On 2010-12-14, at 17:38, Rami Ojares wrote:

> 15.12.2010 0:20, Rami Ojares kirjoitti:
>> And I think what you are aiming at is not just one possible representation 
>> but
>> the canonical string representation of the value.
> This remark was bs.
> Because clearly the type imlementor can support many different strings to 
> represent the same value.
> So you were talking about a possible string representation of a value.

Yes, although there might be a canonical representation, or perhaps there is a 
way to ask for a particular representation.  The LispM had a gesture that was 
called "re-present differently" -- you could click on the representation of a 
value and it would be replaced by the value in a different representation.  For 
instance you could click on a number to see it in decimal/hec/octal.

> The point being that this representation is always of type string.
> So another name candidate for the attribute name could be "stringvalue".
> That clearly states the difference to value which can be also some other type 
> than string.

Don't get hung up too much on the fact that the internal value of an external 
representation is stored as a string.  It's more like the concept of a number 
(abstract) and the numerals that we use to represent it.  In that sense, 
`representationvalue` would be a good name; but it is rather long.  I'd like 
something shorter.


Reply via email to