[GRASS-dev] terminology issues in grass7

[Michael Barton]

Hi Martin,

See below.


On Jun 12, 2009, at 11:27 AM, Martin Landa wrote:

> Hi,
>
> 2009/6/12 Michael Barton <michael.barton at asu.edu>:
>
> [...]
>
>> I probably shouldn't add more, but I will anyway.
>>
>> I like calling vector and raster files maps. It is really easy for  
>> users to
>> understand what these files are. Maps can be added to display  
>> layers (i.e.,
>> like layers in a CAD or drawing package) for display and  
>> visualization.
>
> that can be also confusing, data can be stored e.g. in the database as
> PostGIS tables - vector and also raster data (see wktraster) instead
> of files. I would hesitate to use "files" in this connection. Also
> "files maps" seems to be strange to me - I still see "map" as
> something related to the cartography. I would call it "data layers".

 From this perspective, data layers seems sensible and I even talk  
about geospatial data when I teach GIS. I also understand the  
cartographic perspective that maps are the final, often paper, result  
of combining multiple geospatial data layers. Nonetheless, most users  
will find it less confusing if we just continue to call them maps-- 
with the idea that we've moved maps from paper to digital media. Note  
that this was the original usage in one of the world's oldest GIS  
systems still in use (i.e., GRASS). And looking at the 1980's video  
that someone rediscovered, the parallels between paper maps and  
digital maps were made so that potential users could better understand  
a GIS. From a personal perspective, I really don't mind data layers at  
all. I just think that map is easer for most users to understand even  
if it seems somewhat inaccurate from a more technical perspective.


>
>> The features that are currently called vector "layers" really serve a
>> database function. Given that, my preference is that they be called
>> something in database jargon that is also very easily recognizable.  
>> AFAIK,
>> the term "layer" is not a term commonly used for DBMS files and  
>> functions.
>> The closest common term for what our "layer" does is a key field.  
>> Whether or
>> not the key field is use to connect the vector to an attribute  
>> table, that
>> is what it is good for ultimately. So that is why I favor some  
>> version of
>> "key" for this feature.
>
> It's not always related to the database function, but you are right in
> the most cases it is. I still see "something-set" as good choice,
> because it's grouping cats/keys/ids to the set. E.g. if we use
> 'keyset', then we should call 'cat' as 'key'. Note that we already use
> 'feature id' for different meaning - every feature has unique fid.

Keyset is fine with me too. We are talking about a single column of  
values, or database field. So simply "key" is OK too-as in the vector  
has 2 keys, one of which is linked with attribute table A.

I agree with you that feature ID is incorrect. Each vector has a  
unique feature ID, but this is not what links it with an attribute  
table. The ID relates to the vector object. Cat values can be  
identical to the ID values, but this is not at all necessary.

Michael

>
> Martin
>
> -- 
> Martin Landa <landa.martin gmail.com> * http://gama.fsv.cvut.cz/~landa


______________________________
C. Michael Barton, Professor of Anthropology
Director of Graduate Studies, School of Human Evolution & Social Change
Director, Center for Social Dynamics & Complexity
Arizona State University
Tempe, AZ  85287-2402
USA

voice: 480-965-6262; fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton