From: rpa@netcom.com (Ramin Firoozye)
Message-Id: <9310190147.AA24833@netcom2.netcom.com>
Subject: Re: Black Boxes
To: kevin@scic.intel.com (Kevin Altis)
Date: Mon, 18 Oct 93 18:47:13 PDT
In-Reply-To: <9310190105.AA19237@rs042.scic.intel.com>; from "Kevin Altis" at Oct 18, 93 6:01 pm
[Kevin Altis writes:]
>
> >debates on %20 vs. quotes and wrappers vs. terminating triple-colons
> >is like arguing about the color of internal engine components, irrelevant
> >and esoteric...
>
> Ramin,
> your comments on the class library will be good for discussion. I just
> wanted to point out that while some of the syntax issues may seem esoteric,
> they aren't! They are not the same as "arguing about the color of internal
> engine components" since everyone is going to read and type and quote these
> UR* identifiers every day. That's why it was important that the original
> URLs be "human readable" and that's why people like me don't want to see
> common characters have to be escaped, especially not just so they'll be
> pre-escaped for SGML. If UR* end up looking like math equations people are
> not going to use them or when they do they'll make lots of typos and all
> our efforts will be pointless. Right now, our domain of users is largely
> Unix based, but that is already changing quickly. UR* cannot be in a form
> that only a parser could love.
>
> ka
>
I completely agree with you, and realize the importance of the syntax,
especially the syntactic form that is exposed to the end-user.
My comment was more to make people realize there are more important
issues that need to be resolved before arguing on small syntactic
discrepancies.
The core of my proposal was to go ahead and create a "canonical" or
"internal" machine-friendly form, but also provide two-way mappings to
all sorts of external forms. One such mapping would be to human-readable
citations. Another would be to WWW namings. Yet another would be for
mapping to local system software (whether it's Unix, Mac, Windows,
or whatever).
My other main point was that when you have so many apparently conflicting
requirements, it's futile to discuss detail implementation issues. This
is why external interfaces become more important than internal formats.
As an application developer, I don't care what the exact TCP protocol
looks like. I *do* care that my platform comply with the Berkeley Socket
interface. If someone replaces the internal opcode settings of the
TCP protocol, I don't care. If they compress the packets, I don't care.
If they reverse the byte-order or change MTU sizes, I still don't care.
My stuff still works. This is the sort of behavior I think people are
asking for from URI's.
Cheers,
Ramin.
-- Ramin Firoozye' rp&A Inc. - San Francisco, California Internet: rpa@netcom.COM - CIS: 70751,252---- Ramin Firoozye' rp&A Inc. - San Francisco, California Internet: rpa@netcom.COM - CIS: 70751,252--
-- Ramin Firoozye' rp&A Inc. - San Francisco, California Internet: rpa@netcom.COM - CIS: 70751,252--