Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!crdgw1!uunet!intercon!news
From: kdb@macaw.intercon.com (Kurt Baumann)
Newsgroups: comp.protocols.appletalk
Subject: Re: MacNFS vs AppleShare
Message-ID: <27486CC5.1A75@intercon.com>
Date: 19 Nov 90 23:37:41 GMT
References: <9011161241.aa27485@Mercury.TWG.COM>
Sender: usenet@intercon.com (USENET The Magnificent)
Reply-To: kdb@macaw.intercon.com (Kurt Baumann)
Organization: InterCon Systems Corporation, Herndon, VA
Lines: 91


Most of the below I have already answered in another note, but I thought
that I would respond to some of the TWG responses directly.

In article <9011161241.aa27485@Mercury.TWG.COM>, lefty@TWG.COM writes:
> Steve Dorner writes:
> >
> >In article <27430546.6B52@intercon.com> kdb@macaw.intercon.com (Kurt Baumann) 
> >writes:
> >>Yes.  Both can be changed in an AppleSingle file.
> >
> >With GREAT pain and suffering, right?  Don't you have to rewrite practically
> >the entire file in the worst case?  So the entire file has to traverse
> >the network twice (or does the NFS protocol allow you to tell the server to do
> >some things locally?).
> >
> >AppleSingle is an icky way to run a filesystem, unless you KNOW you will only
> >be playing with one fork, or so it seems to me.
> 
> Yeah, we pretty much agreed with you on this one.  Too much overhead and
> agita involved in supporting AppleSingle.  Bottom line: we dont do it.  If
> theres sufficient demand (we havent seen it yet) well consider it, but I
> think a better solution would be a UNIX utility to convert those pesky
> AppleSingle files into AppleDouble format.
> 

First, we feel and will always feel that people using a Macintosh don't really
want to be bothered with the nitty grity, they want something that works,
looks, and acts like it should on a Mac.  To those ends, we would NEVER in
a million years consider having a Mac user have to log onto UNIX and manipulate
files with a "pesky" utility. :-)  There are those users who are able to
use both UNIX and a Mac, but the majority of users only know how to use a
Mac.

Second, why didn't you at least give the user the ability to read AppleSingle?

> >>> If so, how do you decide when to do the translation?
> >>Yes, if it is a Macintosh text file we translate CR-to-LF when output and
> >>reverse on input.  The UNIX text files are also treated this way.
> >
> >Uh, how do you decide if a UNIX file is a text file?  Or a mac file for that
> >matter; there are some applications that create text files that have creators
> >other than 'TEXT' (TeachText comes to mind)?
> 
> Again, our product doesnt fiddle with EOL conversion for exactly this
> reason.  I would imagine that Intercon is using the execute bit to
> "determine" what is and isn't a text file; in my view this is both suspect
> and risky.  We offer a TeachText work-alike called NetText, which tries to
> be a little intelligent about EOL conversion: it determines the EOL
> convention in force and will translate to another if requested to do so. 
> Safer all around, to my way of thinking.
> 

The only way you could read a Mac text file written by TWG's NFS would
be to do the following:

a. convert the file
b. edit the file
c. convert the file back

(ever try reading a file with VI that contains too many characters on a single
line?)

The way we do this, all you need to do is step b (edit the file).

I can see why you thought this, as this is they way TWG does it.  We don't 
look at the execute bit as many scientific datasets do not have the bit set, 
yet are binary; and some files have the bit set but are text files (such as  
login, etc.).

NetText is, IMHO, a hack.  Why would a Mac user want to use a special
text editor to edit files on the NFS side, when they should be able to use
the editor that they are familiar with.  For exactly this reason, we decided
to provide invisible text file conversion.

> >Now, for my own question: Does your product work with MacTCP?  If it
> >does, I may be a customer.  If it doesn't, I have suddenly lost interest...
> 
> Of course it does.  I wouldnt have it any other way.
> 

Neither would Apple :-).

I am not trying to start a flame war, but I did want to point out a fews
things here, that were supposition on TWG's part.  Hopefully this helps to
clarify those few guesses.

--
Kurt Baumann                       InterCon Systems Corporation
703.709.9890                      Creators of fine TCP/IP products
703.709.9896 FAX               for the Macintosh.