v.digit allows to add comments but the new digitizing tool does not.

Does there already exist a module to add metadata, is this simply missing yet or will metadata be handled totally different in GRASS 5.7?

  thanks a lot
   Otto

2

On Thursday 02 September 2004 10:34, Otto Dassau wrote:

Dear list,

I would like to add metadata to maps in GRASS 5.7 but I can't find any
modules to edit related history files in the directory /hist.

For raster data in GRASS 5.0 and 5.3 r.support allows to add additional
comments, but this module doesn't exist in GRASS 5.7 anymore.

For vector data, the old v.digit allows to add comments but the new
digitizing tool does not.

Does there already exist a module to add metadata, is this simply missing
yet or will metadata be handled totally different in GRASS 5.7?

I think that vector metadata should be stored in 'head', not in 'hist'.
Currently there are no tools for manipulation of user metadata in 'head' file.
Which kind of metadata?

Radim

3

On Thu, 2 Sep 2004 13:36:55 +0200
Radim Blazek <blazek@itc.it> wrote:

On Thursday 02 September 2004 10:34, Otto Dassau wrote:
> Dear list,
>
> I would like to add metadata to maps in GRASS 5.7 but I can't find any
> modules to edit related history files in the directory /hist.
>
> For raster data in GRASS 5.0 and 5.3 r.support allows to add additional
> comments, but this module doesn't exist in GRASS 5.7 anymore.
>
> For vector data, the old v.digit allows to add comments but the new
> digitizing tool does not.
>
> Does there already exist a module to add metadata, is this simply missing
> yet or will metadata be handled totally different in GRASS 5.7?

I think that vector metadata should be stored in 'head', not in 'hist'.
Currently there are no tools for manipulation of user metadata in 'head' file.
Which kind of metadata?

Hi Radim,

I was using vmap0 data and thought of adding some comments to them. But I was missing a place to put those meta information about creator, scale, quality, special notes, time...

I thought about the way it is done for the spearfish data set.

thanks
  Otto

4

Radim,

metadata can be quite large so it would be useful to have a separate file for
it. I don't know whether it is possible to add it to GRASS5.7 at this point
but they are very important, as they often include information about the
data accuracy, when exactly the survey was made, who is the point of contact,
what kind of processing was done to the data and anything else that the data creator
may consider important. US has a standard metadata form that is provided
for each data set - attached is an example of standard metadata file.
It does not have to be so detailed, but a lot of the information provided
in this file is essential when this data are to be combined with data
from other sources.
I am sure that for many users it is important to be able to edit the metadata,
especially if one wants to add some comments about the processing that was done
to the file within GRASS. So at least to keep the capabilities for handling
metadata as in 5.3 is needed.

Helena

Radim Blazek wrote:

On Thursday 02 September 2004 10:34, Otto Dassau wrote:

Dear list,

I would like to add metadata to maps in GRASS 5.7 but I can't find any
modules to edit related history files in the directory /hist.

For raster data in GRASS 5.0 and 5.3 r.support allows to add additional
comments, but this module doesn't exist in GRASS 5.7 anymore.

For vector data, the old v.digit allows to add comments but the new
digitizing tool does not.

Does there already exist a module to add metadata, is this simply missing
yet or will metadata be handled totally different in GRASS 5.7?

I think that vector metadata should be stored in 'head', not in 'hist'.
Currently there are no tools for manipulation of user metadata in 'head' file.
Which kind of metadata?

Radim

_______________________________________________
grass5 mailing list
grass5@grass.itc.it
http://grass.itc.it/mailman/listinfo/grass5

(attachments)

hmitaso30088_raw.met (11.9 KB)

5

Helena wrote:
[...]

It does not have to be so detailed, but a lot of the information provided
in this file is essential when this data are to be combined with data
from other sources.
I am sure that for many users it is important to be able to edit the metadata,
especially if one wants to add some comments about the processing that was done
to the file within GRASS. So at least to keep the capabilities for handling
metadata as in 5.3 is needed.

Why not use the ISO standards for metadata in Grass ? BTW, it seems that XML
is becomming a good choice for structuring this kind od data.

Just my 0.02 €

--
Michel Wurtz - Auzeville-Tolosane

6

On Thursday 02 September 2004 22:37, Michel Wurtz wrote:

Helena wrote:
[...]

> It does not have to be so detailed, but a lot of the information provided
> in this file is essential when this data are to be combined with data
> from other sources.
> I am sure that for many users it is important to be able to edit the
> metadata, especially if one wants to add some comments about the
> processing that was done to the file within GRASS. So at least to keep
> the capabilities for handling metadata as in 5.3 is needed.

