It is nice. Especially as it's supposed to become a free addon for other CADs as well.
Noticed though the comments on that blog page: the vle-remove-all function is superfluous as the existing VLisp vl-remove does exactly the same thing, only faster.
Some of the others might be very efficient though. Especially those extracting and/or modifying only the dxf code you wanted instead of the entire list as entget/entmod does. Should speed up these things drastically - perhaps even faster than the vla methods since there's no need to vlax-ename->vla-object anymore. I wonder if this would get around the issue of annotative text scaling each time an entmod is performed on it.
Some things which might also help is (e.g.) if the vle-entget could retrieve more than one code at a time. E.g. if the dxf code is a list, then retrieve all values assoc'ed in that list. Same goes for the vle-entmod. And if those codes with defaults could be included in the vle-entget (e.g. bylayer colour/linetype) it might save yet some more coding.
Not too sure about the predicator functions for variable type though. Is (vle-integerp obj) a lot faster than simply typing (= (type obj) 'Int)? It's not a lot of saving on keypresses.
The vle-file->list seems awesome! Finally a full read of a file in one go! Must be a lot faster than reading character for character / line by line! How does it handle binary files? Could it be set to keep all \N \R combinations at the end of each line?
One function I can suggest is a vle-insert: Would allow an item (or list of items) to be inserted at an index position. E.g.
(setq lst '
(1 2 3 4 5 6 7)) (vle-insert lst 4 'A) ;==> (1 2 3 4 A 5 6 7)
(vle-insert lst 3 '(A B)) ;==> (1 2 3 A B 4 5 6 7)
And a whole set of functions which would be a great help is some decent string handling routines which works on UniCode values as well (e.g. sending unicode strings to the vl-string-* functions result in "?" wherever non-ANSI characters are placed.) Also some sort of RegularExpression search / replace would be awesome!
