octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

problems with optimset option name completion in 3.8.0


From: Olaf Till
Subject: problems with optimset option name completion in 3.8.0
Date: Fri, 3 Jan 2014 23:34:46 +0100
User-agent: Mutt/1.5.21 (2010-09-15)

Wish I had realized this before the 3.8.0 release.

'optimset' since 3.8.0 completes option names if only their first
characters are given and warns if these characters are ambiguous.

But up to now it was legitimate to have e.g. both the following (full)
option names:

'abc'
'abcde'

Some newer optimization frontends in the optim package use optimset
and I had chosen options for them as in the above artificial
example. Now I get warnings like these:

optimset ("dfdp", "a")
warning: ambiguous option: dfdp (dfdp, dfdp_pstruct)
ans =

  scalar structure containing the fields:

    dfdp = a

"dfdp" and "dfdp_pstructs" are not the only options with this problem
in these functions of the optim package, there are many.

Now, apart from just let the user turn off the warning (but that does
not seem to be possible at the moment), I see three possibilities:

1) consider it illegitimate to have option names whose first
   characters are equal to a different full option name (means I have
   to change the option names in the said functions of the optim
   package!),

2) change optimset so that it always matches an option name if it is
   given in full, even if there is a longer option name which starts
   with the same characters,

3) drop option name completion in optimset.


Personally, I'd think that 1) is a non-intuitive restriction, while 2)
seems a bit bizarr. I fear I'd favour 3), because I don't see so much
benefit from option name completion that it is worth these problems.

Some decision should be made soon, for a new optim package version has
to be made which doesn't duplicate fminsearch (which is in Octave
since 3.8.0).

Olaf

-- 
public key id EAFE0591, e.g. on x-hkp://pool.sks-keyservers.net

Attachment: signature.asc
Description: Digital signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]