Message-Id: <9310181943.AA16363@rs042.scic.intel.com>
Date: Mon, 18 Oct 1993 12:39:25 -0800
To: Terry Allen <terry@ora.com>
From: kevin@scic.intel.com (Kevin Altis)
Subject: Re: <'s for URLs
At 12:06 PM 10/18/93 -0700, Terry Allen wrote:
>Do you have a text editor that does automatic conversion of angle
>brackets to character entities? I sure don't, and as *any other
>choice of character but < > will eliminate the problem* I
>don't see why this oversight can't be remedied.
No, I don't have an editor today that recognizes angle brackets, but I
expect to. We could use some characters other than < and > if that solves
an immediate problem. I agree that < > and : get used for way to many
things, but the question is whether using & or some other character makes
detection any easier; if you used & based on your earlier example,
>So now your HTML text says <URL:foo&rt;. Why not do the
>simple thing and choose delimiters that don't create a widespread
>problem?
wouldn't you get &URL:foo& and is that any better?! Remember, the
brackets (whatever brackets they end up being) are for PLAIN TEXT, not HTML
(SGML). Regardless of what you have in plain text, when you put it into an
HTML document you're going to have to do some character conversion and
document validation. You're conversion tools (texttohtml, rtftohtml...) or
editing software should help with or hide this conversion.
In my native environment, when I paste styled text into a plain text field,
the formatting is stripped out, when I copy a picture and then paste it
somewhere else, it is often converted automatically if needed, when I copy
a set of spreadsheet cells and paste them into another non-spreadsheet
document, they are converted to a valid table form within that application.
We shouldn't ruin URLs or UR*s in general by tying them exclusively to use
within SGML! Within SGML, an URL has to meet certain guidelines, including
the addition of <A HREF=...> or some such tag, but within plain text, or
gopher+, or news, or portable document format the URL only has to conform
to the URL specification, not the HTML or SGML spec! Even if the HTML
(SGML) tools never got any better, more people (by an order of magnitude at
least) will be using URLs and UR*s in plain text or other non-HTML formats
than people will be entering UR*s into HTML.
ka