Tony,
Of course you are correct. It is interesting, however, that the object model refers to them as "attribute" objects and what we know as attributes, it calls "attributereference" objects.
SAM90,
Sounds like someone exploded the blocks and then edited the attributes to have the text they wanted or copied attributes around as if they had been dtext and changed the tag values. If you are interested in learning the lisp necessary to perform the task, then you can obtain a selection set for attribute definitions by (setq ss (ssget (list (cons 0 "attdef"))))
Then you can work through the selection set in a repeat or while loop and (entget (ssname ss i)) each attdef.
For each of those entities, you need to change only a few of the codes in the entity list of the attribute definition to make a valid text enitity list. Once you have the entity list, then you can create the new text entity and delete the old attribute definition.
(entmake <entity list>) and then
(entdel <attribute ename>)
The codes to change are 0, some of the 100 codes, the tag code 2 becomes code 1, and justification code 74 becomes
73.
There is no discernable difference on a plot, however,
between what you have and the text entities they would
be converted into.
Regards,
Doug
Tony Tanzillo said:
There's no such thing as "loose attributes not in a block".
Attributes can only exist within an insert/seqend sequence, so perhaps you are referring to attribute definitions rather than attributes?
attributes not in a block. - Kev