Newsgroups: comp.lang.smalltalk
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!hookup!nstn.ns.ca!dragon.acadiau.ca!car407.acadiau.ca!ivan
From: ivan@ace.acadiau.ca (IVAN TOMEK)
Subject: Re: Standards and "Private" methods
Message-ID: <ivan.117.2FD59252@ace.acadiau.ca>
Lines: 49
Sender: news@relay.acadiau.ca
Nntp-Posting-Host: car407.acadiau.ca
Organization: Acadia University
X-Newsreader: Trumpet for Windows [Version 1.0 Rev B final beta #1]
References: <3qga8f$d2t@pipe6.nyc.pipeline.com> <3qjbvq$24t@ocean.CAM.ORG> <1995Jun1.182047.8501@medicus.com> <3qm152$gfa@warp.cris.com> <3qnfm4$bi9@ornews.intel.com> <3qokmf$ec8@warp.cris.com> <3quftm$kkc@zen.hursley.ibm.com>
Date: Wed, 7 Jun 1995 11:49:38 GMT
Lines: 49

In article <3quftm$kkc@zen.hursley.ibm.com> Paul_Gover@uk.ibm.com writes:
>Path: dragon.acadiau.ca!nstn.ns.ca!newsflash.concordia.ca!uunet!usc!howland.reston.ans.net!math.ohio-state.edu!jussieu.fr!oleane!pipex!hursley.ibm.com!news
>From: Paul_Gover@uk.ibm.com
>Newsgroups: comp.lang.smalltalk
>Subject: Re: Standards and "Private" methods
>Date: 5 Jun 1995 08:40:22 GMT
>Organization: IBM Warwick Development Group, UK
>Lines: 24
>Message-ID: <3quftm$kkc@zen.hursley.ibm.com>
>References: <3qga8f$d2t@pipe6.nyc.pipeline.com> <3qjbvq$24t@ocean.CAM.ORG> <1995Jun1.182047.8501@medicus.com> <3qm152$gfa@warp.cris.com> <3qnfm4$bi9@ornews.intel.com> <3qokmf$ec8@warp.cris.com>
>Reply-To: Paul_Gover@uk.ibm.com
>NNTP-Posting-Host: plum.wdg.uk.ibm.com
>X-Newsreader: IBM NewsReader/2 v1.01


>In <3qokmf$ec8@warp.cris.com>, Douglas Camp <dcamp@acm.org> writes:
>>pdlogan@ornews.intel.com (Patrick Logan) wrote:

>> ...
>>I think that, if the vendors could provide an implemention
>>of 'private' methods, which didn't immediately break all existing
>>code - something which would allow developers to mark
>>methods as truly private if they chose to do so - I think that 
>>would be useful. At the least, it would avoid the sort of ad-hoc
>>attempts to enforce private methods which (I think) started this 
>>thread.
>>
>>...

>What worries me about enforcing "private"  is that I may know enough
>about the private method to want to use it myself.  I think the Smalltalk code I've
>seen tends to apply "Private" designations in various inconsistent ways, usually
>meaning "we don't think you should want to do this", but not necessarily "it's
>dangerous to do this".  Of course, the next step is to make the private method's
>implementation private - ie, ship no source <g>.  I'm happy, overjoyed even, if the
>code _warns_ me not to use it, but displeased if it _forces_ me...

>Paul Gover, IBM Warwick Development Group.
>X.400: G=paul; S=gover; P=ibmmail; A=ibmx400; C=gb
>I speak for myself, not IBM, of course.

I think that access to all code is one of ther greatest features of ST and I 
am against making any methods inaccessible. I have on several occasions used 
methods labeled as private because I needed them - I think that the designer 
cannot know for certain which of the methods that he or she writes may be 
needed by others in the future.

Ivan Tomek

