Makes sense to me. I will convert the tests to use the surviving models. Then we can either remove these now or in the next release.
-----Original Message----- From: Ted Dunning [mailto:[email protected]] Sent: Sunday, April 24, 2011 10:40 AM To: [email protected] Subject: Re: Can L1Model, NormalModel, AsymmetricSampledNormalModel be removed? Might be worthwhile targeting them at the surviving classes you mean? On Sat, Apr 23, 2011 at 11:31 PM, Grant Ingersoll <[email protected]>wrote: > Is there anything worthwhile in the tests to keep? > > > On Apr 23, 2011, at 11:48 PM, Jeff Eastman wrote: > > > I marked them as deprecated recently. They are superceded by models > within the Cluster hierarchy and were only used in early testing of > Dirichlet, afaict. I think the GaussianCluster and DistanceMeasureCluster > can replace them all. > > Jeff > > > > > > On 4/23/11 11:41 AM, Ted Dunning wrote: > >> Let me take a look. > >> > >> Sent from my iPhone > >> > >> On Apr 23, 2011, at 1:31 AM, Sean Owen<[email protected]> wrote: > >> > >>> Someone marked them deprecated recently. And they are not used except > in > >>> tests and a corresponding "Distribution" class, which are likewise > unused > >>> AFAICT. > > > > >
