From tpeters@mixcom.com Wed Dec 24 15:24:43 2008
From: tpeters@mixcom.com
To: test-drb@ccmp.vtda.org
Subject: [personal] Re: PDF datasheets (was Re: Sources for 8b TTL
 keyboards(Keytronics))
Date: Wed, 24 Dec 2008 15:24:43 -0600
Message-ID: <5.1.0.14.2.20081224151948.00c0b9c0@localhost>
In-Reply-To: <20081224123644.G90174@shell.lmi.net>
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="===============1440356232190308756=="

--===============1440356232190308756==
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable

At 01:04 PM 12/24/2008 -0800, you wrote:
> > > I wish that I knew a good way in HTML to express offsets from beginning
> > > of a document.
>
>On Wed, 24 Dec 2008, Mike Hatch wrote:
> > It is a variant of the <a href=3D tag
> > To another page it would be <a href=3D"pagename.htm">, for a position in =
that
> > page it would be <a href=3D"pagename.htm#position">, where 'position' is =

> a tag
> > in that page <a name=3D"position">.
>
>Yes, you can ADD 'position's to get a form of symbolic labeling.
>That is a very useful feature for identifying positions within your own
>files.  (Although it left a bit to be desired when implementing full
>indices.  That is only for before-the-fact access.
>
>I don't own every website that I reference.  In some of my indexing
>projects, instead of storing both a URL AND a file offset, I would like
>a URL that dereferences to an offset within the file, such as
>"notice the wording where he says 'xxx' (684 bytes into that file)".
>http://www.foo.bar/RFC.html*684
>(where '*' is some unique punctuation provided for in the HTML spec)
>"Dear webmaster, is that a type at xxxx.html/*yyy ?"
>
>
>Unfortunately, as far as I can tell, the current implementations of HTML
>(and SGML?) do not provide an "offset" operator.  If I am wrong, please
>tell me!  It took me weeks to find <pre></pre> to be able to do
>non-proportional ASCII art, and to be able to create websites that discuss
>HTML.
>
>Inclusion of an "offset operator" would permit handing a URL to somebody
>and having THEIR browser take them to the desired location within the
>file, WITHOUT edit rights to the file.
>
>It would permit creating an index to a document WITHOUT requiring
>annotation of the original, nor captive environment for display.
>The pointers within the index could then be actual access URLs
>"RFC.html*684"  or http://www.foo.bar/RFC.html*684
>I do not like providing a proprietary viewer that must be used to be able
>to see the referenced items within the file.
>
>
>BTW, I do NOT see variations in word size or character set, as being
>relevant.

HTML can be rendered so radically different on various platforms. What=20
would the offset measure? Bytes? Characters? (those two would be very=20
different) Lines? How could you possibly compute that? Would a byte offset=20
be meaningful? A byte offset that lands you in the middle of a script or=20
code of some other type would be the same screen position as a byte offset=20
that's hundreds of bytes (or thousands) different; how would that be=20
useful? Ever see a web page rendered in Lynx? It's just text and nothing more.

The whole point of a web browser is that the content is rendering engine=20
independent. I think that some sort of offset is a step backwards and is=20
unlikely to be implemented.

Just my $0.02




-----
311. [Philosophy] No, I don't think smoking should be illegal. Smokers are
already getting the death penalty! --from Stev0's "rant of the week" on
smoking
--... ...--  -.. .  -. ----. --.- --.- -...
tpeters at nospam.mixcom.com   (remove "nospam") N9QQB (amateur radio)
"HEY YOU" (loud shouting)    WEB: http://www.mixweb.com/tpeters
43? 7' 17.2" N by 88? 6' 28.9" W,  Elevation 815',  Grid Square EN53wc
WAN/LAN/Telcom Analyst, Tech Writer, MCP, CCNA, Registered Linux User 385531

--===============1440356232190308756==--