Re: TMQL: Minor issue: Plurals for paths

Lars Heuer wrote:
> Hi Benjamin,
>
>    
>>> So "t/occurrences" does not match well with the mental model.
>>>        
> Rereading [1] I found at slide 76: "Consistent naming of axes - Xuân
> wants them all to be plural".
>
> Can you or Xuân explain the switch to plural terms? Why do they fit
> any better than the (commonly) used singular terms?
>
> [1]<http://www.itscj.ipsj.or.jp/sc34/open/1536.pdf>
>    
Well, as far as I can recall, the idea is that an axis is actually a 
function, and functions returning a set of something should be named in 
plural. This is common (and I think good) practice, like in Python, the 
method to retrieve all keys of a dict object is called "keys" (see 
http://docs.python.org/library/stdtypes.html#dict.keys ), not "key".

Actually sticking to this naming convention allows to distinguish 
between "Role#player" and "Association#players", and also enables 
Robert's idea.

Robert Barta wrote:
> [...]
> While I have no strong inclinations here, I could think that
>
>      t / occurrences   # list, collection
>
> could mean something different from
>
>      t / occurrence    # pick exactly one
>
> Whether this is useful per se, and/or can be exploited in the TMQL
> semantics is a question to resolve.
>
> \rho
>    
I like this idea in general. However, this should only be done if we 
have succeeded in defining a generic "exactly_one" function (or axis, if 
you will) which picks along some still-to-be-defined rules an element of 
a set|collection (and raises an error of the collection is empty). 
(Picking "right" elements is actually a non-trivial question, as the 
notion of "the most right one" may actually depend on the user locale, 
for example, something which may not be accessible by default to the 
TMQL processor.) Then it could be trivial to specify the singular 
versions by means of the "exactly_one" function (or axis).

> Best regards,
> Lars
>    
ciao,
Xuân.
_______________________________________________
topicmapmail mailing list
[email protected]
http://www.infoloom.com/mailman/listinfo/topicmapmail