Mailing-List: contact dev-help@commons.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Commons Developers List" <dev@commons.apache.org>
Received-SPF: pass (nike.apache.org: domain of sebbaz@gmail.com designates
 74.125.44.156 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:to:subject:in-reply-to:mime-version
         :content-type:content-transfer-encoding:content-disposition
         :references;
        b=j4P0vOgNShHiOCTMm2cbSkZy/3moOyRJXILsvtMr8doCVQQemUfEc6wYqQ0htxeTjE
         uWh1dE+lg687EpBAWqOEPuhy6sppJqdLeA7AyB+VKj0k6/D0dEqHlMLZiaDuGy9tVpku
         03BvYGKUuGAvKSxThARcG0JNy1de9p5PsbxYU=
Message-ID: <25aac9fc0901020806j5c0b6fdy3e8cea14b8f9b9fd@mail.gmail.com>
Date: Fri, 2 Jan 2009 16:06:40 +0000
From: sebb <sebbaz@gmail.com>
To: "Commons Developers List" <dev@commons.apache.org>
Subject: Re: [math] Shoud OpenIntToDoubleHashMap get return NaN instead of 0
 for key not found?
In-Reply-To: <495E3875.8040608@free.fr>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
References: <495E2834.9060409@gmail.com> <495E3875.8040608@free.fr>

On 02/01/2009, Luc Maisonobe <Luc.Maisonobe@free.fr> wrote:
> Phil Steitz a =E9crit :
>
> > I am still working through this class and the sparse matrix class that
>  > it was extracted from (thanks, Ismael and Sugit!), so I am not sure if
>  > changing this would cause problems, but the current setup (returning 0
>  > for missing keys) limits usefulness of this class.   I see how this is
>  > convenient for sparse matrices; but I would see NaN as a more natural
>  > return value for non-existent keys in the general case.  Alternatively=
,
>  > I guess we could add another method get(int key, double missingReturn)=
.
>  >
>  > Thoughts?
>
>
> I had exactly the same thought while extracting the class.
>  I also prefer to use Double.NaN for numbers that have never been
>  initialized explicitly, but I also understand 0 is more logical in the
>  special case of sparse matrices.

+0 (I never used Math, but seems sensible)

>  What about having a configurable value for missing entries ? It should
>  probably be configured at construction time (with a default value to
>  Double.NaN if not specified) and never changed afterwards. In the case
>  of sparse matrices, we should configure this value to 0.0.

+1 to never changing the value - making it final would be best for
thread safety.

>
>  Luc
>
>
>  >
>  > Phil
>  >
>  > ---------------------------------------------------------------------
>  > To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>  > For additional commands, e-mail: dev-help@commons.apache.org
>  >
>
>
>  ---------------------------------------------------------------------
>  To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>  For additional commands, e-mail: dev-help@commons.apache.org
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org