Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id EAA13195; Thu, 18 Jan 1996 04:09:09 -0500 Received: by CS.UTK.EDU (bulk_mailer v1.3); Thu, 18 Jan 1996 04:08:45 -0500 Received: from ester.dsv.su.se by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id EAA13177; Thu, 18 Jan 1996 04:08:43 -0500 Received: (from jpalme@localhost) by ester.dsv.su.se (8.7.1/8.7.1) id KAA07805; Thu, 18 Jan 1996 10:08:39 +0100 (MET) Date: Thu, 18 Jan 1996 10:08:38 +0100 (MET) From: Jacob Palme To: ietf-drums Subject: Re: "be liberal in what you accept" considered harmful Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII RANDY@MPA15AB.MV.UNISYS.COM wrote: > > "Be liberal in what you accept" can actually be a quite harmful > > practice. It's use papers over problems which otherwise would be > > noticed and fixed. > > (John goes on to provide examples of how being "liberal" can be > dangerous, and one where being "conservative" resulted in widespread > broken software being fixed.) > > This is a good point, but if someone implements a new mail system > today, and rejects messages that do not conform to the standards, it > will be seen as unable to interoperate. Wbat's a software developer > to do? The best answer to your question is "a software developer should follow the standards". Only by specifying in the standards in what cases you should be liberal, and in what cases you should not be liberal, can we get convergent software. This is what I said when this discussion started: The liberal/conservative commandmend is bad because it is vague and does not tell you when to be liberal and when to be conservative. ------------------------------------------------------------------------ Jacob Palme (Stockholm University and KTH) for more info see URL: http://www.dsv.su.se/~jpalme