Why not use the ISO standards for metadata in Grass ? BTW, it seems that
XML is becomming a good choice for structuring this kind od data.

XML seems to be optimal for data like that sent by Helena.
The current structure of 'head' file is a list of key:value
pairs, which probably is not enough.
I don't have any experience with metadata and currently I don't need that,
so I leave this for somebody else.

Radim

7

On Thu, Sep 02, 2004 at 10:37:58PM +0200, Michel Wurtz wrote:

Helena wrote:

[...]
>It does not have to be so detailed, but a lot of the information provided
>in this file is essential when this data are to be combined with data
>from other sources.
>I am sure that for many users it is important to be able to edit
>the metadata, especially if one wants to add some comments about
>the processing that was done to the file within GRASS. So at
>least to keep the capabilities for handling metadata as in 5.3 is
>needed.

Why not use the ISO standards for metadata in Grass ? BTW, it
seems that XML is becomming a good choice for structuring this
kind od data.

I am an XML skeptic, because of the hype.
(XML only solves the character set encoding problem, it does not
help you with the contents and its sematics.)

Also I am a meta data standards skeptic because
you cannot really enforce a number of fields to be filled all the time.
But the ability to store annotations (which are meta data in a way)
is very important.

Thus having a place where free text for annotations
could be saved would be an important advantages.

8

Bernhard Reiter wrote:

> [...]
> >It does not have to be so detailed, but a lot of the information provided
> >in this file is essential when this data are to be combined with data
> >from other sources.
> >I am sure that for many users it is important to be able to edit
> >the metadata, especially if one wants to add some comments about
> >the processing that was done to the file within GRASS. So at
> >least to keep the capabilities for handling metadata as in 5.3 is
> >needed.
>
> Why not use the ISO standards for metadata in Grass ? BTW, it
> seems that XML is becomming a good choice for structuring this
> kind od data.

I am an XML skeptic, because of the hype.
(XML only solves the character set encoding problem, it does not
help you with the contents and its sematics.)

It also helps solve structuring issues, e.g. how to delimit sections
and subsections.

Also I am a meta data standards skeptic because
you cannot really enforce a number of fields to be filled all the time.
But the ability to store annotations (which are meta data in a way)
is very important.

Thus having a place where free text for annotations
could be saved would be an important advantages.

By using a standardised encapsulation mechanism such as XML, you can
ensure that specific types of information can be identified, without
limiting the type of data which can be stored.

E.g. you can specify that a particular tag is used to delimit
sections, so that an application can add a new section, containing
whatever data it wants, without interfering with existing sections.

You can define additional tags for common fields such as date, author
etc, a general purpose tag for "human-readable text" which has no
specific interpretation, a tag for unspecified types of potentially
machine-readable data (e.g. <data type="...">blah blah</data>).

Even the most basic schema would be an improvement over the current
history file, which consists of 8 specific fields, each consisting of
a single line of up to 80 characters, plus up to 50 additional
80-character lines with no internal structure.

--
Glynn Clements <glynn.clements@virgin.net>

9

> > It does not have to be so detailed, but a lot of the information
> > provided in this file is essential when this data are to be
> > combined with data from other sources.
> > I am sure that for many users it is important to be able to edit
> > the metadata, especially if one wants to add some comments about
> > the processing that was done to the file within GRASS. So at least
> > to keep the capabilities for handling metadata as in 5.3 is
> > needed.

...

I don't have any experience with metadata and currently I don't need
that, so I leave this for somebody else.

What would be nice is something like v.touch/v.comments to add text from
stdin or a file into the 'v.info -h' history. It is often useful to
write some notes down somewhere.

As for r.support, it needs to be rewritten, IMO. It's pretty much
unusable in 5.7 without at minimum adding a title="" command line option.
As Glynn noted, the raster 'hist' comments are fairly limited in size
and you need to use struct History;G_short_history();G_write_history()
which pretty much limits you to C modules during initial map creation.
(and I thought it was 80cols x 40lines? shrug)

Hamish

10

On Fri, Sep 03, 2004 at 10:54:31AM +0100, Glynn Clements wrote:

Bernhard Reiter wrote:

> Thus having a place where free text for annotations
> could be saved would be an important advantages.

By using a standardised encapsulation mechanism such as XML, you can
ensure that specific types of information can be identified, without
limiting the type of data which can be stored.

Even the most basic schema would be an improvement over the current
history file, which consists of 8 specific fields, each consisting of
a single line of up to 80 characters, plus up to 50 additional
80-character lines with no internal structure.

Most important is to have a place where just text (without any other
contstrains can go in). Of course further improvements about the
structure are welcome (and XML might help here).