Path: utzoo!attcan!uunet!munnari!bhpese!andrew
From: andrew@bhpese.bhpese.oz (Andrew Steele)
Newsgroups: comp.sys.amiga
Subject: Re: Filtering the source/binaries without bottleneck
Summary: Comments for a standard posting format
Message-ID: <158@bhpese.bhpese.oz>
Date: 10 Jan 89 00:57:16 GMT
References: <34235@bbn.COM> <10831@s.ms.uky.edu> <279@lakesys.UUCP>
Organization: BHP ESE Newcastle, Australia
Lines: 59


	My Comments on the subject :-

I think Bob Page is doing a great job, keep it up.

The frustration with things not working would be alleviated if the time
	it takes to cut above a non standard cut line, unshar, join, uudecode
	and unzooing a file could be reduced.  If this whole process could 
	be automated then people would not waste large amounts of time getting
	files to a usable form.  I realise that some things to do this have
	come over the net of late but they either were OS dependant
	(BSD/SysV), or only worked for some cases, or both.

I generally get all the parts for a program and put them into a temperary
	directory before I start undoing them.  All programs I would think
	should only assume that they are to go into the current directory
	and if a program requires subdirectories it should only assume that
	they are to be created below the current directory.

The comment made that a piece of source code should be used to teach ideas
	is good, if it works then that is a bonus.

The idea of a test site already exists, its called the net.  If something
	doesn't work it doesn't take long to hear about it.  In the case of 
	a program like rgb if you don't know whether it's worth the time to
	download, wait a week ( usually less ), if there are no comments
	about it and you are still not sure try asking for comments on it.
	There is no need to inflict this work on the moderator.

The idea of bit bucketing anything that hasn't come back from a test site
	within a week or so is not a good idea.

MY SUGGESTION :
	Someone ( Probably Bob Page as he is in the position to do it ),
	should come up with a standard packaging method for files and
	an accompanying algorithm to undo them.  The shar format is almost
	right but it could be improved.
		The cut line should always be EXACTLY the same to allow
			for it to be searched for.
		A header to the first/only file should list the parts
			needed for the complete package, where they are
			( comp.binaries.amiga/comp.sources.amiga ), and 
			what format they are in ( .uu/.zuu/.zoo/need joining
			before undoing/etc. ).
		An simple algorithm to go from one/many news items to
			one .zoo file that contains all parts (binaries,
			sources,docs,etc) assuming all required news items
			are present in the same place.  Actual code for
			such an undoer would be nice but may be at the
			expense of portability (need at least a version
			for U*NIX BSD & Sys V and the Amiga).

Any other comments ?

Andrew Steele                      _____     Spengat Technologies, 
                                  /_   _\    c/o Electrical Computer Services,
ACSnet  : andrew@bhpese.oz        __| |__    BHP Rod & Bar Products Division,
INTERNET: andrew@bhpese.oz.au    /__| |__\   Newcastle, NSW, Australia.
UUCP    : ...!{uunet,mcvax}!munnari!bhpese.oz!andrew