Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!mstar!mstar.morningstar.com!bob From: bob@MorningStar.Com (Bob Sutterfield) Newsgroups: news.software.b Subject: Re: Time for 8 bit news, isn't it?????. Message-ID: Date: 24 Jul 90 14:05:57 GMT References: <777@hades.ausonics.oz.au> <15688@bfmny0.BFM.COM> <1990Jul21.054016.10409@looking.on.ca> <15692@bfmny0.BFM.COM> Sender: usenet@MorningStar.COM (USENET Administrator) Reply-To: bob@MorningStar.Com (Bob Sutterfield) Organization: Morning Star Technologies Lines: 39 In-Reply-To: tneff@bfmny0.BFM.COM's message of 23 Jul 90 06:55:12 GMT In article <15692@bfmny0.BFM.COM> tneff@bfmny0.BFM.COM (Tom Neff) writes: In article <1990Jul21.054016.10409@looking.on.ca> brad@looking.on.ca (Brad Templeton) writes: If I'm going to lose parts of a multi-part binary, I may as well lose the whole thing. I agree completely. Binaries shouldn't be split. If a binary is bigger than 50K, it has NO BUSINESS BEING BROADCAST AS NEWS! Post a pointer for FTP and anonymous UUCP, and let those who want it pay for it. THAT'S how Usenet ought to work. If the binary thing is a computer program, it has no business being transmitted as news. Binaries are inherently nonportable, and of little value to a majority of sites passing them through. If the binary thing is a multimedia document, then (at the user interface level) it should certainly appear in one chunk. ...when a properly packaged source archive (i.e., what Usenet SHOULD be broadcasting)... Yes, if the thing you're talking about is a program for a computer, it certainly should be distributed as source. Does anyone actually trust an encoded binary that they found in some newsgroup? How quaint, how naive! However, there are (as has been abundantly pointed out) plenty of examples of things that aren't program binaries but that still should be transmissible via news-like mechanisms and that break the current news implementations. ...arrives in a dozen pieces, installation is VERY easy on most platforms. When a part is lost or corrupted, the penalty for rebroadcast is a lot smaller than it would be if the whole kit had to be resent. Sequencing and reassembly are problems for the session layer, not the user interface layer. Users should never see that a document was split into transport layer-sized chunks, which is what the 50K article-size limit really is.