Subject: Re: [Gnu-arch-users] Re: File naming conventions
From: Zenaan Harkness <zen@cashpoolcomps.com>
To: arch <gnu-arch-users@gnu.org>
In-Reply-To: <buovfcy8eyf.fsf@mctpc71.ucom.lsi.nec.co.jp>
References: <1098141434.19575.42.camel@whiskas.cashpoolcomps.com>
	<20041018233511.GA31642@fencepost>
	<1098156854.19575.65.camel@whiskas.cashpoolcomps.com>
	<buok6tn1foz.fsf@mctpc71.ucom.lsi.nec.co.jp>
	<1098163496.19575.134.camel@whiskas.cashpoolcomps.com>
	<20041019060152.GC18852@wisq.net>
	<1098311382.11967.35.camel@nemesis.xlii.org>
	<1098313564.5336.29.camel@whiskas.cashpoolcomps.com>
	<vpqmzyhhrhi.fsf@ecrins.imag.fr>
	<1098319598.5336.46.camel@whiskas.cashpoolcomps.com>
	<20041021123218.GA30989@fencepost>
	<1098398014.5336.118.camel@whiskas.cashpoolcomps.com>
	<200410252009.i9PK9pP3066424@xl2.seyza.com>
	<buovfcy8eyf.fsf@mctpc71.ucom.lsi.nec.co.jp>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Message-Id: <1098771438.3124.23.camel@localhost.localdomain>
Mime-Version: 1.0
Date: Tue, 26 Oct 2004 16:17:18 +1000
Precedence: list
Sender: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Errors-To: gnu-arch-users-bounces+migo=homemail.com@gnu.org
Status: RO
Content-Length: 2800
Lines: 67
On Tue, 2004-10-26 at 11:07, Miles Bader wrote:
> Thomas Lord <lord@emf.net> writes:
> > Can't you use a filter program to generate the output format you want
> > from the output format provided?
> 
> This is perhaps harder than it should be because tla's output escaping
> format seems very complicated -- e.g., unicode encoding etc -- and it

Unicode encoding is actually a good thing. It would indeed be nice to
have a tla-encoding filter/munger to convert to "xargs" format. Keeps
everything nice and orthogonal etc...

> seems annoying to have to duplicate any decoding smarts that already
> have to be in tla.

Well, conversion between two unique encoding formats, one of which is a
logical superset of the other, technically does not "have to be in tla",
which I think is Tom's point.

Orthogonal filters/programs is fundamental to Unix design philosophy,
and probably due to my impatience last week, I seemed to totally miss
this point...

> What I'd personally like (i.e., it would be convenient for me) is a
> dumbed down "half-encoded" output format: use pika escaping for a small
> fixed set of problematic characters (e.g. NUL, SPC, NL, CR, backslash
> [to protect the pika escapes]), but for anything else, output decoded
> "native" characters.

Sort of like UTF-8 (kinda sorta...).

I would argue against YAEF (yet another encoding format). It is best to
have a canonical format, that can cater to all languages (ie unicode
aware) and then convert to and from that format as-needed.
That's The Unix Way (TM).

> [I've posted this before, but nobody seemed to care; won't anybody think
> of the scripters... :-...? ]

I am a scripter, and scripting is the exact reason I brought this up. So
I'm fully on the same page as you here. And understand that
orthogonality does a few things:
 - minimizes the namespace (eg. "random" options on every other app,
   instead use a single filter (or two))
 - by doing so, maximizes the inter-application connections _possible_
 - thereby maximizes utility of the system overall
 - (oh, and of course, all this simplifies writing scripts :)

Seriously, it is The Right Thing. It's just that some may need a quick
fix, or have to make do without until the filters are available. I can
fully empathize with Tom (or anyone really) not wanting to duplicate
functionality, or "mess up" a perfectly clean (and orthogonal)
codespace, namespace and options space (backwards compatibility
expectations and the like). Of course it's easy for me to say all this
now since 1) it doesn't effect me, 2) I'm in a good mood today :).

regards
zen


_______________________________________________
Gnu-arch-users mailing list
Gnu-arch-users@gnu.org
http://lists.gnu.org/mailman/listinfo/gnu-arch-users

GNU arch home page:
http://savannah.gnu.org/projects/gnu-arch/